SlideShare ist ein Scribd-Unternehmen logo
1 von 59
Downloaden Sie, um offline zu lesen
© 2017 by Daffodil International University
Computation of Multi-Agent Based Relative
Direction Learning Specification
Submitted By
S. Rayhan Kabir
ID: 133-35-561
Bachelor of Science in Software Engineering
Department of Software Engineering
Daffodil International University
rayhan561@diu.edu.bd
Supervised by
Dr. Shaikh Muhammad Allayear
Associate Professor
Department of Software Engineering
Associate Professor and Head
Department of Multimedia & Creative Technology
Daffodil International University
Submitted Date: September 2017
Daffodil International University
Dhaka, Bangladesh
I
© 2017 by Daffodil International University
DECLARATION
It hereby announces that, this bachelor thesis under the supervision of Dr. Shaikh Muhammad
Allayear, Associate Professor, Department of Software Engineering, Daffodil International
University. It is also declared that neither this thesis nor any part of this has been submitted
elsewhere for award of any degree.
Submitted by:
………………………………………
S. Rayhan Kabir
ID: 133-35-561
Batch: 12th
Department of Software Engineering
Daffodil International University
Faculty of Science & Information Technology
Daffodil International University
Certified by:
………………………………………
Dr. Shaikh Muhammad Allayear
Associate Professor
Department of Software Engineering
Associate Professor and Head
Daffodil International University
Department of Multimedia & Creative Technology
Faculty of Science & Information Technology
II
© 2017 by Daffodil International University
APPROVAL
This bachelor thesis titled “Computation of Multi-Agent Based Relative Direction Learning
Specification”, submitted by S. Rayhan Kabir, ID: 133-35-561 to the Department of Software
Engineering, Daffodil International University has been accepted as satisfactory for the partial
fulfillment of the requirements for the degree of B.Sc. in Software Engineering (SWE) and
approved as to its style and contents.
BOARD OF EXAMINERS
-----------------------------------------------
Dr. TouhidBhuiyan
Professor and Head
Department of Software Engineering
Faculty of Science and Information Technology
Daffodil International University
Chairman
-----------------------------------------------
Dr.Md. Asraf Ali
Associate Professor
Department of Software Engineering
Faculty of Science and Information Technology
Daffodil International University
Internal Examiner 1
-----------------------------------------------
Manan Binth Taj Noor
Lecturer
Department of Software Engineering
Faculty of Science and Information Technology
Daffodil International University
Internal Examiner 2
-----------------------------------------------
Dr. Md. Nasim Akhtar
Professor and Chairman
Department of Computer Science Engineering
Dhaka University of Engineering & Technology, Gazipur
External Examiner
III
© 2017 by Daffodil International University
Abstract
The most widely recognized relative directions are left, right, up, down, backward and forward.
This research paper presents another algorithm for computing the relative directions between two
agents, where one agent can learn another agent’s relative directions. We exhibit a study
contrasting direction construct guidelines and relative direction instructions with respect to
people on foot in a genuine city condition, measuring both goal and subjective achievement.
Eyewitnesses commonly depict their condition by determining the relative directions in which
they see different items or other individuals from their perspective. Be that as it may, it is
surprisingly difficult to integrate relative directions got from various observers between two
agents. In this paper, we introduce a novel subjective portrayal and representation of this
MultiAgent Relative Direction (MARD) algorithm can solve these problems. For handling and
recognizing relative direction, its executed work method or computation sets aside more
opportunity for the result. Its actualized computation has few stages for distinguishing relative
direction. Endeavor to diminish one stage and expected the outcome will be quicker.
IV
© 2017 by Daffodil International University
Acknowledgments
Firstly, I might want to thank my supervisor, Associate Professor Dr. Shaikh Muhammad
Allayear. I owe such a great amount to his motivating direction over the span of this venture, for
his recommendations on papers to peruse, and for his endless hours of accommodating
exchanges and assessment. He gives me an opportunity to work in Smart Data Science Center
(SDSC) for complete my research. SDSC is a computer research laboratory of Daffodil
International University. I might likewise want to demonstrate appreciation to my committee,
including Associate Professor Dr. Md. Asraf Ali, Chairman, Project/Thesis Committee,
Department of Software Engineering and my noteworthy our department head Professor
Dr.Touhid Bhuiyan for their profitable instructions.
All the more by and large, I can't exaggerate the amount Daffodil International University's
software engineering offices have helped me develop as an understudy. Uncommon much
gratitude goes Assistant Professor Imran Mahmud for putting me on the way to seeking after
hypothetical software engineering research and for being a uniquely rousing coach and to
lecturer Ms. Manan Binth Taj Noor for filling in as my scholarly consultant. I might likewise
want to thank lecturer Ms. Fouzia Rahman for teaching one incredible course that truly got me
amped up for a few complex factors. I would also like to thank lecturer Ms. Rubaida Easmin
for her incredibly extensive and important input on the evidence of my principle result. I'm
particularly appreciative for Research Associate MD. Tahsir Ahmed Munna and lecturer
Mirza Mohtashim Alam for being regular teammates on issue sets. I'm particularly thankful for
my paper observers Assistant Professor K. M. Imtiaz-Ud-Din and lecturer Md. Anwar Hossen
for being continuous collaborators on issue sets and their suggestions on research report to
peruse or read.
Some extended and customized parts of this thesis recently got an opportunity to present at
International Conference on Intelligent Sustainable Systems (ICISS 2017), in India and
publisher is IEEE Xplore [35]. There I have described how intelligent computer learn and
identify human's relative directions which based on this thesis dissertation. I hope I can
successfully showcase my research in this conference.
Lastly, I might want to thank my parents for bringing me into this world and making everything
conceivable. They were the reason I initially began to look all starry eyed at learning, and I am
appreciative consistently for what they have done to raise me up to be simply the best form.
V
© 2017 by Daffodil International University
Table of Contents
Abstract......................................................................................................................................... III
Acknowledgments.........................................................................................................................IV
Introduction..................................................................................................................................... 1
1.1 Overview ........................................................................................................................... 1
1.2 Research Objectives .......................................................................................................... 2
1.3 Research Questions............................................................................................................ 2
1.4 Organization ...................................................................................................................... 3
1.5 Definitions ......................................................................................................................... 4
1.6 Motivation of Research ..................................................................................................... 4
Background and Literature Review ................................................................................................ 6
2.1 Relative Direction Fundamentals ...................................................................................... 6
2.2 Multi-Agent System Environment..................................................................................... 7
2.3 Previous Research and Work............................................................................................. 8
2.4 Left-Right Confusion....................................................................................................... 13
2.4.1 Research about Left-Right Confusion........................................................................ 14
2.4.2 Artificial Intelligence Perspective .............................................................................. 14
Proposed Algorithm Model........................................................................................................... 18
3.1 Induction.......................................................................................................................... 18
3.2 Tracking Agent and Handoff Agent ................................................................................ 19
3.2.1 Direction Points.......................................................................................................... 19
3.2.2 Various 3D Aspects.................................................................................................... 20
3.2.3 Realistic Inputs ........................................................................................................... 21
3.3 Mathematical Exploration ............................................................................................... 22
3.4 Structure of Algorithm..................................................................................................... 22
3.5 Identify Relative Direction .............................................................................................. 25
3.6 Machine Learning Aspect................................................................................................ 26
3.7 Alternative MARD Algorithm Approach........................................................................ 26
3.8 Multi-Agent Route Direction........................................................................................... 31
Methodology................................................................................................................................. 33
4.1 Algorithm Engineering Method....................................................................................... 33
4.2 Experiment Control ......................................................................................................... 35
Results and Analysis..................................................................................................................... 36
5.1 Accuracy Result............................................................................................................... 36
5.2 Comparison...................................................................................................................... 40
5.2.1 Comparison with UAV Relative Attitude Estimation Algorithm .............................. 40
5.2.2 Comparison with Others Algorithms.......................................................................... 42
Discussion..................................................................................................................................... 44
6.1 Summary.......................................................................................................................... 44
6.2 Conclusion....................................................................................................................... 45
6.3 Future Work..................................................................................................................... 46
Bibliography ................................................................................................................................. 47
VI
© 2017 by Daffodil International University
List of Tables
Table 1: Relative directions and their numerical values..................................................... 21
Table 2: Condition of Cases................................................................................................ 27
Table 3: Order of Cases. ..................................................................................................... 28
Table 4: Accuracy result ..................................................................................................... 39
Table 5: Comparison of case solution between two algorithms. ........................................ 40
Table 6: Multi-agent based and single agent based algorithm............................................ 43
VII
© 2017 by Daffodil International University
List of Figures
Figure 1: Different types of relative directions. ................................................................ 7
Figure 2: Characteristics of Multi-agent System Environment......................................... 8
Figure 3: Tracking UAV and Handoff UAV.................................................................... 9
Figure 4: Relative directions (right and left) of two agents............................................. 15
Figure 5: Agent A ordered agent R to find the mobile object ................................... 16
Figure 6: Agent A calculated that his right direction is equal to agent .......................... 16
Figure 7: Agent A ordered agent R ................................................................................. 17
Figure 8: Direction points of Handoff Agent and Tracking Agent. ................................ 19
Figure 9: Different cases of tracking agent in 3D aspect. ............................................... 20
Figure 10: Tracking agent's direction identification........................................................ 25
Figure 11: Multi-agent based route direction in mapping. .............................................. 32
Figure 12: Algorithm Engineering methodology cycle................................................... 34
Figure 13: Comparison graph of case solution between two algorithms. ....................... 41
Figure 14: Use of area between two algorithms.............................................................. 41
VIII
© 2017 by Daffodil International University
List of Algorithms
Algorithm 1: Multi-Agent Relative Direction Algorithm.................................................. 24
Algorithm 2: Direction Identification Algorithm............................................................... 26
Algorithm 3: Alternative MARD Algorithm...................................................................... 30
1
© 2017 by Daffodil International University
Chapter 1
Introduction
The most generally perceived relative directions are left, right, up, down, backward and forward.
A man giving relative directions will utilize center terms. A multi-agent framework is an
automated framework made out of different associating keen operators inside a domain. Multi-
agent frameworks can be utilized to take care of issues that are troublesome or inconceivable for
an individual agent or a solid framework to understand. This paper explores another
computation for assessing the relative direction between two agents.
1.1 Overview
Learning and identify relative directions in the multi-agent based system depends on three pairs
of relative directions which are forward and backward, left and right, up and down. The
experiment or research of multi-agent based relative direction learning the algorithmic process
for dealing with computational issues, while one agent wants to learn another agent's relative
directions. In here we show a new algorithm for processing the relative directions between two
agents, how one agent can learn another agent’s relative headings in 3D perspective. In this
paper, we introduce about Multi-Agent Relative Direction (MARD) algorithm concept which can
represent this issue.
Since the disclosure of "Plaid Motion Coherence on Component Grating Directions" by
Jeounghoon Kim and Hugh R.Wilson in 1993 and its understanding unmistakably demonstrates
that coherence of movement for 2D designs in various spatial scales depends basically on the
relative direction of movement of part gratings [13]. Intellectual mapping research has generally
centered on how people explore and procure spatial data about genuine situations. An old
research which presents two examinations that analyze how people learn relative direction
between landmarks in a desktop virtual condition (William S. Albert, Ronald A. Rensink and
2
© 2017 by Daffodil International University
Jack M. Beusmans, 1999). This relative direction test included showing the course of concealed
landmarks from various vantage focuses in nature [14].
In 2017 a research which researchers are Attiya Mahmood, Jon W. Wallace, and Michael A.
Jensen. They reveal estimation that, given any announcement in Unmanned Aerial Vehicles
(UAV), where estimation of relative attitude between two unmanned flying vehicles which in
view of different information and various yield radio recurrence transmissions between the two
flying machine. Specialist demonstrated that three Euler points required depicting the relative
attitude [1]. Yet, computer researchers are regularly concerned about settings where agents are
asset limited, in which case a few questions remain: does there exist a multi-agent system
condition, given any announcement relative direction logic, either creates a short verification or
infers that no short evidence exists? Comparably, it approaches whether the class of issues for
which one can rapidly check a proposed arrangement is the same as the class of issues for which
one can rapidly discover such an answer.
This absence of advance may urge to look for new algorithmic procedures for identify or
learning relative directions for multi-agent based system. As we will find in this thesis,
specification of learning total six relative directions between two the agents and identify relative
directions by using of two directions.
1.2 Research Objectives
The main objective of this thesis is how agents are identify or learn relative directions among
them or each one another by using our propose MARD algorithm where MARD algorithm
performs from multi-agent viewpoints.
This research report exhibits about three different algorithms. This research report exhibits about
three different algorithms. The purpose of first algorithm is learn relative direction between two
agents, purpose of second algorithm is identify a agent and also detect relative direction and
purpose of third algorithm is how one agent learn another agent's relative direction by using of
two direction value. All thought of these propose algorithm approach show with different issues
are additionally talked about in chapter 3 (Proposed Algorithm Model).
1.3 Research Questions
The thesis with titles showing that how relative directions are works between two agents from
computer aspects. There have exactly some research questions and this will enable to understand
some features of this thesis.
3
© 2017 by Daffodil International University
 Why need relative direction and MARD algorithm?
 What direction learning opportunities between two agents in computer science aspects?
 How relative direction works in 3D environment and route direction?
 Is it a new or modify approach?
 Why use numerical value in the approach?
 Why numerical values have been used for indicate relative direction?
1.4 Organization
In Chapter 1, we revolve our work around the thesis and quickly present about foundation
overview and purpose of this algorithmic research.
In Chapter 2, we concentrate on a particular issue for building algorithm. In here we
demonstrate which literature is audited. We at that point exhibit the evidence of our motivation,
which utilizes relative direction of agents, before showing our own confirmation of a similar
outcome and our outcome on endless groups of limit parts. This chapter briefly foundation or
background of thesis, past research of relative direction, artificial intelligence based concept.
In Chapter 3, we introduced about MARD algorithm. In here we have also displayed various 3D
cases of agent and another different approach of MARD algorithm which will helps in machine
learning sector. We also demonstrate multi-agent route directions which present interpretations
mapping over the different area.
In Chapter 4, we talk about our research methodology and exploration technique. This part
shows research design and Algorithm engineering method for this algorithmic research.
In Chapter 5, we analyse our algorithm, result. This part demonstrates explore examination and
execution assessment with time and histogram. Different type of solving strategy analysis we
exhibit in this section.
In Chapter 6, we endeavor to settle on a choice about legitimacy of relative direction algorithm
for multi-agent environment. This part indicates summary of this thesis and finishing.
In Bibliography, we try to show proper references which help for complete this research.
In Appendix, we incorporate audits of the fundamental many-sided quality classes, the portrayal
of code, programming structure and evidences of some minor subtle elements specified in the
body of the thesis.
4
© 2017 by Daffodil International University
1.5 Definitions
MARD algorithm: Multi-agent relative direction (MARD) algorithm refers an algorithm
strategy which identify or learning relative direction among the agents where one agent can be
learning other agent’s relative directions.
UAV: Unmanned aerial vehicle (UAV) normally acquainted as a drone. It is an airplane or
aircraft without any human pilot on board. UAVs are a segment of an unmanned flying machine
which incorporates a UAV, a ground-based controller, and an arrangement of interchanges
between the two. The flight of UAVs may work with different degrees of self-rule: either
according to remote control by a human administrator or independently by locally available
computers.
Multi-agent: A multi-agent framework or system is an automated manner made out of various
associating smart operators inside a domain. Multi-agent system can be utilized to take care of
issues that are troublesome or incomprehensible for an individual operator or a solid framework
to solve. Intelligence may involve some methodic, utilitarian, procedural approach, algorithmic
inquiry. Despite the fact that there is impressive cover, a multi-agent process is not generally the
same as an agent-based model.
Route Directions: A route is regularly part into a few fragments that are then verbalized. These
verbalized directions can be guidelines to make a specific move, for example, "walk" or "turn",
or portrayals of the map.
Direction of Arrival: In signal technology writing, direction of arrival (DOA) means the course
from which as a rule an engendering wave touches base at a point, where as a rule an
arrangement of sensors are found. DOA discovers the direction in relative to the cluster where
the sound source is found.
OPRA: The Oriented Point Relation Algebra (OPRA) distributes for subjective spatial
description and reasoning. OPRA is an introduction calculus math with movable granularity.
OPRA depends on objects which are spoken to as oriented points. Oriented points are indicated
as match of a point and a direction on the 2D-plane.
1.6 Motivation of Research
The ―Direction Learning‖ has been a well known research point among specialists and
researchers of arithmetic and computer science. With the objective of limiting the relative
direction of the agent in estimating agent's relative direction problem. Arrangements of this
issue can be connected to an extensive variety of improvement issues.
5
© 2017 by Daffodil International University
Estimation of relative attitude between two UAV [1] is the latest research of direction learning
calculation which can play numerous info various yield radio recurrence transmissions between
the two aircraft. Most likely this recent research initially indicates multi-operator idea based
relative state of attitude learning calculation. Our analysis is to attempt to give some calculation
which takes after a few exercises of this recent research. Generally this recent research
motivated us to doing our research. The structures we are searching for are just thickness
varieties in the computation. Contrasting landmark based guidelines and relative direction
directions on people on foot in a genuine city condition, measuring both goal and subjective
achievement [4]. We find that at some choice focuses, multi agent based relative direction
algorithm work better for route direction in mapping. We show a strategy that how our research
gives better instruction in different computer science area.
6
© 2017 by Daffodil International University
Chapter 2
Background and Literature Review
The background gives a prologue to Multi-Agent Relative Direction (MARD) algorithm. It
additionally presents some artificial intelligence (AI) perspective concept about this algorithm
which is of intrigue while talking about multi-agent based relative direction learning process.
This section describes the sequencing relative direction in any multi-agent based issue.
Highlights of an algorithm are examined with a specific end goal to order the issue into sub-
issues like object finding problem, structure issue. This Chapter additionally gives a presentation
of the unmanned aerial vehicles and talks about the similarities and contrasts between past
research and this thesis.
2.1 Relative Direction Fundamentals
The most well-known relative directions are right, left, up, down, forward and backward. There
are definite connections between the relative directions. Forward-backward, left-right, and up-
down are three sets of integral relative directions. Relative directions are otherwise called
egocentric coordinates. Relative directions can be helpful to individuals who are new to the area
of cardinal directions. Since meanings of left and right in view of the geometry of the natural
habitat are inconvenient. The importance of relative direction words is passed on through
custom, cultural assimilation, training, and direct reference. One normal meaning of up and
down utilizes gravity and the globe as an edge of reference. Up is then characterized as the other
way of down. Another normal definition utilizes a human body, standing upright, as an edge of
reference.
Forward and backward might be characterized by alluding to a question's or individual's
movement. Forward is characterized as the bearing in which the question is moving. Backward
is then characterized as the other way to forward. Then again, forward might be the direction
pointed by the onlooker's nose, characterizing backward as the heading from the nose to the
sagittal fringe in the eyewitness skull. Concerning a ship forward would show the relative
7
© 2017 by Daffodil International University
position of any protest lying toward the path the ship is pointing. In Figure 1 illustrates different
relative directions from human perspective.
Figure 1: Different types of relative directions.
For symmetrical purpose, it is additionally important to characterize forward and backward
regarding expected course. Many mass travel trains are constructed symmetrically with matched
control corners, and meanings of forward, in reverse, left, and right are brief.
2.2 Multi-Agent System Environment
Multi-agent structure is a computerized way made out of different partner keen administrators
inside a space. Multi-specialist framework can be used to deal with issues that are troublesome
or inconceivable for an individual agent or a strong system to solve. Multi-agent frameworks
comprise of agents and their condition. Regularly multi-agent frameworks examine alludes to
programming operators. Nonetheless, the specialists in a multi-agent system could similarly well
8
© 2017 by Daffodil International University
be robots, people or human groups. A multi-specialist framework may contain consolidated
human-agent groups.
The inspiration for considering multi-agent system regularly originates from enthusiasm for
programming or software agents. The material traverses teaches as assorted as software
engineering (artificial intelligence, hypothesis, and distributed computing), financial matters
(essentially microeconomics concept), research, scientific rationality, and phonetics. In
understanding the determination made here, it is valuable to remember the accompanying
algorithms [27].
Figure 2: Characteristics of Multi-agent System Environment.
Multi-agent environment can include specialists making arrangements for a shared objective, an
agent organizing the plans or arranging of others, or specialists refining their own particular
designs while consulting over errands or assets. The theme likewise includes how agents can do
this progressively while executing designs. Multi-agent booking varies from multi-agent
planning a similar way arranging and vary in planning regularly the undertakings that should be
performed are as of now chose, and by and by, planning tends to concentrate on algorithms for
particular issue areas.
2.3 Previous Research and Work
To represent the idea of multi-agent relative direction (MARD) algorithm previously formally
characterizing every one of the parts that go into the several approaches, we give a genuinely
9
© 2017 by Daffodil International University
casual exploration of one of its more eminent examples of success stories. All things considered,
this subsection might be skipped if the peruser likes to jump straight into definitions.
 Relative Attitude Estimation of UAV (Mahmood, Wallace and Jensen, 2017):
They propose a renewed algorithm for evaluating the relative attitude between two unmanned
aerial vehicles (UAV) in view of various multiple input and output radio recurrence
transmissions between the two airplanes. The strategy can evaluate each of the three Euler
points required to portray the relative disposition [1].
Figure 3: Tracking UAV and Handoff UAV in
Relative Attitude Estimation of UAV’s paper
In here the system show comprising of Tracking and Handoff UAVs and their comparing nearby
arrange outlines and also the connection between the new organize outlines at the two UAVs
framed from direction of entry estimates. Each UAV has its own particular nearby facilitate
outline characterized by the unit-standard vectors where i ∈ {h, t} for Handoff and Tracking,
individually. The analysts of this paper shows a novel calculation that consolidates direction of
arrival (DOA) gauges with polarimetric multi-antenna apparatus channel appraisals to figure the
relative attitude between two UAVs.
 Relative Direction Change (Hahn, Bethge and Döllner, 2017):
A topology-based metric for design solidness in treemaps of the Relative Direction Change
(RDC) introduced metric considers the nearness and course of action of single shapes in a
treemap, and takes into consideration a rotation-invariant portrayal of format alterations between
two snapshots of a dataset delineated with treemaps [2].
10
© 2017 by Daffodil International University
 Fast Contour-Tracing Algorithm Based on a Pixel-Following Method (Seo, Chae, Shim,
Kim, Cheong and Han, 2016):
The experiment introduces a novel contour-tracing algorithm for quick and exact contour
following. This algorithm orders the sort of contour pixel, in light of its neighborhood design. At
that point, it traces the following contour utilizing the past pixel’s model. The algorithm follows
shape pixels along the clockwise direction from the present pixel, i.e., it consecutively seeks
nearby dark pixels of the present pixel utilizing a relative directional request, for example, left,
front-left, front, front-right, right, rear right and back. To decide the contour point, which might
be a contour pixel, the tracer recognizes the power of its adjoining pixel Pr and the new absolute
direction dr for Pr by utilizing relative direction data r ∈ { front, front − left, left, rear − left,
rear, rear − right, right, r ∈ { front − right} [3].
 Relative Directions Work Better Than Landmarks (Götze and Boye, 2015):
People make broad utilization of landmarks while depicting the best approach to others and are
more effective after directions that cover landmarks. It exhibit an examination contrasting
landmark based guidelines and relative direction indication on people on foot in a genuine city
environment, measuring both target and subjective achievement. Researchers find that at some
choice focuses, relative direction guidelines work better. Specifically, guidelines that avoid a
landmark and use just a relative direction like "left" or "right", appear to be favored at some
decision focuses, especially those with a basic arrangement where streets meet at right edges [4].
 Complexity of Reasoning with Relative Directions (Lee, 2014):
In the case of reasoning upon relative directions can be performed in NP has been an open issue
in subjective spatial reasoning. Effective reasoning with relative directions is fundamental, for
instance, rule-compliant agent navigation [5]. In this research reasoning upon relative directions
is ∃R-complete. As a result, reasoning with relative directions is not in NP, if not NP = ∃R,
where ∃R is a many-sided complexity class.
 Effective Reasoning about Relative Directions (Lee, Renz and Wolter, 2013):
Eyewitnesses commonly depict environment by indicating the relative directions in which they
see different articles or other people from their perspective [6]. They demonstrate that reasoning
in StarVars is in NP and present the primary algorithm that enables to viably coordinate relative
direction data from various observers. They built up a spatial portrayal, StarVars, which
increases cardinal direction relations to illustrate to relative directional information.
 Relative Identification and Direction for Wireless Network (Weng and Lai, 2013):
A less intricate, more productive routing algorithm called as relative identification and direction-
based sensor routing (RIDSR) algorithm [7]. RIDSR influences sensor hubs to set up more
11
© 2017 by Daffodil International University
dependable and vitality effective routing path for information transmission. RDSR calculation
not just tackles the routing loop problem inside the algorithm yet in addition encourages the
immediate choice of a shorter way for routing by the sensor node. Moreover, it saves energy and
broadens the existence of the sensor hubs.
 Relative Direction of Oriented Points (Mossakowski and Moratz, 2012):
An imperative problem in qualitative spatial reasoning is the portrayal of relative directions. In
this paper, basic geometric tenets that empower reasoning about the relative direction into
oriented points [8]. This structure arranged a oriented point algebra OPRAm, has a versatile
granularity m. In this paper a basic algorithm for figuring the OPRAm synthesis tables and
demonstrates its rightness.
 Verbal Navigational Directions in Relative Frames (Mossakowski and Moratz, 2008):
This examination inspected how people utilize verbal route directions conferred in relative and
absolute edges of reference in genuine route, especially contrasts or likenesses in cognitional
load postured by the two frames of reference [9]. This instruction took a gander at how people
utilize verbal route directions offered in two sorts of frames of reference, relative and absolute,
in genuine route. Specifically, inspected the distinctions or likenesses in the trouble of utilizing
and preparing data given in favored and non preferred casings of reference, and whether
individuals could adjust to or switch between the two frames of reference.
 Triangular Multiple Flow Direction Algorithm (Seibert and McGlynn, 2007):
Gridded digital elevation data (DEMs) frequently alluded to as DEMs, are a standout amongst
the most broadly accessible types of natural information. Here a give an account of a stream
routing algorithm and contrast it with three regular classes of calculations at present in across
the board utilize. The upside of the algorithm is that unrealistic dispersion on planar or curved
hillslopes is dodged, while numerous flow directions are permitted on raised hillslopes. The
steepest directions point fairly to one left and right. Be that as it may, since there must be one
outflow direction in the algorithm, just a single of these two directions gets territory, while the
two directions ought to get region [10].
 An ant colony optimisation algorithm for the 2D and 3D (Shmygelska and Hoos, 2005):
The protein folding issue is a principal issue in computerized molecular science and biochemical
physics. In this work, research demonstrated that ant colony optimisation (ACO) can be
connected in a somewhat straight-forward path to the 2D and 3D HP Protein Folding Problems.
Despite the fact that our ACO-HPPFP-3 calculation depends on exceptionally straightforward
structure parts (single relative directions) and a basic backup neighborhood seek strategy [11].
12
© 2017 by Daffodil International University
 Relative Direction as a Binary Relation (Moratz, 2006):
A central issue in robotics is the representation of relative orientation. This paper introduces a
new calculus about oriented points which has a scalable granularity [12]. In this calculus, named
OPRA, simple rules can generate the minimal composition. Furthermore, the algebraic closure
for a set of OPRA statements is sufficient to solve knowledge integration tasks in robotics.
 Plaid Motion Coherence on Component Grating Directions (Kim and Wilson, 1993):
A few element motion directions were created little to vast angular contrasts. In here a
confirmation obviously demonstrates that coherence of movement for 2D designs in various
spatial scales depends fundamentally on the relative direction of motion of component gratings
and is moderately autonomous of difference and speed. It is likewise free of the SF distinction
between two parts as long as the proportion is more prominent than around 3:1 [13]. It would
appear to be environmentally more conceivable for the visual system to decide inflexibility of
movement in view of the relative directions of neighborhood motion vectors.
 Relative Directions between Landmarks (Albert, Rensink and Beusmans, 1999):
This examination presents two tests that inspect how people learn relative directions between
landmarks in a desktop virtual condition. Subjects were introduced preview pictures of various
virtual environments containing recognizing points of landmarks and a road network. The
introduction of each virtual environment, subjects were given a relative direction test [14]. The
relative direction test included demonstrating the direction of concealed landmarks from various
vantage focuses in the environment.
 Robot kinematics (1998):
Kinematics is the connections between the positions, speeds, and increasing velocities of the
connections of amanipulator, where a controller is an arm, finger, or leg. This exploration
characterizing the arrangement as far as elbow up or down, left or right handed [15]. A matrix
portrays the change from the base to the hand of the controller, a succession called the forward
kinematic change of the manipulator.
 Maps and Relative Direction (Foster, 2016):
It's quite basic to portray direction in connection to area on a map. Go up that path, down here,
or over yonder. Up, down, and over are relative directions given from a perspective, regularly
physical topographic change [16]. Up stream, down the slope, and over to the lake. The words
up and down can be held in respect to gravity. Unless people are alluding to up and down in
connection to geology, or in relative to a specific area.
 Relative Main Line layout algorithm:
The Relative Main Line format algorithm works from traits that enable the calculation to
distinguish the straight lines that is, the principle lines and root schematic nodes from which
13
© 2017 by Daffodil International University
those straight lines begin. Root schematic nodes can be set utilizing the Set Schematic Root tool
[17]. Set Schematic Root to determine the beginning stages of the straight lines. The algorithm
initially looks nodes to observe candidates to be the root node that is, node associated with a
single link that can be considered as the beginning point for a straight line.
 Dragonfly Algorithm for Solving Multi-objective Problems (Mirjalili, 2016):
Dragonfly calculation is a fiction swarm intelligence streamlining technics. The progression
vector of this algorithm demonstrates the direction of the development of the dragonflies and
characterized as takes after: Xr + 1 = ( sSi + aAi + cCi + fFi + eEi ) + wXt where s
demonstrates the division weight, Si shows the partition of the i-th individual, a is the
arrangement weight, Ai is the arrangement of i-th singular, c shows the attachment weight, Ci is
the union of the i-th singular, f is the sustenance factor, Fi is the nourishment wellspring of the i-
th singular, e is the foe factor, Ei is the position of foe of the i-th singular, w is the dormancy
weight, and t is the cycle counter [18].
 DOA Estimation of Animal Vocalizations (Hedley, Huang and Yao, 2017):
A recording system built from two Wildlife Acoustics SM3 recording units that can calculate
the direction-of-arrival (DOA) of an approaching signal with high precision [20]. Signal
processing algorithms, similar to the MUSIC algorithm utilized their analysis, they utilize these
stage contrasts to decide the angle from which each sound arrived (α and β for the red and blue
winged creatures, individually, relative to a self- assertive mention angle, marked 0). The system
utilizes four all the while recording receivers to evaluate the direction from which a sound
arrived, in view of the stage contrasts of the approaching sound waves at the microphones.
 Discrete-State-Based Vision Navigation Control Algorithm (Wei, 2015):
To set out a principled dialog of the exactness and productivity of navigation algorithms,
entirely quantitative meanings of following error, actuator impact, and time proficiency are built
up. The navigation algorithm would control the robot following the particular direction [21]. In
the wake of characterizing the relative angle between desired velocity vector and the real
velocity vector as course blunder signified by e. Most extreme Steering Angle max. This value
is the practical steering angle pushing forward in a direction. That implies the robot can move
forward with the direction scope of [−max, max]. Negative esteem implies the robot turns right
while the positive value implies it turns left with respect to the robot.
2.4 Left-Right Confusion
Left-right confusion is the inconvenience a few people have in recognizing the distinction
between the headings left and right. These individuals can as a rule typically perform every day
14
© 2017 by Daffodil International University
exercises, for example, driving as indicated by signs and exploring as indicated by a map,
however will regularly go astray when advised to turn left or right and may experience issues
performing activities that require exact comprehension of directional orders.
2.4.1 Research about Left-Right Confusion
Challenges in left–right discrimination (LRD) are usually experienced in regular day to day
existence circumstances. An examination demonstrates that the neurocognitive components of
left– right separation and the particular part of left precise gyrus [29]. In later an examination
surveyed the connection between self-appraised right–left confusability and execution on the
Money Road-Map Test (MRMT). Eighty-six understudies appraised right–left subjective
confusability utilizing a poll, and afterward finished the Money Road-Map Test.
Another examination researches the connection between the view of bilateral symmetry and left-
right bewilderment in neurologically in place grown-ups by utilizing tachistoscopic introduction
of boosts and a decision response time technique. Scientists found a little yet predictable pattern
toward snappier symmetry judgments in left-right distracted subjects. The legitimacy of such
self-report measures in foreseeing real execution on right-left segregation undertakings is
addressed since the outcomes, in any event as a component of handedness, relied upon the
inquiry [30, 31, 32].
2.4.2 Artificial Intelligence Perspective
Artificial intelligence (AI) is insight displayed by machines, instead of people or different
creatures (natural intelligence). In software engineering, the field of AI investigate characterizes
itself as the investigation of intelligent agents are any gadget that sees its condition and takes
activities that expand its risk of achievement at some objective. Conversationally, the expression
computerized intelligence is connected when a machine impersonates subjective capacities that
humans connect with other human personalities, for example, "learning" and "critical thinking".
Learning or identify direction is an exploration range in software engineering and computer
science, with territories, for example, unraveling, trouble and era. One of the critical choice rule
for picking MARD algorithm for this proposal have hence been the algorithm hidden strategy
for crossing the hunt space, for this situation deterministic and stochastic strategies.
Possible left-right confusion with artificial intelligence perspective show in Figure 4, 5, 6 and 7,
each agent has its own local relative directions. Where Agent 1 = R, Agent 2 = A and Mobile
phone is a object for find mobile phone object. Primary objective of these figures are to find the
object and defining the left and right direction where relative directions (right and left) of two
agents show from different perspective. The process for achieving this purpose is as follows:
15
© 2017 by Daffodil International University
Figure 4: Relative directions (right and left) of two agents
from different perspective. In here a mobile phone is a
example of object.
1) In Figure 4, there are two agent respectivelty agent A and agent R and their right and left
directions different from different perspective. Mobile phone is an example of object for
understanding left-right confusion in human and artificial intelligence both perpective.
The Mobile object is located on its left side of agent R and on the right side of agent A.
2) In Figure 5, agent A ordered agent R to find the mobile object on the right direction.
When agent R searches the mobile phone object in his right direction, he could not find
any mobile phone, because relative directions between two agents are different. It means
that agent A’s right direction is not equal to agent R’s right direction. Relative directions
between two agents are different.
3) In Figure 6, agent A think and calculated that his right direction is equal to agent R’s left
direction. Agent A also calculated that his left direction is equal to agent R’s right
direction. At last agent A learn that Agent R’s relative directions.
4) Lastly in Figure 7, agent A ordered agent R to find a mobile phone on agent R’s left
direction and agent R find the mobile phone object.
16
© 2017 by Daffodil International University
Figure 5: Agent A ordered agent R to find the mobile object on the
right direction but agent R could not find any mobile phone on his right
direction, because relative directions between two agents are different.
Figure 6: Agent A calculated that his right direction is equal to agent
R’s left direction and left direction is equal to agent R’s right direction.
17
© 2017 by Daffodil International University
Figure 7: Agent A ordered agent R to find a mobile phone on
agent R’s left direction. At last agent R find the mobile object.
Basically our MARD algorithm will work like this way. As a result one agent easily can
understand anodher agen’s left-right or relative directions. Through this concept we can devlope
our new multi-agent relative direction (MARD) algorithm and with the help of this algorithm we
can solve left-right confusion problem in compuer or artificial intelligence perspective.
18
© 2017 by Daffodil International University
Chapter 3
Proposed Algorithm Model
To represent the idea of multi-agent relative direction (MARD) algorithm previously formally
characterizing every one of the parts that go into the approach, we give a genuinely casual
exploration of one of its more eminent examples of success stories. All things considered, this
subsection might be skipped if the peruser likes to jump straight into definitions.
In this chapter, we present the fundamental theory of MARD algorithm. We work through an
illustrative case in this segment before formalizing the model. At that point we talk about the
characterizing numerical properties of MARD algorithm and give a unique re-detailing of these
properties that will demonstrate basic to demonstrating our primary hypothesis. At long last, we
give a programming structure of MARD algorithm as far as spinor assortments and drawing
ideas.
3.1 Induction
This exploration section introduces propose algorithm for figuring the relative directions
between two agents, where how one agent can take in another agent's relative directions through
in computer or programming perspective. We display an examination of relative direction. In
this section demonstrate our experiment design, realistic input, mathematical concept and
structure of MARD algorithm. We present a novel subjective depiction and portrayal this
MARD algorithm can tackle these issues for dealing with and perceiving relative directions, its
executed work strategy or computation puts aside greater open door for result. In order to
MARD algorithm experiment, we focus on Algorithm Engineering technic. Algorithm
engineering revolves around the outline, examination, analysis, implementation, optimization
and exploratory appraisal of algorithms. One significant though all too every now and again
overlooked issue when driving analyses in Computer Science is to ensure MARD algorithm.
19
© 2017 by Daffodil International University
3.2 Tracking Agent and Handoff Agent
With past work focusing on attitude estimation, an algorithm using Tracking UAV and Handoff
UAV for estimating relative attitude between two unmanned aerial vehicles (UAV) [1]. In the
MARD algorithm we use ―Tracking Agent‖ and ―Handoff Agent‖ (see Figure 8). We consider in
the algorithm where one tracking agent is following an objective on the ground, and it is wanted
to have a moment handoff agent have the capacity to learning the relative direction of the
tracking agent.
3.2.1 Direction Points
We use total six riletive directions in our algorithm. Each agent has 6 direction points for handoff
agent direction points are a, b, c, d, e, f and for tracking agent direction points are same level but
reverse for that direction points are a1, b1, c1, d1, e1, f1. Firstly handoff agent knows his relative
directions but in the beginning handoff agent not knows about tracking agent’s relative
directions, which has been displayed in Figure 8.
Figure 8: Direction points of Handoff Agent and Tracking Agent.
Handoff Agent
Tracking Agent
RightLeft
Up
Down
Forward
Backward
c1
b1 a1
d1
e1
f1
ab
c
d
e
f
20
© 2017 by Daffodil International University
3.2.2 Various 3D Aspects
3D PC illustrations are regularly alluded to as 3D models. The view of relative directions seems
from different 3D aspects. In Figure 9 we show 24 different cases of relative direction aspects of
tracking agent perspective.
Figure 9: Different cases of tracking agent in 3D aspect.
Right
a1
Left
Up
Down
Forward
Backward
b1
c1
d1
e1
f1
Case 1
Case 6
Case 11
Case 16
Case 21
Up
Down
LeftRight
ForwardBackward
Case 2
Down
Up
Left Right
Backward
Case 3
Forward Up
Down
Backward
RightLeft
Forward
Case 4
Up
Down
Forward
Backward
Left
Right
Case 5
Up
Down
Forward
Backward
Left
Right
Down
Up
BackwardLeft
Right
Case 7
Backward
Forward
Down
Up
Left
Right
Case 8
Forward
Backward
Up
Down
Right
Left
Case 9
Backward
Forward
Up
Down
Left
Right
Case 10
Up
Down
Forward
Backward
Left
Right
Backward
Forward
Left
Right
Up
Down
Forward
Case 12
Backward
Forward
Right
Down
Up
Left
Right
Forward
Backward
Case 13
Down
Up
Backward
Forward
Left
Right
Up
Down
Right
Left
Case 15Case 14
Up
Down
Forward
Backward
Left
Left
Forward
Backward
LeftRight
Forward
Backward
Forward
Backward
Down
Up
Left
Right
Up
Down
Backward
Forward
Left
Right
Up Up
Down DownRight
Case 17 Case 18 Case 19
Case 24Case 22 Case 23
Case 20
Up
DownLeft
Right
Forward
Backward
Up
DownRight
Left
Forward
Backward
Backward
Forward
Up
Down
Right
Left
Up
Down
Forward
Backward
Left
Right
c1
c1
c1
c1
c1
c1 c1 c1 c1 c1
c1
c1
c1 c1 c1
f1
c1 c1 c1 c1 c1
c1
c1
c1
b1 b1 b1
b1
b1 b1
b1 b1
b1 b1 b1
b1 b1 b1 b1 b1
b1 b1
b1 b1
a1
a1 a1 a1
a1 a1 a1 a1 a1
a1 a1 a1 a1 a1
a1 a1 a1 a1 a1
a1 a1
a1 a1
e1
e1 e1
e1 e1 e1e1 e1
e1
e1 e1 e1 e1
e1 e1 e1 e1 e1
e1
e1
e1 e1
f1f1 f1
f1 f1 f1 f1 f1
b1 b1 b1
f1 f1 f1 f1 f1
f1 f1 f1 f1 f1
f1
f1 f1 f1
d1 d1 d1
f1
d1
d1 d1 d1
d1
d1 d1 d1 d1
d1
d1
d1
d1
d1
d1
d1d1
d1d1
21
© 2017 by Daffodil International University
From Figure 8 and Figure 9 we can say that relative direction of tracking agent always
changeable but direction points of tracking agent are constant. Here we find total twenty four
relative direction movement. So we can say that,
Total relative direction movement (M) = Number of relative direction (n) * 4
Or, M = 6*4 = 24
3.2.3 Realistic Inputs
At first handoff agent knows his relative directions but handoff agent not knows about tracking
agent’s relative directions. Moreover handoff agent compare with tracking agent’s relative
directions by tracking agent’s directions points.
For MARD algorithm development we use some niumerical value for identify relative
directions. Every relative direction point has own value which depends on the relative
directions such as 0 for Right, 1 for Left, 2 for Up, 3 for Down, 4 for Forward and 5 for
Backward, which have been shown in Table 1. We have used these values because the reason in
our proposed algorithm we have used two array agent which contains these direction variables.
Relative Direction Direction Point Value
(For relative direction)
Right 0
Left 1
Up 2
Down 3
Forward 4
Backward 5
Table 1: Relative directions and their numerical values.
22
© 2017 by Daffodil International University
3.3 Mathematical Exploration
In this area, we will numerically define the MARD algorithm with every one of the imperatives
specified. The main navigating auto in the computer science serves an arrangement of benefit
arrange from various wellsprings of the algorithm design. The procedure for mathematical
computation of this algorithm is as follow:
Ti  ( Hj )
In here T refers tracking agent, where T = {Right, Left, Up, Down, Forward, Backward} and H
refers handoff agent, where H = {right, left, up, down, forward, backward}. Besides i is the set of
index T’s relative direction, where i = 0, 1, 2, 3, 4, 5 and j is the set of index H’s relative
direction, where j = 0, 1, 2, 3, 4, 5. For i, handoff agent H’s j relative direction is assigned into
tracking-human T’s i relative direction.
3.4 Structure of Algorithm
This investigation part presents propose algorithm for figuring the relative directions between
two agents, where how one agent's can learn in another agent's relative directions through in
programming point of view. In this segment exhibit our analysis plan, practical information,
programming idea and structure of MARD algorithm for application. We very briefly develop
MARD algorithm. The process of MARD algorithm for programming or application perspective
goal is as per the following:
1) Step1: Creat a HandoffAgent class
 An array Direction = [right, left, up, down, forward, backward].
 Direction point variables of handoff agent are a, b, c, d, e and f.
 Points are containing direction values (see Figure 8 and Table 1). So that a = 0, b = 1,
c = 2, d = 3, e = 4 and f = 5.
 Direction point variables are involving relative directions. So that, right = a, left = b,
c = up, d = down, e = forward and f = backward (see Figure 8 and Table 1).
2) Step2: Creat a TrackingAgent class
 Direction = [Right, Left, Up, Down, Forward, Backward].
 Direction point variables of tracking agent are a1, b1, c1, d1, e1 and f1. Direction points
contain different direction value. These values are 0, 1, 2, 3, 4 or 5 (see Figure 8, Figure 9
and Table 1).
23
© 2017 by Daffodil International University
3) Step3: Creat a Main class, main function and object
 handoffAgent is a object of HandoffAgent class
 trackingAgent is a object of HandoffAgent class
4) Step4: Creat a Loop
 A variable i = 0 to 5.
 The condition of the loop is i <= 5.
 After every loop value of i will increase (i++).
 If complete all possible loops then go to step 6 for end process.
5) Step5: Declare j and find relative directions
 if i == a1 then,
j = 0; // j = 0 means right //
 else if i == b1 then,
j = 1; // j = 1 means left //
 else if i == c1 then,
j = 2; // j = 2 means up //
 else if i == d1 then,
j = 3; // j = 3 means down //
 else if i == e1 then,
j = 4; // j = 4 means forward //
 else if i == f1 then,
j = 5; // j = 5 means backward //
 trackingAgent.Direction[ i ] = handoffAgent[ j ];
 Print ―Tracking Agent’s‖ + i direction name == ―Hanoff Agent’s‖ + j direction name;
 Go to step 4 for complete loop.
6) Step6: End
 End execution of algorithm.
We present a pseudocode which in view of objects oriented programming (OOP) formation that
recognizes relative directions and correspondences between two agents. The pseudocode of the
MARD algorithm is given below:
24
© 2017 by Daffodil International University
Algorithm 1: Multi-Agent Relative Direction Algorithm
1: HandoffAgent class {
2: Direction = [ right, left, up, down, forward, backward ];
3: directions points: a; b; c; d; e; f;
4: //points are contains values (see Figure 8 and Table 1)//
5: right = a; left = b; up = c; down = d; forward = e; backward = f1;
6: }
7:
8: TrackingAgent class {
9: Direction = [ Right, Left, Up, Down, Forward, Backward ];
10: directions points: a1; b1; c1; d1; e1; f1;
11: //points are contains values (see Figure 8, Figure 9 and Table 1)//
12: }
13:
14: Main class {
15: void function main( ) {
16: HandoffAgent handoffAgent = new HandoffAgent();
17: TrackingAgent trackingAgen = new TrackingAgent();
18:
19: int j;
20:
21: for i = 0 to 5
22: if i == a1 then, j = 0;
23: else if i == b1 then, j = 1;
24: else if i == c1 then, j = 2;
25: else if i == d1 then, j = 3;
26: else if i == e1 then, j = 4;
27: else if i == f1 then, j = 5;
28:
29: trackingAgent.Direction[ i ] = handoffAgent.Direction[ j ];
30:
31: end for
32: }
33: }
25
© 2017 by Daffodil International University
3.5 Identify Relative Direction
Identify tracking agent and its relative directions are very important aspect. Figure 10 represents
a simple concept of tracking agent's relative directions identification.
Figure 10: Tracking agent's direction identification. (a) Color location
tracking agent’s direction point. (b) Shape are indicates different
relative directions. (c) Handoff agent takes an image for identifying
tracking agent’s relative directions.
We have shown a sample pseudocode of this idea for identify the tracking agent which has been
shown in Figure 10. At first handoff agent’s camera takes an image of tracking agent. There are
six color locations to find out tracking-human’s direction points. An array Shape contains
different shapes which are Circle, Diamond, Triangle, Hexagon, Rectangle and Pentagon. These
shapes refers different relative directions for track a tracking agent’s relative directions. Variable
k for completing the loop and also has been used for identifying the relative directions among
the directional points (a1, b1, c1, d1, e1 and f1) of tracking agent. If specific color is equal to a
specific Shape, then direction points of tracking agent will contain a relative direction value
which is k (see table 1).
The structure of identify tracking agent’s relative direction in programming perspective goal is
as per the following:
Handoff Agent
Camera
Image
Red
Yellow
Green
Orange
Blue
Violet
d1
c1 e1
f1
a1b1
(a)
Right
Left
Up
Down
Forward
Backward
Tracking Agent
(b)
(c)
26
© 2017 by Daffodil International University
Algorithm 2: Direction Identification Algorithm
1: Camera Image of trackingAgent;
2: directions points: a1; b1; c1; d1; e1; f1;
3: location of direction points in image: Green; Blue; Red; Yellow; Orange; Violet;
4: Shape = [Circle, Triangle, Rectangle, Diamond, Pentagon, Hexagon];
5: for k := 0 to 5
6: if Green == Shape[k] then, a1 = k ;
7: else if Blue == Shape[k] then, b1 = k ;
8: else if Red == Shape[k] then, c1 = k ;
9: else if Yellow == Shape[k] then, d1 = k ;
10: else if Orange == Shape[k] then, e1 = k ;
11: else if Violet == Shape[k] then, f1 = k ;
12: end for
3.6 Machine Learning Aspect
Machine learning is an area of computer science and software engineering that gives PCs the
capacity to learn without being expressly programmed. Machine learning explores the study and
construction of algorithms that can learn from and make predictions on data, such algorithms
overcome following strictly static program instructions by making data-driven predictions or
decisions, through building a model from sample inputs. The pseudocodes of Algorithm 2 can
be used for various purposes of machine learning for identify multi-agent based relative
directions. In the next part we represent an alternative algorithm approach of multi-agent based
relative direction identification algorithm, where realistic inputs can be worked better at
machine learning in future.
3.7 Alternative MARD Algorithm Approach
In this section, we give another alternative algorithm approach of MARD algorithm. Firstly, if
we look deeply in Figure 9 we can understand that if we find ―Forward‖ and ―Up‖ direction then
we can easily estimate Left, Right, Down and Backward directions. In Figure 9 there have total
24 case befor go to the alternative MARD algorithm we want to show the condition of these
cases. The condition for 3D cases is summarized in Table 2. In here i = 3, 5 indicates that the
value of Up and Forward are depends on i.
27
© 2017 by Daffodil International University
Case Tracking
agent
direction
Equal
to
Hanoff
Agent
direction
And Tracking
agent
direction
Equal
to
Hanoff
Agent
direction
1 Up == up && Forward == forward
2 Up == up && Forward == backward
3 Up == down && Forward == forward
4 Up == down && Forward == backward
5 Up == left && Forward == up
6 Up == left && Forward == down
7 Up == right && Forward == up
8 Up == right && Forward == down
9 Up == forward && Forward == up
10 Up == forward && Forward == down
11 Up == backward && Forward == up
12 Up == backward && Forward == down
13 Up == right && Forward == forward
14 Up == right && Forward == backward
15 Up == left && Forward == forward
16 Up == left && Forward == backward
17 Up == down && Forward == left
18 Up == down && Forward == right
19 Up == up && Forward == left
20 Up == up && Forward == right
21 Up == backward && Forward == right
22 Up == backward && Forward == left
23 Up == forward && Forward == right
24 Up == forward && Forward == left
Table 2: Condition of Cases.
There are total 12 orders, two particular orders for one individual case. It means that, Total order
= total number of direction 6*2 = 12. These orders show the left and right direction of these
cases. Table 3 shows the orders.
28
© 2017 by Daffodil International University
Order Tracking agent
direction
Assignment
(=)
Handoff agent
direction
1 Right = Right
2 Left = Left
3 Right = Left
4 Left = Right
5 Right = Backward
6 Left = Forward
7 Right = Forward
8 Left = Backward
9 Right = Down
10 Left = Up
11 Right = Up
12 Left = Down
Table 3: Order of Cases.
Now, we have proposed an alternative MARD algorithm approach that is based on only two
directions (Up and Forward). The algorithm has twenty four cases for learning all possible
directions of tracking agent which based on three dimension space. The process of alternative
MARD algorithm approach for programming or application perspective goal is as per the
following:
1) Step1: Creat a HandoffAgent class
 Direction = [ right, left, up, down, forward, backward ].
 Direction point variables of handoff agent are a, b, c, d, e and f.
 Points are containing direction values (see Figure 8 and Table 1). So that a = 0, b = 1,
c = 2, d = 3, e = 4 and f = 5.
 Direction point variables are involving relative directions. So that, right = a, left = b,
c = up, d = down, e = forward and f = backward (see Figure 8 and Table 1).
2) Step2: Creat a TrackingAgent class
 Direction = [Right, Left, Up, Down, Forward, Backward].
 Direction point variables of tracking agent are a1, b1, c1, d1, e1 and f1. Direction points
contain different direction value. These values are 0, 1, 2, 3, 4 or 5 (see Figure 8, Figure 9
and Table 1).
29
© 2017 by Daffodil International University
3) Step2: Creat a Main class, main function and object
 handoffAgent is a object of HandoffAgent class
 trackingAgent is a object of HandoffAgent class
4) Step4: Creat a Loop
 A variable i = 2 and 4.
 Loop condition is i < 5.
 The limitation of loop is i = i + 2.
 If complete all possible loops thent go to step 7.
5) Step5: Declare j and Find Up and Forward Direction
 if i == a1 then, j = 0; // j = 0 means right //
 else if i == b1 then, j = 1; // j = 1 means left //
 else if i == c1 then, j = 2; // j = 2 means up //
 else if i == d1 then, j = 3; // j = 3 means down //
 else if i == e1 then, j = 4; // j = 4 means forward //
 else if i == f1 then, j = 5; // j = 5 means backward //
 trackingAgent.Direction[ i ] = handoffAgent[ j ]; // for learn Up and Forward direction //
 trackingAgent[ i+1] = handoffAgent[ j +1 ]; // for learn Down and Backward direction //
 Print ―Tracking Agent’s‖ + i direction name == ―Hanoff Agent’s‖ + j direction name;
 Go to step 3 for complete all loops.
6) Step6: Find Right and Left Direction
 if Case 1 then, Order 1 and Order 2.
 else if Case 2 then, Order 3 and Order 4.
 else if Case 3 then, Order 3 and Order 4.
 else if Case 4 then, Order 1 and Order 2.
 else if Case 5 then, Order 5 and Order 6.
 else if Case 6 then, Order 7 and Order 8.
 else if Case 7 then, Order 7 and Order 8.
 else if Case 8 then, Order 5 and Order 6.
 else if Case 9 then, Order 3 and Order 4.
 else if Case 10 then, Order 1 and Order 2.
 else if Case 11 then, Order 1 and Order 2.
 else if Case 12 then, Order 3 and Order 4.
 else if Case 13 then, Order 9 and Order 10.
 else if Case 14 then, Order 11 and Order 12.
 else if Case 15 then, Order 11 and Order 12.
30
© 2017 by Daffodil International University
 else if Case 16 then, Order 9 and Order 10.
 else if Case 17 then, Order 5 and Order 6.
 else if Case 18 then, Order 7and Order 8.
 else if Case 19 then, Order 7 and Order 8.
 else if Case 20 then, Order 5 and Order 6.
 else if Case 21 then, Order 9 and Order 10.
 else if Case 22 then, Order 11 and Order 12.
 else if Case 23 then, Order 11 and Order 12.
 else if Case 24 then, Order 9 and Order 10.
7) Step7: End Process
 After complete all steps then end process.
The learning approach of the relative direction for three dimension modeling purpose is look
complicated but is concept is easy to identify relative direction when handoff agent can learn
tracking agent's all relative directions by using Up and Forward direction of tracking agent. The
process of Alternative MARD algorithm for programming or application perspective goal is as
per the following:
Algorithm 3: Alternative MARD Algorithm
1: HandoffAgent class {
2: Direction = [ right, left, up, down, forward, backward ];
3: directions points: a; b; c; d; e; f;
4: //points are contains values (see Figure 8 and Table 1)//
5: right = a; left = b; up = c; down = d; forward = e; backward = f1;
6: }
7: TrackingAgent class {
8: Direction = [ Right, Left, Up, Down, Forward, Backward ];
9: directions points: a1; b1; c1; d1; e1; f1;
10: //points are contains values (see Figure 8, Figure 9 and Table 1)//
11: }
12: Main class {
13: void function main( ) {
14: HandoffAgent handoffAgent = new HandoffAgent();
15: TrackingAgent trackingAgen = new TrackingAgent();
16: int j;
17: for i = 2 and 4
18: if i == a1 then, j = 0;
19: else if i == b1 then, j = 1;
20: else if i == c1 then, j = 2;
31
© 2017 by Daffodil International University
21: else if i == d1 then, j = 3;
22: else if i == e1 then, j = 4;
23: else if i == f1 then, j = 5;
24: trackingAgent[ i ] = handoffAgent[ j ];
25: // for learn Up and Forward direction //
26: trackingAgent[ i+1] = handoffAgent[ j +1 ];
27: // for learn Down and Backward direction //
28: end for
29:
30: if Case 1 then, Order 1 and Order 2.
31: else if Case 2 then, Order 3 and Order 4.
32: else if Case 3 then, Order 3 and Order 4.
33: else if Case 4 then, Order 1 and Order 2.
34: else if Case 5 then, Order 5 and Order 6.
35: else if Case 6 then, Order 7 and Order 8.
36: else if Case 7 then, Order 7 and Order 8.
37: else if Case 8 then, Order 5 and Order 6.
38: else if Case 9 then, Order 3 and Order 4.
39: else if Case 10 then, Order 1 and Order 2.
40: else if Case 11 then, Order 1 and Order 2.
41: else if Case 12 then, Order 3 and Order 4.
42: else if Case 13 then, Order 9 and Order 10.
43: else if Case 14 then, Order 11 and Order 12.
44: else if Case 15 then, Order 11 and Order 12.
45: else if Case 16 then, Order 9 and Order 10.
46: else if Case 17 then, Order 5 and Order 6.
47: else if Case 18 then, Order 7and Order 8.
48: else if Case 19 then, Order 7 and Order 8.
49: else if Case 20 then, Order 5 and Order 6.
50: else if Case 21 then, Order 9 and Order 10.
51: else if Case 22 then, Order 11 and Order 12.
52: else if Case 23 then, Order 11 and Order 12.
53: else if Case 24 then, Order 9 and Order 10.
3.8 Multi-Agent Route Direction
A route direction is ordinarily part into a few sections that are then verbalized. These verbalized
bearings can be guidelines to make a specific move, for example, "move" or "turn", or portrayals
of the earth like "There is a market to one left side". The request of the headings ought to mirror
32
© 2017 by Daffodil International University
the straight request in which the course is crossed. An investigation contrasting landmark point
based guidelines and relative directions on people on foot in a genuine city condition, measuring
both goal and subjective achievement and find that at some choice focuses, relative direction
indication work better [4].
In Figure 11 a mapping image illustrate that there have two agents who are stand in different
area. Handoff agent stayed in area 1 and tracking agent stayed in area 2. Tracking want to find
―School‖. Handoff agent knows about the school location. So hanoff agent helps tracking agent
to find school location by relative directions. MARD algorithm can solve this type of issue.
That MARD algorithm can assume an uncommon part in the correspondence of route direction
has been exhibited in a few investigations. Multi-agent based computation is a way to recognize
critical focuses along the course where turning moves should be made or could be taken and in
addition to find the start and the finish of the route. MARD algorithm additionally has a part in
the clear impact of route directions, and to affirm that the devotee has accurately executed a turn.
A image for multi-agent based route directions where use relative directions is shown in Figure
11.
Figure 11: Multi-agent based route direction in mapping.
Area 1 Area 2
School Up
Down
Foward Backward
Right
Left
RightLeft
Up
Down
Backward
Foward
Tracking Agent
Handoff Agent
33
© 2017 by Daffodil International University
Chapter 4
Methodology
This part concentrates on how process of choosing algorithm engineering method. This part
gives an outline of the measurable investigation which was performed on this algorithmic
research. This additionally incorporates what computational restrictions were available and how
this affected the outcomes. Methodology is the efficient, theoretical examination of the
strategies connected to a field of study. It involves the theoretical investigation of the methods
and standards related with a branch of information.
4.1 Algorithm Engineering Method
Algorithm engineering (AE) is a usual methodology for algorithmic research. It centers on the
analysis, design, examination, execution, enhancement, implementation and test assessment of
algorithm. In our paper we used Algorithm engineering methodology [22, 24, 25, 26].
First, we introduce the relative direction concept on multi-agent best system structure. At that
point we designed the MARD algorithm. The primary part of this algorithmic research is set
realistic models and set real inputes which helps test the algorithm on various relative direction
cases. We use some numerical values for identify relative directions, which helps to demonstrate
a realistic model for develop the MARD algorithm.
In this manner, we designed the algorithm. Before design MARD algorithm we create "Tracking
agent" and "Handoff agent", after that we are keen on an effective algorithm. First algorithm
design depends on proper inputs, mathematical exploration and human perspective concept.
Second algorithm structure construct on image processing aspects. At last alternative algorithm
approach has been designed on machine learning aspects and two direction based identification.
34
© 2017 by Daffodil International University
This algorithm was analyzed by pseudocode and theoretical knowledge of mathematics. It
concerns the result of the algorithm configuration stage in the picked programming dialect.
Experiment of algorithm by programming in application level which shows different results and
various preferences. All execution processes were measured and promote analysis. Since there
may be varieties in execution and a component of randomness in algorithm implementation,
various tests were performed on each relative direction.
Figure 12: Algorithm Engineering methodology cycle.
An essential objective of AE in our research is additionally to accelerate the exchange of
algorithmic information into applications. In this experiment period of algorithmic research
picking the correct issue cases for testing is continually testing until get correct output.
Applications
Real
Inputs
Realistic
Model
Design
ExperimentsAnalysis
Implementation
35
© 2017 by Daffodil International University
4.2 Experiment Control
Algorithm Engineering is constantly determined by certifiable applications. The application
situation decides the equipment which must be demonstrated generally sensibly. The
consequences of an experimentation stage may then later on request an update of the
demonstrating stage, in light of the fact that the picked models are not appropriate.
Once in a while an examination of the picked model would already be able to uncover its
deficiency. Apart from delimiting the negative effects of round-off by controlling the
accumulation of numerical errors, numerical analysis also helps us in assessing their actual
magnitude at runtime. In doing so it allows us to check at runtime whether computed solutions
are reliable or not. This is an important part of the basis of reliable computing.
36
© 2017 by Daffodil International University
Chapter 5
Results and Analysis
In this chapter various outcomes are given together an exchange about how the outcomes could
be translated. This section is dedicated to exhibiting how algorithm performs. This segment
demonstrates how algorithm performs in respect to the each other and talks about various part of
correlation. In here we investigate the possibility of trouble rating and the idea of relative
directions being inalienably troublesome. One of the principle tasks of this part is to characterize
test cases for MARD algorithm to get setups of algorithm when apply to sequencing relative
direction learning issue.
5.1 Accuracy Result
The accuracy of MARD algorithm involves determining how accurately learning the relative
directions, and we measure it by counting the number of relative directions and different
directional movement case. First, we apply the algorithm to the test three dimensional movement
cases and mark the output on the directions. Then, we count all of direction cases. Table 5 shows
the results of the relative directions of the proposed algorithm. We find total 24 cases for relative
direction movement which is shown in Figure 9 and in Table 4 we have displayed these different
results of different tracking agent 3D perspective.
Case
No.
Tracking Agent Handoff Agent
Direction
Point
Direction Point
value
Relative
Direction
Directio
n Point
Relative
Direction
Direction Point
Value
1
a1 0 Right a right 0
b1 1 Left b left 1
c1 2 Up c up 2
d1 3 Down d down 3
e1 4 Forward e forward 4
f1 5 Backward f backward 5
37
© 2017 by Daffodil International University
2
a1 1 Left a right 0
b1 0 Right b left 1
c1 2 Up c up 2
d1 3 Down d down 3
e1 5 Backward e forward 4
f1 4 Forward f backward 5
3
a1 1 Left a right 0
b1 0 Right b left 1
c1 3 Down c up 2
d1 2 Up d down 3
e1 4 Forward e forward 4
f1 5 Backward f backward 5
4
a1 0 Right a right 0
b1 1 Left b left 1
c1 3 Down c up 2
d1 2 Up d down 3
e1 5 Backward e forward 4
f1 4 Forward f backward 5
5
a1 3 Down a right 0
b1 2 Up b left 1
c1 4 Forward c up 2
d1 5 Backward d down 3
e1 1 Left e forward 4
f1 0 Right f backward 5
6
a1 3 Down a right 0
b1 2 Up b left 1
c1 5 Backward c up 2
d1 4 Forward d down 3
e1 0 Right e forward 4
f1 1 Left f backward 5
7
a1 2 Up a right 0
b1 3 Down b left 1
c1 4 Forward c up 2
d1 5 Backward d down 3
e1 0 Right e forward 4
f1 1 Left f backward 5
8
a1 2 Up a right 0
b1 3 Down b left 1
c1 5 Backward c up 2
d1 4 Forward d down 3
e1 1 Left e forward 4
f1 0 Right f backward 5
9
a1 1 Left a right 0
b1 0 Right b left 1
c1 4 Forward c up 2
d1 5 Backward d down 3
e1 2 Up e forward 4
f1 3 Down f backward 5
38
© 2017 by Daffodil International University
10
a1 0 Right a right 0
b1 1 Left b left 1
c1 5 Backward c up 2
d1 4 Forward d down 3
e1 2 Up e forward 4
f1 3 Down f backward 5
11
a1 0 Right a right 0
b1 1 Left b left 1
c1 4 Forward c up 2
d1 5 Backward d down 3
e1 3 Down e forward 4
f1 2 Up f backward 5
12
a1 1 Left a right 0
b1 0 Right b left 1
c1 5 Backward c up 2
d1 4 Forward d down 3
e1 3 Down e forward 4
f1 2 Up f backward 5
13
a1 2 Up a right 0
b1 3 Down b left 1
c1 1 Left c up 2
d1 0 Right d down 3
e1 4 Forward e forward 4
f1 5 Backward f backward 5
14
a1 2 Up a right 0
b1 3 Down b left 1
c1 0 Right c up 2
d1 1 Left d down 3
e1 5 Backward e forward 4
f1 4 Forward f backward 5
15
a1 3 Down a right 0
b1 2 Up b left 1
c1 0 Right c up 2
d1 1 Left d down 3
e1 4 Forward e forward 4
f1 5 Backward f backward 5
16
a1 3 Down a right 0
b1 2 Up b left 1
c1 1 Left c up 2
d1 0 Right d down 3
e1 5 Backward e forward 4
f1 4 Forward f backward 5
17
a1 5 Backward a right 0
b1 4 Forward b left 1
c1 3 Down c up 2
d1 2 Up d down 3
e1 1 Left e forward 4
f1 0 Right f backward 5
39
© 2017 by Daffodil International University
18
a1 4 Forward a right 0
b1 5 Backward b left 1
c1 3 Down c up 2
d1 2 Up d down 3
e1 0 Right e forward 4
f1 1 Left f backward 5
19
a1 5 Backward a right 0
b1 4 Forward b left 1
c1 2 Up c up 2
d1 3 Down d down 3
e1 0 Right e forward 4
f1 1 Left f backward 5
20
a1 4 Forward a right 0
b1 5 Backward b left 1
c1 2 Up c up 2
d1 3 Down d down 3
e1 1 Left e forward 4
f1 0 Right f backward 5
21
a1 4 Forward a right 0
b1 5 Backward b left 1
c1 1 Left c up 2
d1 0 Right d down 3
e1 3 Down e forward 4
f1 2 Up f backward 5
22
a1 5 Backward a right 0
b1 4 Forward b left 1
c1 0 Right c up 2
d1 1 Left d down 3
e1 3 Down e forward 4
f1 2 Up f backward 5
23
a1 4 Forward a right 0
b1 5 Backward b left 1
c1 0 Right c up 2
d1 1 Left d down 3
e1 2 Up e forward 4
f1 3 Down f backward 5
24
a1 5 Backward a right 0
b1 4 Forward b left 1
c1 1 Left c up 2
d1 0 Right d down 3
e1 2 Up e forward 4
f1 3 Down f backward 5
Table 4: Accuracy result of tracking agent’s direction
which compare with handoff agent’s directions
In Table 4, we can see that there have no changes in handoff agent’s directions because handoff
agent’s directions are constant. We can see a change of directions and direction points in tracking
40
© 2017 by Daffodil International University
agent part because these results represented by the perpective of hanoff agent’s directions and
direction points.
5.2 Comparison
To get a thought of how every algorithm performs it is reasonable to plot tackling times in a
histogram. Another way for showing the execution is to sort the unraveling times and plots
confuse record as opposed to explaining time. Both of these are of intrigue however since they
can uncover distinctive things about the algorithms execution.
5.2.1 Comparison with UAV Relative Attitude Estimation Algorithm
UAV relative attitude estimation algorithm that consolidates direction of arrival gauges with
polarimetric multi-radio wire channel appraisals to figure the relative attitude between two
unmanned aerial vehicles [1]. This approach sustainable only for radio frequency based
environment but our propose MARD algorithm can estimate relative direction which approach
can any computer programming environment.
From Figure 13 and Table 5 we can learn that, there are 24 different cases of a tracking agent’s
relative direction aspects. Our MARD algorithm can solve these 24 different cases but UAV
relative attitude estimation algorithm can not solve all cases because this algorithm can not solve
reverse and front cases or reverse and front direction such as, case 2, 3, 4, 6, 8, 10, 12, 14, 16, 17
and 18 which is showed in Table 5 and Figure 13 show a bar graph where this comparison
percentages are shown.
Algorithm Total Case Solved Failed Case
MARD Algorithm 24 Null
UAV Relative Attitude
Estimation Algorithm
13 Case 2, 3, 4, 6, 8,10, 12, 14, 16,
17 and 18
Table 5: Comparison of case solution between two algorithms.
Figure 14 represented the use of area between two algorithms in computer science environment.
UAV Relative Attitude Estimation Algorithm can be used only in in UAV or unmanned aerial
41
© 2017 by Daffodil International University
vehicle area in computer science. Our MARD algorithm approach can be used in any section of
computer science.
Figure 13: Comparison graph of case solution between two algorithms.
Figure 14: Use of area between two algorithms.
UAV
Section
Any Computer Sicence
Section
42
© 2017 by Daffodil International University
5.2.2 Comparison with Others Algorithms
So far fewer researches have been done on relative direction. Most of the research work about
single agent based relative direction but only one research work about multi-agent based relative
direction which published in recent year [1]. So it is very hard to present comparision among
those algorithm and approach. Table 6 represents some multi-agent based and single agent based
relative direction approaches. Before viewing Table 6, first we need to know about Multi-agent
approach and Single agent approach.
 Single agent Approach:
In single agent approach an agent is to look for informative knowledge into the aggregate
conduct of agents which don't really need to be intelligent.
 Multi-agent Approach:
Multi-agent based approach refers a multi-agent system which composed of multiple intelligent
agents inside a domain. The main advantages about multi-agent approach is that when more than
one agent are perform multiple interacting in one computer environment. In multi-agent opinion
isolated agents who cooperate together to a target but in single agent approach could not be
accomplished it because single agent acting alone.
Algorithm Multi-agent
Approach
Single Agent
Approach
MARD Algorithm Multi-agent based
Approach
UAV Relative Attitude
Estimation Algorithm
Multi-agent based
approach
Relative Direction Change
Algorithm
Single agent based
approach
Complexity of Reasoning
with Relative Directions
Algorithm
Single agent based
approach
43
© 2017 by Daffodil International University
Energy-Efficient
Routing Algorithm Based on
Relative Identification
Single agent based
approach
Relative Direction of
Oriented Points Algorithm
Single agent based
approach
An ant colony
optimisation algorithm
for the 2D and 3D
Single agent based
approach
Relative Direction Algorithm
as a Binary Relation
Single agent based
approach
Table 6: Multi-agent based and single agent based algorithm.
44
© 2017 by Daffodil International University
Chapter 6
Discussion
This experiment depended on a greatly rich thought. By setting up a algorithm to keep running
out of sight of the relative direction learning purpose, we could accomplish uncertain measures
of watching time. This enabled us to endeavor a more profound hunt than would have been
conceivable in a period apportioned circumstance. With an exceptionally restricted spending we
assembled and introduced a beneficiary, spectrometer and control programming, all of which
have performed honorably.
6.1 Summary
The objective of this proposal is to look at execution of multi-agent relative direction algorithm
for the getting the hang of learning direction in a multi-agent scheme environment. Our
proposed approach which utilizes diverse sorts of relative directions and two agent. This work
parts into two four principle parts. The initial segment is tied in with examining qualities of
relative direction in displaying a straightforward algorithm. The algorithm fills in as a stage for
getting the hang of steering algorithm. The second piece of this work focuses on scientific
investigation, and how they unravel MARD algorithm. In the third part, this proposition has
tweaked and execute depth view aspects algorithm keeping in mind the end goal to get answer
the principle objective toward the start of this work. In the last part, to make it conceivable to
look at algorithm. This examination shows a modern algorithm for processing the relative
directions between two agents. MARD algorithm examination differentiating direction develops
rules and relative direction guidelines regarding applications computer environment, measuring
both objective and subjective accomplishment. The MARD algorithm can dealing with and
perceiving relative directions, its executed work strategy or computation puts aside greater open
door for result. Its completed computation has few phases for recognizing relative directions.
45
© 2017 by Daffodil International University
6.2 Conclusion
The research gives an introduction to Multi-Agent Relative Direction (MARD) algorithm and
the diverse approaches to manage making profitable solvers. It also displays some hypothetical
establishment about this computation which is of interest while discussing and picking
estimation. Finally the algorithm that will be considered in this suggestion is shown. This
segment portrays the sequencing relative directions in any multi-agent based issue. Features of
an algorithm are analyzed with a particular true objective to arrange the issue into sub-issues
like protest discovering issue, structure issue. This thesis moreover gives an introduction of the
unmanned elevated vehicles and discusses the similitudes and differences between past research
and this theory.
This investigation experiment presents bring in a algorithm for figuring the relative direcrions
between two agenrs, where how one agent can realize in another agent's relative directions
through in PC or programming point of view. In this thesis exhibit our trial outline, practical
info, scientific idea and structure of MARD algorithm. We introduce a novel subjective
delineation and depiction this MARD algorithm can handle these issues for managing and
seeing relative directions, its executed work technique or calculation sets aside more prominent
open entryway for result. General we have seen that the MARD Algorithm is better than other
said ideas. It is the best Algorithm in finding the relative directions and the second-best of
quickest runtime. Versatility of the MARD Algorithm is likewise great.
The target of this thesis is to take a vision at execution of MARD algorithm for the getting the
hang of learning directions in a multi-agent conspire condition. Our proposed approach which
uses different sorts of relative headings and two operators. This work parts into two four
guideline parts. The underlying portion is tied in with looking at characteristics of relative
heading in showing a direct algorithm. The algorithm fills in as a phase for getting the hang of
guiding algorithm. The second bit of this work concentrates on logical examination, and how
they disentangle MARD algorithm. This suggestion has changed and execute profundity see
viewpoints algorithm remembering the true objective to get answer the rule objective toward the
begin of this work. To make it possible to take a algorithm. This examination demonstrates an
advanced algorithm for learninf the relative side between.
MARD algorithm examination separating heading creates tenets and relative directions rules
with respect to applications PC condition, measuring both target and subjective achievement.
The MARD algorithm can managing and seeing relative bearings, its executed work system or
algorithm sets aside more prominent open entryway for result. Its finished calculation has few
stages for perceiving relative directions.
46
© 2017 by Daffodil International University
6.3 Future Work
Future work incorporates examining the conduct of this algorithm in connection to the last
dissemination of execution times. The extensive difference and stochastic conduct no doubt
requests an investigation with access to a lot of computational power. It is additionally
fascinating to think about the impact of various temperature plunge strategies utilized as a part
of relative direction, with restarting being a reasonable contrasting option to unendingly
diminishing temperatures.
Because of significance development of MARD algorithm on the point of view of ―Three
Dimension (3D)‖, ―Machine Learning‖ and ―Route Direction‖, this is fundamental for better
understanding what it takes to find relative directions depiction theoretic hindrances. The
farthest point part issue winds up being solidly related to the set up request of finding relative
direction on the measurement and mapping. In future we will optimistic do a research about
machine learning and route direction perceptive of multi-agent based relative direction learning
computation.
47
© 2017 by Daffodil International University
Bibliography
[1] Attiya Mahmood, Jon W. Wallace, Michael A. Jensen, "Radio Frequency UAV Attitude Estimation
Using Direction of Arrival and Polarization," in 2017 11th European Conference on Antennas and
Propagation (EUCAP), Paris, France, 2017, pp. 1857-1859.
[2] Sebastian Hahn, Joseph Bethge, Jürgen Döllner, "Relative Direction Change - A Topology-based
Metric for Layout Stability in Treemaps," in 12th International Joint Conference on Computer
Vision, Imaging and Computer Graphics Theory and Applications, vol. 3, Porto, Portugal, 2017, pp.
88- 95.
[3] Jonghoon Seo, Seungho Chae, Jinwook Shim, Dongchul Kim, Cheolho Cheong and Tack-Don Han,
"Fast Contour-Tracing Algorithm Based on a Pixel-Following Method for Image Sensors," Sensors,
vol. 16, no. 3, p. 353, 9 March 2016.
[4] Jana Götze, Johan Boye, "―Turn left‖ vs. ―Walk towards the café‖: When relative directions work
better than landmarks," in AGILE 2015: Geographic Information Science as an Enabler of Smarter
Cities and Communities, Lisboa, Portugal, 2015, pp. 253-267.
[5] Jae Hee Lee, "The Complexity of Reasoning with Relative Directions," in Frontiers in Artificial
Intelligence and Applications, 21st European Conference on Artificial Intelligence (ECAI 2014), vol.
263, Prague, Czech Republic, 2014, pp. 507–512.
[6] Jae Hee Lee, Jochen Renz and Diedrich Wolter, "StarVars—Effective Reasoning about Relative
Directions," in Proceedings of the Twenty-Third International Joint Conference on Artificial
Intelligence (IJCAI 2013), Beijing, China, 2013, pp. 976–982.
[7] Chien-Erh Weng and Tsung-Wen Lai, "An Energy-Efficient Routing Algorithm Based on Relative
Identification and Direction for Wireless Sensor Networks," Wireless Personal Communications,
vol. 69, no. 1, pp. 253–268, March 2013.
[8] Till Mossakowski and Reinhard Moratz, "Qualitative Reasoning about Relative Direction of
Oriented Points," Artificial Intelligence, vol. 180–181, pp. 34-45, April 2012.
[9] Toru Ishikawa and Mika Kiyomoto, "Turn to the Left or to the West: Verbal Navigational Directions
in Relative and Absolute Frames of Reference," in 5th International Conference, Geographic
Information Science 2008, Park City, UT, USA, 2008, pp. 119-132.
48
© 2017 by Daffodil International University
[10] Jan Seibert and Brian L. McGlynn, "A new triangular multiple flow direction algorithm for
computing upslope areas from gridded digital elevation models," Water Resources Journal, vol. 43,
no. 4, p. W04501, April 2007.
[11] Alena Shmygelska and Holger H Hoos, "An ant colony optimisation algorithm for the 2D and 3D
hydrophobic polar protein folding problem," BMC Bioinformatics, vol. 6, no. 1, p. 30, 14 February
2005.
[12] Reinhard Moratz, "Representing Relative Direction as a Binary Relation of Oriented Points," in
ECAI 2006, 17th European Conference on Artificial Intelligence, Riva del Garda, Italy, 2006, pp.
Pages 407-411.
[13] Jeounghoon Kim and Hugh R.Wilson, "Dependence of Plaid Motion Coherence on Component
Grating Directions," Vision Research, vol. 33, no. 17, pp. 2479-2489, December 1993.
[14] William S. Albert, Ronald A. Rensink, Jack M. Beusmans, "Learning relative directions between
landmarks in a desktop virtual environment," Spatial Cognition and Computation, vol. 1, no. 2, pp.
131–144, June 1999.
[15] R M Crowder. (1998, January) University of Southampton. [Online]. HYPERLINK
"http://www.southampton.ac.uk/~rmc1/robotics/arkinematics.htm"
[16] Mike Foster. (2016, December) Graphicarto.com. [Online]. HYPERLINK
"http://www.graphicarto.com/directional-cartography-maps-and-relative-direction/"
[17] ArcGIS Desktop. [Online]. HYPERLINK
"http://desktop.arcgis.com/en/arcmap/latest/extensions/schematics/relative-main-line-layout-
algorithm-properties-page.htm"
[18] Seyedali Mirjalili, "Dragonfly algorithm: a new meta-heuristic optimization technique for solving
single-objective, discrete, and multi-objective problems," Neural Computing and Applications, vol.
27, no. 4, pp. 1053–1073, May 2016.
[19] Zhi Li and and Kris Hauser Jianqiao Li, "A Study of Bidirectionally Telepresent Tele-action During
Robot-Mediated Handover," in 2017 IEEE International Conference on Robotics and Automation
(ICRA), Singapore, 2017, pp. 2890-2896.
[20] Richard W. Hedley, Yiwei Huang and Kung Yao, "Direction-of-arrival estimation of animal
vocalizations for monitoring animal behavior and improving estimates of abundance," Avian
Conservation & Ecology, vol. 12, no. 1, Article 6, pp. 116-128, June 2017.
[21] Dunwen Wei, "Discrete-State-Based Vision Navigation Control Algorithm for One Bipedal Robot,"
Mathematical Problems in Engineering, vol. 2015, Article ID 168645, p. 12, May 2015.
[22] Andrew V. Goldberg, Giuseppe F. Italiano, David S. Johnson and Dorothea Wagner, "Algorithm
Engineering (Dagstuhl Seminar 13391)," Dagstuhl Reports, vol. 3, no. 9, pp. 169--189, 2014.
49
© 2017 by Daffodil International University
[23] Marc Goerigk and Anita Schöbel, "Algorithm Engineering in Robust Optimization," in Algorithm
Engineering. Lecture Notes in Computer Science, Lasse Kliemann and Peter Sanders, Ed. Cham:
Springer, 2016, vol. 9220, pp. 245-279.
[24] Markus Chimani and Karsten Klein, "Algorithm Engineering: Concepts and Practice," in
Experimental Methods for the Analysis of Optimization Algorithms, Thomas Bartz-Beielstein, Marco
Chiarandini, Luís Paquete and Mike Preuss, Ed. Berlin, Heidelberg, Germany: Springer, 2010, pp.
131-158.
[25] Matthias Müller-Hannemann and Stefan Schirra, Ed., Algorithm Engineering. Berlin, Heidelberg,
Germany: Springer, 2010, vol. 5971.
[26] Peter Sanders, "Algorithm Engineering –An Attempt at a Definition," in Efficient Algorithms.
Lecture Notes in Computer Science, Susanne Albers, Helmut Alt and Stefan Näher, Ed. Berlin,
Heidelberg, Germany: Springer, 2009, vol. 5760, pp. 321-340.
[27] Yoav Shoham and Kevin Leyton-Brown, Multiagent Systems: Algorithmic, Game-Theoretic, and
Logical Foundations, 1st ed. New York, USA: Cambridge University Press, December 2008.
[28] Mevludin Glavic, "Agents and Multi-Agent Systems: A Short Introduction for Power Engineers,"
University of Liege, Electrical Engineering and Computer Science Department, Liege, Belgium,
Technical Report May 2006.
[29] Hjelmervik H, Westerhausen , Hirnstein M, Specht K and Hausmann M, "The neural correlates of
sex differences in left–right confusion," NeuroImage, vol. 113, pp. 196–206, June 2015.
[30] Brandt J and Mackavey W, "Left-right confusion and the perception of bilateral symmetry,"
International Journal of Neuroscience, vol. 12, no. 2, pp. 87-94, 1981.
[31] Hannay HJ, Ciaccia PJ, Kerr JW and Barrett D, "Self-report of right-left confusion in college men
and women," Perceptual and Motor Skills, vol. 70, no. 2, pp. 451-457, April 1990.
[32] Hikari Yamashita, "Self-rated right–left confusability and performance on the Money Road-Map
Test," Psychological Research, vol. 77, no. 5, pp. 575–582, September 2013.
[33] E.B. Lum, A. Stompel and Kwan-Liu Ma, "Using Motion to Illustrate Static 3D Shape - Kinetic
Visualization," IEEE Transactions on Visualization and Computer Graphics (TVCG), vol. 9, no. 2,
pp. 115 - 126, April-June 2003.
[34] C.R. Karanam and Y. Mostofi, "3D Through-Wall Imagingwith Unmanned Aerial Vehicles Using
WiFi," in the proceedings of the 16th ACM/IEEE International Conference on Information
Processing in Sensor Networks (IPSN), Pittsburgh, Pennsylvania, USA, April 2017, pp. 131-142.
[35] S. Rayhan Kabir, Shaikh Muhammad Allayear, Mirza Mohtashim Alam and Md. Tahsir Ahmed
Munna, ―A Computational Technique for Intelligent Computers to learn and identify the Human’s
Relative Directions,‖ proceeding in International Conference on Intelligent Sustainable Systems
(ICISS 2017), India, 2017, pp. 1037-1040.
Computation of Multi-Agent Based Relative Direction Learning Specification

Weitere ähnliche Inhalte

Was ist angesagt?

LF MVNIET - Computer Science Engineering
LF MVNIET - Computer Science EngineeringLF MVNIET - Computer Science Engineering
LF MVNIET - Computer Science EngineeringLFMVNIET
 
A black-box-approach-for-response-quality-evaluation-of-conversational-agent-...
A black-box-approach-for-response-quality-evaluation-of-conversational-agent-...A black-box-approach-for-response-quality-evaluation-of-conversational-agent-...
A black-box-approach-for-response-quality-evaluation-of-conversational-agent-...Cemal Ardil
 
Dr. SUJATHA RAMAKRISHNAN CURRICULUM VITAE
Dr. SUJATHA RAMAKRISHNAN CURRICULUM VITAEDr. SUJATHA RAMAKRISHNAN CURRICULUM VITAE
Dr. SUJATHA RAMAKRISHNAN CURRICULUM VITAEDrSUJATHARAMAKRISHNA
 
IRJET- College Enquiry Chatbot System(DMCE)
IRJET-  	  College Enquiry Chatbot System(DMCE)IRJET-  	  College Enquiry Chatbot System(DMCE)
IRJET- College Enquiry Chatbot System(DMCE)IRJET Journal
 
Virtual education system
Virtual education systemVirtual education system
Virtual education systemDhara024
 
iBaTs: Interactive Bash Shell Adaptive Tutoring System
iBaTs: Interactive Bash Shell Adaptive Tutoring SystemiBaTs: Interactive Bash Shell Adaptive Tutoring System
iBaTs: Interactive Bash Shell Adaptive Tutoring SystemCSCJournals
 
Top cited articles 2020 - Advanced Computational Intelligence: An Internation...
Top cited articles 2020 - Advanced Computational Intelligence: An Internation...Top cited articles 2020 - Advanced Computational Intelligence: An Internation...
Top cited articles 2020 - Advanced Computational Intelligence: An Internation...aciijournal
 
A Survey on Using Artificial Intelligence Techniques in the Software Developm...
A Survey on Using Artificial Intelligence Techniques in the Software Developm...A Survey on Using Artificial Intelligence Techniques in the Software Developm...
A Survey on Using Artificial Intelligence Techniques in the Software Developm...IJERA Editor
 
CV_Salim_August-2016
CV_Salim_August-2016CV_Salim_August-2016
CV_Salim_August-2016SALIM ISTYAQ
 
IRJET - Automatic Lip Reading: Classification of Words and Phrases using Conv...
IRJET - Automatic Lip Reading: Classification of Words and Phrases using Conv...IRJET - Automatic Lip Reading: Classification of Words and Phrases using Conv...
IRJET - Automatic Lip Reading: Classification of Words and Phrases using Conv...IRJET Journal
 
Dr. Parkavi.A , Resume working-2020-v4
Dr. Parkavi.A , Resume working-2020-v4Dr. Parkavi.A , Resume working-2020-v4
Dr. Parkavi.A , Resume working-2020-v4Parkavi A
 
Multimedia based IoT-centric smart framework for eLearning paradigm Muhammad ...
Multimedia based IoT-centric smart framework for eLearning paradigm Muhammad ...Multimedia based IoT-centric smart framework for eLearning paradigm Muhammad ...
Multimedia based IoT-centric smart framework for eLearning paradigm Muhammad ...eraser Juan José Calderón
 
Top research articles in 2020 in the area of Computer Science & Engineering
Top research articles in 2020 in the area of Computer Science &  EngineeringTop research articles in 2020 in the area of Computer Science &  Engineering
Top research articles in 2020 in the area of Computer Science & Engineeringgerogepatton
 

Was ist angesagt? (18)

Multiple Instance E-Learning Behavioural Coding
Multiple Instance E-Learning Behavioural CodingMultiple Instance E-Learning Behavioural Coding
Multiple Instance E-Learning Behavioural Coding
 
FINALCV - Copy
FINALCV - CopyFINALCV - Copy
FINALCV - Copy
 
LF MVNIET - Computer Science Engineering
LF MVNIET - Computer Science EngineeringLF MVNIET - Computer Science Engineering
LF MVNIET - Computer Science Engineering
 
A black-box-approach-for-response-quality-evaluation-of-conversational-agent-...
A black-box-approach-for-response-quality-evaluation-of-conversational-agent-...A black-box-approach-for-response-quality-evaluation-of-conversational-agent-...
A black-box-approach-for-response-quality-evaluation-of-conversational-agent-...
 
Dr. SUJATHA RAMAKRISHNAN CURRICULUM VITAE
Dr. SUJATHA RAMAKRISHNAN CURRICULUM VITAEDr. SUJATHA RAMAKRISHNAN CURRICULUM VITAE
Dr. SUJATHA RAMAKRISHNAN CURRICULUM VITAE
 
resume -1-
resume -1-resume -1-
resume -1-
 
IRJET- College Enquiry Chatbot System(DMCE)
IRJET-  	  College Enquiry Chatbot System(DMCE)IRJET-  	  College Enquiry Chatbot System(DMCE)
IRJET- College Enquiry Chatbot System(DMCE)
 
Virtual education system
Virtual education systemVirtual education system
Virtual education system
 
iBaTs: Interactive Bash Shell Adaptive Tutoring System
iBaTs: Interactive Bash Shell Adaptive Tutoring SystemiBaTs: Interactive Bash Shell Adaptive Tutoring System
iBaTs: Interactive Bash Shell Adaptive Tutoring System
 
Top cited articles 2020 - Advanced Computational Intelligence: An Internation...
Top cited articles 2020 - Advanced Computational Intelligence: An Internation...Top cited articles 2020 - Advanced Computational Intelligence: An Internation...
Top cited articles 2020 - Advanced Computational Intelligence: An Internation...
 
A Survey on Using Artificial Intelligence Techniques in the Software Developm...
A Survey on Using Artificial Intelligence Techniques in the Software Developm...A Survey on Using Artificial Intelligence Techniques in the Software Developm...
A Survey on Using Artificial Intelligence Techniques in the Software Developm...
 
CV_Salim_August-2016
CV_Salim_August-2016CV_Salim_August-2016
CV_Salim_August-2016
 
IRJET - Automatic Lip Reading: Classification of Words and Phrases using Conv...
IRJET - Automatic Lip Reading: Classification of Words and Phrases using Conv...IRJET - Automatic Lip Reading: Classification of Words and Phrases using Conv...
IRJET - Automatic Lip Reading: Classification of Words and Phrases using Conv...
 
Dr. Parkavi.A , Resume working-2020-v4
Dr. Parkavi.A , Resume working-2020-v4Dr. Parkavi.A , Resume working-2020-v4
Dr. Parkavi.A , Resume working-2020-v4
 
Multimedia based IoT-centric smart framework for eLearning paradigm Muhammad ...
Multimedia based IoT-centric smart framework for eLearning paradigm Muhammad ...Multimedia based IoT-centric smart framework for eLearning paradigm Muhammad ...
Multimedia based IoT-centric smart framework for eLearning paradigm Muhammad ...
 
Top research articles in 2020 in the area of Computer Science & Engineering
Top research articles in 2020 in the area of Computer Science &  EngineeringTop research articles in 2020 in the area of Computer Science &  Engineering
Top research articles in 2020 in the area of Computer Science & Engineering
 
Nabhiraijain cdac
Nabhiraijain cdac Nabhiraijain cdac
Nabhiraijain cdac
 
Surface Analysis Techniques Feb & April 2013
Surface Analysis Techniques Feb & April 2013Surface Analysis Techniques Feb & April 2013
Surface Analysis Techniques Feb & April 2013
 

Ähnlich wie Computation of Multi-Agent Based Relative Direction Learning Specification

Iimsr student management system
Iimsr student management systemIimsr student management system
Iimsr student management systemSHUJA SHABBIR
 
CS499_JULIUS_J_FINAL_YEAR_PROJETCT_L_DRAFT
CS499_JULIUS_J_FINAL_YEAR_PROJETCT_L_DRAFTCS499_JULIUS_J_FINAL_YEAR_PROJETCT_L_DRAFT
CS499_JULIUS_J_FINAL_YEAR_PROJETCT_L_DRAFTJosephat Julius
 
Document Archiving & Sharing System
Document Archiving & Sharing SystemDocument Archiving & Sharing System
Document Archiving & Sharing SystemAshik Iqbal
 
Student portal system application -Project Book
Student portal system application -Project BookStudent portal system application -Project Book
Student portal system application -Project BookS.M. Fazla Rabbi
 
Portfolio investment diversification.pdf
Portfolio investment diversification.pdfPortfolio investment diversification.pdf
Portfolio investment diversification.pdfSalmanKhan222894
 
DRDO PROJECT REPORT1
DRDO PROJECT REPORT1DRDO PROJECT REPORT1
DRDO PROJECT REPORT1Dikshya Rath
 
VTU final year project report
VTU final year project reportVTU final year project report
VTU final year project reportathiathi3
 
A Physicomimetics Desynchronization Algorithm without Global Time Knowledge f...
A Physicomimetics Desynchronization Algorithm without Global Time Knowledge f...A Physicomimetics Desynchronization Algorithm without Global Time Knowledge f...
A Physicomimetics Desynchronization Algorithm without Global Time Knowledge f...Supasate Choochaisri
 
A Dynamic Middleware-based Instrumentation Framework to Assist the Understand...
A Dynamic Middleware-based Instrumentation Framework to Assist the Understand...A Dynamic Middleware-based Instrumentation Framework to Assist the Understand...
A Dynamic Middleware-based Instrumentation Framework to Assist the Understand...Luz Martinez
 
ANSYS Fluent - CFD Final year thesis
ANSYS Fluent - CFD Final year thesisANSYS Fluent - CFD Final year thesis
ANSYS Fluent - CFD Final year thesisDibyajyoti Laha
 

Ähnlich wie Computation of Multi-Agent Based Relative Direction Learning Specification (20)

Internship at SELISE
Internship at SELISEInternship at SELISE
Internship at SELISE
 
Iimsr student management system
Iimsr student management systemIimsr student management system
Iimsr student management system
 
CS499_JULIUS_J_FINAL_YEAR_PROJETCT_L_DRAFT
CS499_JULIUS_J_FINAL_YEAR_PROJETCT_L_DRAFTCS499_JULIUS_J_FINAL_YEAR_PROJETCT_L_DRAFT
CS499_JULIUS_J_FINAL_YEAR_PROJETCT_L_DRAFT
 
Document Archiving & Sharing System
Document Archiving & Sharing SystemDocument Archiving & Sharing System
Document Archiving & Sharing System
 
Student portal system application -Project Book
Student portal system application -Project BookStudent portal system application -Project Book
Student portal system application -Project Book
 
Kumar_Akshat
Kumar_AkshatKumar_Akshat
Kumar_Akshat
 
Portfolio investment diversification.pdf
Portfolio investment diversification.pdfPortfolio investment diversification.pdf
Portfolio investment diversification.pdf
 
INSTITUTION -WEBSITE
INSTITUTION -WEBSITEINSTITUTION -WEBSITE
INSTITUTION -WEBSITE
 
1
11
1
 
november001 (1)
november001 (1)november001 (1)
november001 (1)
 
DRDO PROJECT REPORT1
DRDO PROJECT REPORT1DRDO PROJECT REPORT1
DRDO PROJECT REPORT1
 
VTU final year project report
VTU final year project reportVTU final year project report
VTU final year project report
 
3 job adda doc 1
3 job adda doc 13 job adda doc 1
3 job adda doc 1
 
A Physicomimetics Desynchronization Algorithm without Global Time Knowledge f...
A Physicomimetics Desynchronization Algorithm without Global Time Knowledge f...A Physicomimetics Desynchronization Algorithm without Global Time Knowledge f...
A Physicomimetics Desynchronization Algorithm without Global Time Knowledge f...
 
Bitwise Vol 1
Bitwise Vol 1Bitwise Vol 1
Bitwise Vol 1
 
A Dynamic Middleware-based Instrumentation Framework to Assist the Understand...
A Dynamic Middleware-based Instrumentation Framework to Assist the Understand...A Dynamic Middleware-based Instrumentation Framework to Assist the Understand...
A Dynamic Middleware-based Instrumentation Framework to Assist the Understand...
 
1227201 Report
1227201 Report1227201 Report
1227201 Report
 
ANSYS Fluent - CFD Final year thesis
ANSYS Fluent - CFD Final year thesisANSYS Fluent - CFD Final year thesis
ANSYS Fluent - CFD Final year thesis
 
Project documentaion sample.docx
Project documentaion sample.docxProject documentaion sample.docx
Project documentaion sample.docx
 
CV_firdaus
CV_firdausCV_firdaus
CV_firdaus
 

Kürzlich hochgeladen

4.16.24 Poverty and Precarity--Desmond.pptx
4.16.24 Poverty and Precarity--Desmond.pptx4.16.24 Poverty and Precarity--Desmond.pptx
4.16.24 Poverty and Precarity--Desmond.pptxmary850239
 
Incoming and Outgoing Shipments in 3 STEPS Using Odoo 17
Incoming and Outgoing Shipments in 3 STEPS Using Odoo 17Incoming and Outgoing Shipments in 3 STEPS Using Odoo 17
Incoming and Outgoing Shipments in 3 STEPS Using Odoo 17Celine George
 
ICS2208 Lecture6 Notes for SL spaces.pdf
ICS2208 Lecture6 Notes for SL spaces.pdfICS2208 Lecture6 Notes for SL spaces.pdf
ICS2208 Lecture6 Notes for SL spaces.pdfVanessa Camilleri
 
4.18.24 Movement Legacies, Reflection, and Review.pptx
4.18.24 Movement Legacies, Reflection, and Review.pptx4.18.24 Movement Legacies, Reflection, and Review.pptx
4.18.24 Movement Legacies, Reflection, and Review.pptxmary850239
 
USPS® Forced Meter Migration - How to Know if Your Postage Meter Will Soon be...
USPS® Forced Meter Migration - How to Know if Your Postage Meter Will Soon be...USPS® Forced Meter Migration - How to Know if Your Postage Meter Will Soon be...
USPS® Forced Meter Migration - How to Know if Your Postage Meter Will Soon be...Postal Advocate Inc.
 
How to do quick user assign in kanban in Odoo 17 ERP
How to do quick user assign in kanban in Odoo 17 ERPHow to do quick user assign in kanban in Odoo 17 ERP
How to do quick user assign in kanban in Odoo 17 ERPCeline George
 
AUDIENCE THEORY -CULTIVATION THEORY - GERBNER.pptx
AUDIENCE THEORY -CULTIVATION THEORY -  GERBNER.pptxAUDIENCE THEORY -CULTIVATION THEORY -  GERBNER.pptx
AUDIENCE THEORY -CULTIVATION THEORY - GERBNER.pptxiammrhaywood
 
Dust Of Snow By Robert Frost Class-X English CBSE
Dust Of Snow By Robert Frost Class-X English CBSEDust Of Snow By Robert Frost Class-X English CBSE
Dust Of Snow By Robert Frost Class-X English CBSEaurabinda banchhor
 
Student Profile Sample - We help schools to connect the data they have, with ...
Student Profile Sample - We help schools to connect the data they have, with ...Student Profile Sample - We help schools to connect the data they have, with ...
Student Profile Sample - We help schools to connect the data they have, with ...Seán Kennedy
 
Keynote by Prof. Wurzer at Nordex about IP-design
Keynote by Prof. Wurzer at Nordex about IP-designKeynote by Prof. Wurzer at Nordex about IP-design
Keynote by Prof. Wurzer at Nordex about IP-designMIPLM
 
Active Learning Strategies (in short ALS).pdf
Active Learning Strategies (in short ALS).pdfActive Learning Strategies (in short ALS).pdf
Active Learning Strategies (in short ALS).pdfPatidar M
 
Integumentary System SMP B. Pharm Sem I.ppt
Integumentary System SMP B. Pharm Sem I.pptIntegumentary System SMP B. Pharm Sem I.ppt
Integumentary System SMP B. Pharm Sem I.pptshraddhaparab530
 
4.16.24 21st Century Movements for Black Lives.pptx
4.16.24 21st Century Movements for Black Lives.pptx4.16.24 21st Century Movements for Black Lives.pptx
4.16.24 21st Century Movements for Black Lives.pptxmary850239
 
GRADE 4 - SUMMATIVE TEST QUARTER 4 ALL SUBJECTS
GRADE 4 - SUMMATIVE TEST QUARTER 4 ALL SUBJECTSGRADE 4 - SUMMATIVE TEST QUARTER 4 ALL SUBJECTS
GRADE 4 - SUMMATIVE TEST QUARTER 4 ALL SUBJECTSJoshuaGantuangco2
 
Q4-PPT-Music9_Lesson-1-Romantic-Opera.pptx
Q4-PPT-Music9_Lesson-1-Romantic-Opera.pptxQ4-PPT-Music9_Lesson-1-Romantic-Opera.pptx
Q4-PPT-Music9_Lesson-1-Romantic-Opera.pptxlancelewisportillo
 
THEORIES OF ORGANIZATION-PUBLIC ADMINISTRATION
THEORIES OF ORGANIZATION-PUBLIC ADMINISTRATIONTHEORIES OF ORGANIZATION-PUBLIC ADMINISTRATION
THEORIES OF ORGANIZATION-PUBLIC ADMINISTRATIONHumphrey A Beña
 
Transaction Management in Database Management System
Transaction Management in Database Management SystemTransaction Management in Database Management System
Transaction Management in Database Management SystemChristalin Nelson
 
ROLES IN A STAGE PRODUCTION in arts.pptx
ROLES IN A STAGE PRODUCTION in arts.pptxROLES IN A STAGE PRODUCTION in arts.pptx
ROLES IN A STAGE PRODUCTION in arts.pptxVanesaIglesias10
 
Concurrency Control in Database Management system
Concurrency Control in Database Management systemConcurrency Control in Database Management system
Concurrency Control in Database Management systemChristalin Nelson
 

Kürzlich hochgeladen (20)

4.16.24 Poverty and Precarity--Desmond.pptx
4.16.24 Poverty and Precarity--Desmond.pptx4.16.24 Poverty and Precarity--Desmond.pptx
4.16.24 Poverty and Precarity--Desmond.pptx
 
Incoming and Outgoing Shipments in 3 STEPS Using Odoo 17
Incoming and Outgoing Shipments in 3 STEPS Using Odoo 17Incoming and Outgoing Shipments in 3 STEPS Using Odoo 17
Incoming and Outgoing Shipments in 3 STEPS Using Odoo 17
 
ICS2208 Lecture6 Notes for SL spaces.pdf
ICS2208 Lecture6 Notes for SL spaces.pdfICS2208 Lecture6 Notes for SL spaces.pdf
ICS2208 Lecture6 Notes for SL spaces.pdf
 
4.18.24 Movement Legacies, Reflection, and Review.pptx
4.18.24 Movement Legacies, Reflection, and Review.pptx4.18.24 Movement Legacies, Reflection, and Review.pptx
4.18.24 Movement Legacies, Reflection, and Review.pptx
 
USPS® Forced Meter Migration - How to Know if Your Postage Meter Will Soon be...
USPS® Forced Meter Migration - How to Know if Your Postage Meter Will Soon be...USPS® Forced Meter Migration - How to Know if Your Postage Meter Will Soon be...
USPS® Forced Meter Migration - How to Know if Your Postage Meter Will Soon be...
 
How to do quick user assign in kanban in Odoo 17 ERP
How to do quick user assign in kanban in Odoo 17 ERPHow to do quick user assign in kanban in Odoo 17 ERP
How to do quick user assign in kanban in Odoo 17 ERP
 
AUDIENCE THEORY -CULTIVATION THEORY - GERBNER.pptx
AUDIENCE THEORY -CULTIVATION THEORY -  GERBNER.pptxAUDIENCE THEORY -CULTIVATION THEORY -  GERBNER.pptx
AUDIENCE THEORY -CULTIVATION THEORY - GERBNER.pptx
 
Dust Of Snow By Robert Frost Class-X English CBSE
Dust Of Snow By Robert Frost Class-X English CBSEDust Of Snow By Robert Frost Class-X English CBSE
Dust Of Snow By Robert Frost Class-X English CBSE
 
Student Profile Sample - We help schools to connect the data they have, with ...
Student Profile Sample - We help schools to connect the data they have, with ...Student Profile Sample - We help schools to connect the data they have, with ...
Student Profile Sample - We help schools to connect the data they have, with ...
 
Keynote by Prof. Wurzer at Nordex about IP-design
Keynote by Prof. Wurzer at Nordex about IP-designKeynote by Prof. Wurzer at Nordex about IP-design
Keynote by Prof. Wurzer at Nordex about IP-design
 
Active Learning Strategies (in short ALS).pdf
Active Learning Strategies (in short ALS).pdfActive Learning Strategies (in short ALS).pdf
Active Learning Strategies (in short ALS).pdf
 
Integumentary System SMP B. Pharm Sem I.ppt
Integumentary System SMP B. Pharm Sem I.pptIntegumentary System SMP B. Pharm Sem I.ppt
Integumentary System SMP B. Pharm Sem I.ppt
 
4.16.24 21st Century Movements for Black Lives.pptx
4.16.24 21st Century Movements for Black Lives.pptx4.16.24 21st Century Movements for Black Lives.pptx
4.16.24 21st Century Movements for Black Lives.pptx
 
GRADE 4 - SUMMATIVE TEST QUARTER 4 ALL SUBJECTS
GRADE 4 - SUMMATIVE TEST QUARTER 4 ALL SUBJECTSGRADE 4 - SUMMATIVE TEST QUARTER 4 ALL SUBJECTS
GRADE 4 - SUMMATIVE TEST QUARTER 4 ALL SUBJECTS
 
Q4-PPT-Music9_Lesson-1-Romantic-Opera.pptx
Q4-PPT-Music9_Lesson-1-Romantic-Opera.pptxQ4-PPT-Music9_Lesson-1-Romantic-Opera.pptx
Q4-PPT-Music9_Lesson-1-Romantic-Opera.pptx
 
THEORIES OF ORGANIZATION-PUBLIC ADMINISTRATION
THEORIES OF ORGANIZATION-PUBLIC ADMINISTRATIONTHEORIES OF ORGANIZATION-PUBLIC ADMINISTRATION
THEORIES OF ORGANIZATION-PUBLIC ADMINISTRATION
 
Transaction Management in Database Management System
Transaction Management in Database Management SystemTransaction Management in Database Management System
Transaction Management in Database Management System
 
FINALS_OF_LEFT_ON_C'N_EL_DORADO_2024.pptx
FINALS_OF_LEFT_ON_C'N_EL_DORADO_2024.pptxFINALS_OF_LEFT_ON_C'N_EL_DORADO_2024.pptx
FINALS_OF_LEFT_ON_C'N_EL_DORADO_2024.pptx
 
ROLES IN A STAGE PRODUCTION in arts.pptx
ROLES IN A STAGE PRODUCTION in arts.pptxROLES IN A STAGE PRODUCTION in arts.pptx
ROLES IN A STAGE PRODUCTION in arts.pptx
 
Concurrency Control in Database Management system
Concurrency Control in Database Management systemConcurrency Control in Database Management system
Concurrency Control in Database Management system
 

Computation of Multi-Agent Based Relative Direction Learning Specification

  • 1. © 2017 by Daffodil International University Computation of Multi-Agent Based Relative Direction Learning Specification Submitted By S. Rayhan Kabir ID: 133-35-561 Bachelor of Science in Software Engineering Department of Software Engineering Daffodil International University rayhan561@diu.edu.bd Supervised by Dr. Shaikh Muhammad Allayear Associate Professor Department of Software Engineering Associate Professor and Head Department of Multimedia & Creative Technology Daffodil International University Submitted Date: September 2017 Daffodil International University Dhaka, Bangladesh
  • 2. I © 2017 by Daffodil International University DECLARATION It hereby announces that, this bachelor thesis under the supervision of Dr. Shaikh Muhammad Allayear, Associate Professor, Department of Software Engineering, Daffodil International University. It is also declared that neither this thesis nor any part of this has been submitted elsewhere for award of any degree. Submitted by: ……………………………………… S. Rayhan Kabir ID: 133-35-561 Batch: 12th Department of Software Engineering Daffodil International University Faculty of Science & Information Technology Daffodil International University Certified by: ……………………………………… Dr. Shaikh Muhammad Allayear Associate Professor Department of Software Engineering Associate Professor and Head Daffodil International University Department of Multimedia & Creative Technology Faculty of Science & Information Technology
  • 3. II © 2017 by Daffodil International University APPROVAL This bachelor thesis titled “Computation of Multi-Agent Based Relative Direction Learning Specification”, submitted by S. Rayhan Kabir, ID: 133-35-561 to the Department of Software Engineering, Daffodil International University has been accepted as satisfactory for the partial fulfillment of the requirements for the degree of B.Sc. in Software Engineering (SWE) and approved as to its style and contents. BOARD OF EXAMINERS ----------------------------------------------- Dr. TouhidBhuiyan Professor and Head Department of Software Engineering Faculty of Science and Information Technology Daffodil International University Chairman ----------------------------------------------- Dr.Md. Asraf Ali Associate Professor Department of Software Engineering Faculty of Science and Information Technology Daffodil International University Internal Examiner 1 ----------------------------------------------- Manan Binth Taj Noor Lecturer Department of Software Engineering Faculty of Science and Information Technology Daffodil International University Internal Examiner 2 ----------------------------------------------- Dr. Md. Nasim Akhtar Professor and Chairman Department of Computer Science Engineering Dhaka University of Engineering & Technology, Gazipur External Examiner
  • 4. III © 2017 by Daffodil International University Abstract The most widely recognized relative directions are left, right, up, down, backward and forward. This research paper presents another algorithm for computing the relative directions between two agents, where one agent can learn another agent’s relative directions. We exhibit a study contrasting direction construct guidelines and relative direction instructions with respect to people on foot in a genuine city condition, measuring both goal and subjective achievement. Eyewitnesses commonly depict their condition by determining the relative directions in which they see different items or other individuals from their perspective. Be that as it may, it is surprisingly difficult to integrate relative directions got from various observers between two agents. In this paper, we introduce a novel subjective portrayal and representation of this MultiAgent Relative Direction (MARD) algorithm can solve these problems. For handling and recognizing relative direction, its executed work method or computation sets aside more opportunity for the result. Its actualized computation has few stages for distinguishing relative direction. Endeavor to diminish one stage and expected the outcome will be quicker.
  • 5. IV © 2017 by Daffodil International University Acknowledgments Firstly, I might want to thank my supervisor, Associate Professor Dr. Shaikh Muhammad Allayear. I owe such a great amount to his motivating direction over the span of this venture, for his recommendations on papers to peruse, and for his endless hours of accommodating exchanges and assessment. He gives me an opportunity to work in Smart Data Science Center (SDSC) for complete my research. SDSC is a computer research laboratory of Daffodil International University. I might likewise want to demonstrate appreciation to my committee, including Associate Professor Dr. Md. Asraf Ali, Chairman, Project/Thesis Committee, Department of Software Engineering and my noteworthy our department head Professor Dr.Touhid Bhuiyan for their profitable instructions. All the more by and large, I can't exaggerate the amount Daffodil International University's software engineering offices have helped me develop as an understudy. Uncommon much gratitude goes Assistant Professor Imran Mahmud for putting me on the way to seeking after hypothetical software engineering research and for being a uniquely rousing coach and to lecturer Ms. Manan Binth Taj Noor for filling in as my scholarly consultant. I might likewise want to thank lecturer Ms. Fouzia Rahman for teaching one incredible course that truly got me amped up for a few complex factors. I would also like to thank lecturer Ms. Rubaida Easmin for her incredibly extensive and important input on the evidence of my principle result. I'm particularly appreciative for Research Associate MD. Tahsir Ahmed Munna and lecturer Mirza Mohtashim Alam for being regular teammates on issue sets. I'm particularly thankful for my paper observers Assistant Professor K. M. Imtiaz-Ud-Din and lecturer Md. Anwar Hossen for being continuous collaborators on issue sets and their suggestions on research report to peruse or read. Some extended and customized parts of this thesis recently got an opportunity to present at International Conference on Intelligent Sustainable Systems (ICISS 2017), in India and publisher is IEEE Xplore [35]. There I have described how intelligent computer learn and identify human's relative directions which based on this thesis dissertation. I hope I can successfully showcase my research in this conference. Lastly, I might want to thank my parents for bringing me into this world and making everything conceivable. They were the reason I initially began to look all starry eyed at learning, and I am appreciative consistently for what they have done to raise me up to be simply the best form.
  • 6. V © 2017 by Daffodil International University Table of Contents Abstract......................................................................................................................................... III Acknowledgments.........................................................................................................................IV Introduction..................................................................................................................................... 1 1.1 Overview ........................................................................................................................... 1 1.2 Research Objectives .......................................................................................................... 2 1.3 Research Questions............................................................................................................ 2 1.4 Organization ...................................................................................................................... 3 1.5 Definitions ......................................................................................................................... 4 1.6 Motivation of Research ..................................................................................................... 4 Background and Literature Review ................................................................................................ 6 2.1 Relative Direction Fundamentals ...................................................................................... 6 2.2 Multi-Agent System Environment..................................................................................... 7 2.3 Previous Research and Work............................................................................................. 8 2.4 Left-Right Confusion....................................................................................................... 13 2.4.1 Research about Left-Right Confusion........................................................................ 14 2.4.2 Artificial Intelligence Perspective .............................................................................. 14 Proposed Algorithm Model........................................................................................................... 18 3.1 Induction.......................................................................................................................... 18 3.2 Tracking Agent and Handoff Agent ................................................................................ 19 3.2.1 Direction Points.......................................................................................................... 19 3.2.2 Various 3D Aspects.................................................................................................... 20 3.2.3 Realistic Inputs ........................................................................................................... 21 3.3 Mathematical Exploration ............................................................................................... 22 3.4 Structure of Algorithm..................................................................................................... 22 3.5 Identify Relative Direction .............................................................................................. 25 3.6 Machine Learning Aspect................................................................................................ 26 3.7 Alternative MARD Algorithm Approach........................................................................ 26 3.8 Multi-Agent Route Direction........................................................................................... 31 Methodology................................................................................................................................. 33 4.1 Algorithm Engineering Method....................................................................................... 33 4.2 Experiment Control ......................................................................................................... 35 Results and Analysis..................................................................................................................... 36 5.1 Accuracy Result............................................................................................................... 36 5.2 Comparison...................................................................................................................... 40 5.2.1 Comparison with UAV Relative Attitude Estimation Algorithm .............................. 40 5.2.2 Comparison with Others Algorithms.......................................................................... 42 Discussion..................................................................................................................................... 44 6.1 Summary.......................................................................................................................... 44 6.2 Conclusion....................................................................................................................... 45 6.3 Future Work..................................................................................................................... 46 Bibliography ................................................................................................................................. 47
  • 7. VI © 2017 by Daffodil International University List of Tables Table 1: Relative directions and their numerical values..................................................... 21 Table 2: Condition of Cases................................................................................................ 27 Table 3: Order of Cases. ..................................................................................................... 28 Table 4: Accuracy result ..................................................................................................... 39 Table 5: Comparison of case solution between two algorithms. ........................................ 40 Table 6: Multi-agent based and single agent based algorithm............................................ 43
  • 8. VII © 2017 by Daffodil International University List of Figures Figure 1: Different types of relative directions. ................................................................ 7 Figure 2: Characteristics of Multi-agent System Environment......................................... 8 Figure 3: Tracking UAV and Handoff UAV.................................................................... 9 Figure 4: Relative directions (right and left) of two agents............................................. 15 Figure 5: Agent A ordered agent R to find the mobile object ................................... 16 Figure 6: Agent A calculated that his right direction is equal to agent .......................... 16 Figure 7: Agent A ordered agent R ................................................................................. 17 Figure 8: Direction points of Handoff Agent and Tracking Agent. ................................ 19 Figure 9: Different cases of tracking agent in 3D aspect. ............................................... 20 Figure 10: Tracking agent's direction identification........................................................ 25 Figure 11: Multi-agent based route direction in mapping. .............................................. 32 Figure 12: Algorithm Engineering methodology cycle................................................... 34 Figure 13: Comparison graph of case solution between two algorithms. ....................... 41 Figure 14: Use of area between two algorithms.............................................................. 41
  • 9. VIII © 2017 by Daffodil International University List of Algorithms Algorithm 1: Multi-Agent Relative Direction Algorithm.................................................. 24 Algorithm 2: Direction Identification Algorithm............................................................... 26 Algorithm 3: Alternative MARD Algorithm...................................................................... 30
  • 10. 1 © 2017 by Daffodil International University Chapter 1 Introduction The most generally perceived relative directions are left, right, up, down, backward and forward. A man giving relative directions will utilize center terms. A multi-agent framework is an automated framework made out of different associating keen operators inside a domain. Multi- agent frameworks can be utilized to take care of issues that are troublesome or inconceivable for an individual agent or a solid framework to understand. This paper explores another computation for assessing the relative direction between two agents. 1.1 Overview Learning and identify relative directions in the multi-agent based system depends on three pairs of relative directions which are forward and backward, left and right, up and down. The experiment or research of multi-agent based relative direction learning the algorithmic process for dealing with computational issues, while one agent wants to learn another agent's relative directions. In here we show a new algorithm for processing the relative directions between two agents, how one agent can learn another agent’s relative headings in 3D perspective. In this paper, we introduce about Multi-Agent Relative Direction (MARD) algorithm concept which can represent this issue. Since the disclosure of "Plaid Motion Coherence on Component Grating Directions" by Jeounghoon Kim and Hugh R.Wilson in 1993 and its understanding unmistakably demonstrates that coherence of movement for 2D designs in various spatial scales depends basically on the relative direction of movement of part gratings [13]. Intellectual mapping research has generally centered on how people explore and procure spatial data about genuine situations. An old research which presents two examinations that analyze how people learn relative direction between landmarks in a desktop virtual condition (William S. Albert, Ronald A. Rensink and
  • 11. 2 © 2017 by Daffodil International University Jack M. Beusmans, 1999). This relative direction test included showing the course of concealed landmarks from various vantage focuses in nature [14]. In 2017 a research which researchers are Attiya Mahmood, Jon W. Wallace, and Michael A. Jensen. They reveal estimation that, given any announcement in Unmanned Aerial Vehicles (UAV), where estimation of relative attitude between two unmanned flying vehicles which in view of different information and various yield radio recurrence transmissions between the two flying machine. Specialist demonstrated that three Euler points required depicting the relative attitude [1]. Yet, computer researchers are regularly concerned about settings where agents are asset limited, in which case a few questions remain: does there exist a multi-agent system condition, given any announcement relative direction logic, either creates a short verification or infers that no short evidence exists? Comparably, it approaches whether the class of issues for which one can rapidly check a proposed arrangement is the same as the class of issues for which one can rapidly discover such an answer. This absence of advance may urge to look for new algorithmic procedures for identify or learning relative directions for multi-agent based system. As we will find in this thesis, specification of learning total six relative directions between two the agents and identify relative directions by using of two directions. 1.2 Research Objectives The main objective of this thesis is how agents are identify or learn relative directions among them or each one another by using our propose MARD algorithm where MARD algorithm performs from multi-agent viewpoints. This research report exhibits about three different algorithms. This research report exhibits about three different algorithms. The purpose of first algorithm is learn relative direction between two agents, purpose of second algorithm is identify a agent and also detect relative direction and purpose of third algorithm is how one agent learn another agent's relative direction by using of two direction value. All thought of these propose algorithm approach show with different issues are additionally talked about in chapter 3 (Proposed Algorithm Model). 1.3 Research Questions The thesis with titles showing that how relative directions are works between two agents from computer aspects. There have exactly some research questions and this will enable to understand some features of this thesis.
  • 12. 3 © 2017 by Daffodil International University  Why need relative direction and MARD algorithm?  What direction learning opportunities between two agents in computer science aspects?  How relative direction works in 3D environment and route direction?  Is it a new or modify approach?  Why use numerical value in the approach?  Why numerical values have been used for indicate relative direction? 1.4 Organization In Chapter 1, we revolve our work around the thesis and quickly present about foundation overview and purpose of this algorithmic research. In Chapter 2, we concentrate on a particular issue for building algorithm. In here we demonstrate which literature is audited. We at that point exhibit the evidence of our motivation, which utilizes relative direction of agents, before showing our own confirmation of a similar outcome and our outcome on endless groups of limit parts. This chapter briefly foundation or background of thesis, past research of relative direction, artificial intelligence based concept. In Chapter 3, we introduced about MARD algorithm. In here we have also displayed various 3D cases of agent and another different approach of MARD algorithm which will helps in machine learning sector. We also demonstrate multi-agent route directions which present interpretations mapping over the different area. In Chapter 4, we talk about our research methodology and exploration technique. This part shows research design and Algorithm engineering method for this algorithmic research. In Chapter 5, we analyse our algorithm, result. This part demonstrates explore examination and execution assessment with time and histogram. Different type of solving strategy analysis we exhibit in this section. In Chapter 6, we endeavor to settle on a choice about legitimacy of relative direction algorithm for multi-agent environment. This part indicates summary of this thesis and finishing. In Bibliography, we try to show proper references which help for complete this research. In Appendix, we incorporate audits of the fundamental many-sided quality classes, the portrayal of code, programming structure and evidences of some minor subtle elements specified in the body of the thesis.
  • 13. 4 © 2017 by Daffodil International University 1.5 Definitions MARD algorithm: Multi-agent relative direction (MARD) algorithm refers an algorithm strategy which identify or learning relative direction among the agents where one agent can be learning other agent’s relative directions. UAV: Unmanned aerial vehicle (UAV) normally acquainted as a drone. It is an airplane or aircraft without any human pilot on board. UAVs are a segment of an unmanned flying machine which incorporates a UAV, a ground-based controller, and an arrangement of interchanges between the two. The flight of UAVs may work with different degrees of self-rule: either according to remote control by a human administrator or independently by locally available computers. Multi-agent: A multi-agent framework or system is an automated manner made out of various associating smart operators inside a domain. Multi-agent system can be utilized to take care of issues that are troublesome or incomprehensible for an individual operator or a solid framework to solve. Intelligence may involve some methodic, utilitarian, procedural approach, algorithmic inquiry. Despite the fact that there is impressive cover, a multi-agent process is not generally the same as an agent-based model. Route Directions: A route is regularly part into a few fragments that are then verbalized. These verbalized directions can be guidelines to make a specific move, for example, "walk" or "turn", or portrayals of the map. Direction of Arrival: In signal technology writing, direction of arrival (DOA) means the course from which as a rule an engendering wave touches base at a point, where as a rule an arrangement of sensors are found. DOA discovers the direction in relative to the cluster where the sound source is found. OPRA: The Oriented Point Relation Algebra (OPRA) distributes for subjective spatial description and reasoning. OPRA is an introduction calculus math with movable granularity. OPRA depends on objects which are spoken to as oriented points. Oriented points are indicated as match of a point and a direction on the 2D-plane. 1.6 Motivation of Research The ―Direction Learning‖ has been a well known research point among specialists and researchers of arithmetic and computer science. With the objective of limiting the relative direction of the agent in estimating agent's relative direction problem. Arrangements of this issue can be connected to an extensive variety of improvement issues.
  • 14. 5 © 2017 by Daffodil International University Estimation of relative attitude between two UAV [1] is the latest research of direction learning calculation which can play numerous info various yield radio recurrence transmissions between the two aircraft. Most likely this recent research initially indicates multi-operator idea based relative state of attitude learning calculation. Our analysis is to attempt to give some calculation which takes after a few exercises of this recent research. Generally this recent research motivated us to doing our research. The structures we are searching for are just thickness varieties in the computation. Contrasting landmark based guidelines and relative direction directions on people on foot in a genuine city condition, measuring both goal and subjective achievement [4]. We find that at some choice focuses, multi agent based relative direction algorithm work better for route direction in mapping. We show a strategy that how our research gives better instruction in different computer science area.
  • 15. 6 © 2017 by Daffodil International University Chapter 2 Background and Literature Review The background gives a prologue to Multi-Agent Relative Direction (MARD) algorithm. It additionally presents some artificial intelligence (AI) perspective concept about this algorithm which is of intrigue while talking about multi-agent based relative direction learning process. This section describes the sequencing relative direction in any multi-agent based issue. Highlights of an algorithm are examined with a specific end goal to order the issue into sub- issues like object finding problem, structure issue. This Chapter additionally gives a presentation of the unmanned aerial vehicles and talks about the similarities and contrasts between past research and this thesis. 2.1 Relative Direction Fundamentals The most well-known relative directions are right, left, up, down, forward and backward. There are definite connections between the relative directions. Forward-backward, left-right, and up- down are three sets of integral relative directions. Relative directions are otherwise called egocentric coordinates. Relative directions can be helpful to individuals who are new to the area of cardinal directions. Since meanings of left and right in view of the geometry of the natural habitat are inconvenient. The importance of relative direction words is passed on through custom, cultural assimilation, training, and direct reference. One normal meaning of up and down utilizes gravity and the globe as an edge of reference. Up is then characterized as the other way of down. Another normal definition utilizes a human body, standing upright, as an edge of reference. Forward and backward might be characterized by alluding to a question's or individual's movement. Forward is characterized as the bearing in which the question is moving. Backward is then characterized as the other way to forward. Then again, forward might be the direction pointed by the onlooker's nose, characterizing backward as the heading from the nose to the sagittal fringe in the eyewitness skull. Concerning a ship forward would show the relative
  • 16. 7 © 2017 by Daffodil International University position of any protest lying toward the path the ship is pointing. In Figure 1 illustrates different relative directions from human perspective. Figure 1: Different types of relative directions. For symmetrical purpose, it is additionally important to characterize forward and backward regarding expected course. Many mass travel trains are constructed symmetrically with matched control corners, and meanings of forward, in reverse, left, and right are brief. 2.2 Multi-Agent System Environment Multi-agent structure is a computerized way made out of different partner keen administrators inside a space. Multi-specialist framework can be used to deal with issues that are troublesome or inconceivable for an individual agent or a strong system to solve. Multi-agent frameworks comprise of agents and their condition. Regularly multi-agent frameworks examine alludes to programming operators. Nonetheless, the specialists in a multi-agent system could similarly well
  • 17. 8 © 2017 by Daffodil International University be robots, people or human groups. A multi-specialist framework may contain consolidated human-agent groups. The inspiration for considering multi-agent system regularly originates from enthusiasm for programming or software agents. The material traverses teaches as assorted as software engineering (artificial intelligence, hypothesis, and distributed computing), financial matters (essentially microeconomics concept), research, scientific rationality, and phonetics. In understanding the determination made here, it is valuable to remember the accompanying algorithms [27]. Figure 2: Characteristics of Multi-agent System Environment. Multi-agent environment can include specialists making arrangements for a shared objective, an agent organizing the plans or arranging of others, or specialists refining their own particular designs while consulting over errands or assets. The theme likewise includes how agents can do this progressively while executing designs. Multi-agent booking varies from multi-agent planning a similar way arranging and vary in planning regularly the undertakings that should be performed are as of now chose, and by and by, planning tends to concentrate on algorithms for particular issue areas. 2.3 Previous Research and Work To represent the idea of multi-agent relative direction (MARD) algorithm previously formally characterizing every one of the parts that go into the several approaches, we give a genuinely
  • 18. 9 © 2017 by Daffodil International University casual exploration of one of its more eminent examples of success stories. All things considered, this subsection might be skipped if the peruser likes to jump straight into definitions.  Relative Attitude Estimation of UAV (Mahmood, Wallace and Jensen, 2017): They propose a renewed algorithm for evaluating the relative attitude between two unmanned aerial vehicles (UAV) in view of various multiple input and output radio recurrence transmissions between the two airplanes. The strategy can evaluate each of the three Euler points required to portray the relative disposition [1]. Figure 3: Tracking UAV and Handoff UAV in Relative Attitude Estimation of UAV’s paper In here the system show comprising of Tracking and Handoff UAVs and their comparing nearby arrange outlines and also the connection between the new organize outlines at the two UAVs framed from direction of entry estimates. Each UAV has its own particular nearby facilitate outline characterized by the unit-standard vectors where i ∈ {h, t} for Handoff and Tracking, individually. The analysts of this paper shows a novel calculation that consolidates direction of arrival (DOA) gauges with polarimetric multi-antenna apparatus channel appraisals to figure the relative attitude between two UAVs.  Relative Direction Change (Hahn, Bethge and Döllner, 2017): A topology-based metric for design solidness in treemaps of the Relative Direction Change (RDC) introduced metric considers the nearness and course of action of single shapes in a treemap, and takes into consideration a rotation-invariant portrayal of format alterations between two snapshots of a dataset delineated with treemaps [2].
  • 19. 10 © 2017 by Daffodil International University  Fast Contour-Tracing Algorithm Based on a Pixel-Following Method (Seo, Chae, Shim, Kim, Cheong and Han, 2016): The experiment introduces a novel contour-tracing algorithm for quick and exact contour following. This algorithm orders the sort of contour pixel, in light of its neighborhood design. At that point, it traces the following contour utilizing the past pixel’s model. The algorithm follows shape pixels along the clockwise direction from the present pixel, i.e., it consecutively seeks nearby dark pixels of the present pixel utilizing a relative directional request, for example, left, front-left, front, front-right, right, rear right and back. To decide the contour point, which might be a contour pixel, the tracer recognizes the power of its adjoining pixel Pr and the new absolute direction dr for Pr by utilizing relative direction data r ∈ { front, front − left, left, rear − left, rear, rear − right, right, r ∈ { front − right} [3].  Relative Directions Work Better Than Landmarks (Götze and Boye, 2015): People make broad utilization of landmarks while depicting the best approach to others and are more effective after directions that cover landmarks. It exhibit an examination contrasting landmark based guidelines and relative direction indication on people on foot in a genuine city environment, measuring both target and subjective achievement. Researchers find that at some choice focuses, relative direction guidelines work better. Specifically, guidelines that avoid a landmark and use just a relative direction like "left" or "right", appear to be favored at some decision focuses, especially those with a basic arrangement where streets meet at right edges [4].  Complexity of Reasoning with Relative Directions (Lee, 2014): In the case of reasoning upon relative directions can be performed in NP has been an open issue in subjective spatial reasoning. Effective reasoning with relative directions is fundamental, for instance, rule-compliant agent navigation [5]. In this research reasoning upon relative directions is ∃R-complete. As a result, reasoning with relative directions is not in NP, if not NP = ∃R, where ∃R is a many-sided complexity class.  Effective Reasoning about Relative Directions (Lee, Renz and Wolter, 2013): Eyewitnesses commonly depict environment by indicating the relative directions in which they see different articles or other people from their perspective [6]. They demonstrate that reasoning in StarVars is in NP and present the primary algorithm that enables to viably coordinate relative direction data from various observers. They built up a spatial portrayal, StarVars, which increases cardinal direction relations to illustrate to relative directional information.  Relative Identification and Direction for Wireless Network (Weng and Lai, 2013): A less intricate, more productive routing algorithm called as relative identification and direction- based sensor routing (RIDSR) algorithm [7]. RIDSR influences sensor hubs to set up more
  • 20. 11 © 2017 by Daffodil International University dependable and vitality effective routing path for information transmission. RDSR calculation not just tackles the routing loop problem inside the algorithm yet in addition encourages the immediate choice of a shorter way for routing by the sensor node. Moreover, it saves energy and broadens the existence of the sensor hubs.  Relative Direction of Oriented Points (Mossakowski and Moratz, 2012): An imperative problem in qualitative spatial reasoning is the portrayal of relative directions. In this paper, basic geometric tenets that empower reasoning about the relative direction into oriented points [8]. This structure arranged a oriented point algebra OPRAm, has a versatile granularity m. In this paper a basic algorithm for figuring the OPRAm synthesis tables and demonstrates its rightness.  Verbal Navigational Directions in Relative Frames (Mossakowski and Moratz, 2008): This examination inspected how people utilize verbal route directions conferred in relative and absolute edges of reference in genuine route, especially contrasts or likenesses in cognitional load postured by the two frames of reference [9]. This instruction took a gander at how people utilize verbal route directions offered in two sorts of frames of reference, relative and absolute, in genuine route. Specifically, inspected the distinctions or likenesses in the trouble of utilizing and preparing data given in favored and non preferred casings of reference, and whether individuals could adjust to or switch between the two frames of reference.  Triangular Multiple Flow Direction Algorithm (Seibert and McGlynn, 2007): Gridded digital elevation data (DEMs) frequently alluded to as DEMs, are a standout amongst the most broadly accessible types of natural information. Here a give an account of a stream routing algorithm and contrast it with three regular classes of calculations at present in across the board utilize. The upside of the algorithm is that unrealistic dispersion on planar or curved hillslopes is dodged, while numerous flow directions are permitted on raised hillslopes. The steepest directions point fairly to one left and right. Be that as it may, since there must be one outflow direction in the algorithm, just a single of these two directions gets territory, while the two directions ought to get region [10].  An ant colony optimisation algorithm for the 2D and 3D (Shmygelska and Hoos, 2005): The protein folding issue is a principal issue in computerized molecular science and biochemical physics. In this work, research demonstrated that ant colony optimisation (ACO) can be connected in a somewhat straight-forward path to the 2D and 3D HP Protein Folding Problems. Despite the fact that our ACO-HPPFP-3 calculation depends on exceptionally straightforward structure parts (single relative directions) and a basic backup neighborhood seek strategy [11].
  • 21. 12 © 2017 by Daffodil International University  Relative Direction as a Binary Relation (Moratz, 2006): A central issue in robotics is the representation of relative orientation. This paper introduces a new calculus about oriented points which has a scalable granularity [12]. In this calculus, named OPRA, simple rules can generate the minimal composition. Furthermore, the algebraic closure for a set of OPRA statements is sufficient to solve knowledge integration tasks in robotics.  Plaid Motion Coherence on Component Grating Directions (Kim and Wilson, 1993): A few element motion directions were created little to vast angular contrasts. In here a confirmation obviously demonstrates that coherence of movement for 2D designs in various spatial scales depends fundamentally on the relative direction of motion of component gratings and is moderately autonomous of difference and speed. It is likewise free of the SF distinction between two parts as long as the proportion is more prominent than around 3:1 [13]. It would appear to be environmentally more conceivable for the visual system to decide inflexibility of movement in view of the relative directions of neighborhood motion vectors.  Relative Directions between Landmarks (Albert, Rensink and Beusmans, 1999): This examination presents two tests that inspect how people learn relative directions between landmarks in a desktop virtual condition. Subjects were introduced preview pictures of various virtual environments containing recognizing points of landmarks and a road network. The introduction of each virtual environment, subjects were given a relative direction test [14]. The relative direction test included demonstrating the direction of concealed landmarks from various vantage focuses in the environment.  Robot kinematics (1998): Kinematics is the connections between the positions, speeds, and increasing velocities of the connections of amanipulator, where a controller is an arm, finger, or leg. This exploration characterizing the arrangement as far as elbow up or down, left or right handed [15]. A matrix portrays the change from the base to the hand of the controller, a succession called the forward kinematic change of the manipulator.  Maps and Relative Direction (Foster, 2016): It's quite basic to portray direction in connection to area on a map. Go up that path, down here, or over yonder. Up, down, and over are relative directions given from a perspective, regularly physical topographic change [16]. Up stream, down the slope, and over to the lake. The words up and down can be held in respect to gravity. Unless people are alluding to up and down in connection to geology, or in relative to a specific area.  Relative Main Line layout algorithm: The Relative Main Line format algorithm works from traits that enable the calculation to distinguish the straight lines that is, the principle lines and root schematic nodes from which
  • 22. 13 © 2017 by Daffodil International University those straight lines begin. Root schematic nodes can be set utilizing the Set Schematic Root tool [17]. Set Schematic Root to determine the beginning stages of the straight lines. The algorithm initially looks nodes to observe candidates to be the root node that is, node associated with a single link that can be considered as the beginning point for a straight line.  Dragonfly Algorithm for Solving Multi-objective Problems (Mirjalili, 2016): Dragonfly calculation is a fiction swarm intelligence streamlining technics. The progression vector of this algorithm demonstrates the direction of the development of the dragonflies and characterized as takes after: Xr + 1 = ( sSi + aAi + cCi + fFi + eEi ) + wXt where s demonstrates the division weight, Si shows the partition of the i-th individual, a is the arrangement weight, Ai is the arrangement of i-th singular, c shows the attachment weight, Ci is the union of the i-th singular, f is the sustenance factor, Fi is the nourishment wellspring of the i- th singular, e is the foe factor, Ei is the position of foe of the i-th singular, w is the dormancy weight, and t is the cycle counter [18].  DOA Estimation of Animal Vocalizations (Hedley, Huang and Yao, 2017): A recording system built from two Wildlife Acoustics SM3 recording units that can calculate the direction-of-arrival (DOA) of an approaching signal with high precision [20]. Signal processing algorithms, similar to the MUSIC algorithm utilized their analysis, they utilize these stage contrasts to decide the angle from which each sound arrived (α and β for the red and blue winged creatures, individually, relative to a self- assertive mention angle, marked 0). The system utilizes four all the while recording receivers to evaluate the direction from which a sound arrived, in view of the stage contrasts of the approaching sound waves at the microphones.  Discrete-State-Based Vision Navigation Control Algorithm (Wei, 2015): To set out a principled dialog of the exactness and productivity of navigation algorithms, entirely quantitative meanings of following error, actuator impact, and time proficiency are built up. The navigation algorithm would control the robot following the particular direction [21]. In the wake of characterizing the relative angle between desired velocity vector and the real velocity vector as course blunder signified by e. Most extreme Steering Angle max. This value is the practical steering angle pushing forward in a direction. That implies the robot can move forward with the direction scope of [−max, max]. Negative esteem implies the robot turns right while the positive value implies it turns left with respect to the robot. 2.4 Left-Right Confusion Left-right confusion is the inconvenience a few people have in recognizing the distinction between the headings left and right. These individuals can as a rule typically perform every day
  • 23. 14 © 2017 by Daffodil International University exercises, for example, driving as indicated by signs and exploring as indicated by a map, however will regularly go astray when advised to turn left or right and may experience issues performing activities that require exact comprehension of directional orders. 2.4.1 Research about Left-Right Confusion Challenges in left–right discrimination (LRD) are usually experienced in regular day to day existence circumstances. An examination demonstrates that the neurocognitive components of left– right separation and the particular part of left precise gyrus [29]. In later an examination surveyed the connection between self-appraised right–left confusability and execution on the Money Road-Map Test (MRMT). Eighty-six understudies appraised right–left subjective confusability utilizing a poll, and afterward finished the Money Road-Map Test. Another examination researches the connection between the view of bilateral symmetry and left- right bewilderment in neurologically in place grown-ups by utilizing tachistoscopic introduction of boosts and a decision response time technique. Scientists found a little yet predictable pattern toward snappier symmetry judgments in left-right distracted subjects. The legitimacy of such self-report measures in foreseeing real execution on right-left segregation undertakings is addressed since the outcomes, in any event as a component of handedness, relied upon the inquiry [30, 31, 32]. 2.4.2 Artificial Intelligence Perspective Artificial intelligence (AI) is insight displayed by machines, instead of people or different creatures (natural intelligence). In software engineering, the field of AI investigate characterizes itself as the investigation of intelligent agents are any gadget that sees its condition and takes activities that expand its risk of achievement at some objective. Conversationally, the expression computerized intelligence is connected when a machine impersonates subjective capacities that humans connect with other human personalities, for example, "learning" and "critical thinking". Learning or identify direction is an exploration range in software engineering and computer science, with territories, for example, unraveling, trouble and era. One of the critical choice rule for picking MARD algorithm for this proposal have hence been the algorithm hidden strategy for crossing the hunt space, for this situation deterministic and stochastic strategies. Possible left-right confusion with artificial intelligence perspective show in Figure 4, 5, 6 and 7, each agent has its own local relative directions. Where Agent 1 = R, Agent 2 = A and Mobile phone is a object for find mobile phone object. Primary objective of these figures are to find the object and defining the left and right direction where relative directions (right and left) of two agents show from different perspective. The process for achieving this purpose is as follows:
  • 24. 15 © 2017 by Daffodil International University Figure 4: Relative directions (right and left) of two agents from different perspective. In here a mobile phone is a example of object. 1) In Figure 4, there are two agent respectivelty agent A and agent R and their right and left directions different from different perspective. Mobile phone is an example of object for understanding left-right confusion in human and artificial intelligence both perpective. The Mobile object is located on its left side of agent R and on the right side of agent A. 2) In Figure 5, agent A ordered agent R to find the mobile object on the right direction. When agent R searches the mobile phone object in his right direction, he could not find any mobile phone, because relative directions between two agents are different. It means that agent A’s right direction is not equal to agent R’s right direction. Relative directions between two agents are different. 3) In Figure 6, agent A think and calculated that his right direction is equal to agent R’s left direction. Agent A also calculated that his left direction is equal to agent R’s right direction. At last agent A learn that Agent R’s relative directions. 4) Lastly in Figure 7, agent A ordered agent R to find a mobile phone on agent R’s left direction and agent R find the mobile phone object.
  • 25. 16 © 2017 by Daffodil International University Figure 5: Agent A ordered agent R to find the mobile object on the right direction but agent R could not find any mobile phone on his right direction, because relative directions between two agents are different. Figure 6: Agent A calculated that his right direction is equal to agent R’s left direction and left direction is equal to agent R’s right direction.
  • 26. 17 © 2017 by Daffodil International University Figure 7: Agent A ordered agent R to find a mobile phone on agent R’s left direction. At last agent R find the mobile object. Basically our MARD algorithm will work like this way. As a result one agent easily can understand anodher agen’s left-right or relative directions. Through this concept we can devlope our new multi-agent relative direction (MARD) algorithm and with the help of this algorithm we can solve left-right confusion problem in compuer or artificial intelligence perspective.
  • 27. 18 © 2017 by Daffodil International University Chapter 3 Proposed Algorithm Model To represent the idea of multi-agent relative direction (MARD) algorithm previously formally characterizing every one of the parts that go into the approach, we give a genuinely casual exploration of one of its more eminent examples of success stories. All things considered, this subsection might be skipped if the peruser likes to jump straight into definitions. In this chapter, we present the fundamental theory of MARD algorithm. We work through an illustrative case in this segment before formalizing the model. At that point we talk about the characterizing numerical properties of MARD algorithm and give a unique re-detailing of these properties that will demonstrate basic to demonstrating our primary hypothesis. At long last, we give a programming structure of MARD algorithm as far as spinor assortments and drawing ideas. 3.1 Induction This exploration section introduces propose algorithm for figuring the relative directions between two agents, where how one agent can take in another agent's relative directions through in computer or programming perspective. We display an examination of relative direction. In this section demonstrate our experiment design, realistic input, mathematical concept and structure of MARD algorithm. We present a novel subjective depiction and portrayal this MARD algorithm can tackle these issues for dealing with and perceiving relative directions, its executed work strategy or computation puts aside greater open door for result. In order to MARD algorithm experiment, we focus on Algorithm Engineering technic. Algorithm engineering revolves around the outline, examination, analysis, implementation, optimization and exploratory appraisal of algorithms. One significant though all too every now and again overlooked issue when driving analyses in Computer Science is to ensure MARD algorithm.
  • 28. 19 © 2017 by Daffodil International University 3.2 Tracking Agent and Handoff Agent With past work focusing on attitude estimation, an algorithm using Tracking UAV and Handoff UAV for estimating relative attitude between two unmanned aerial vehicles (UAV) [1]. In the MARD algorithm we use ―Tracking Agent‖ and ―Handoff Agent‖ (see Figure 8). We consider in the algorithm where one tracking agent is following an objective on the ground, and it is wanted to have a moment handoff agent have the capacity to learning the relative direction of the tracking agent. 3.2.1 Direction Points We use total six riletive directions in our algorithm. Each agent has 6 direction points for handoff agent direction points are a, b, c, d, e, f and for tracking agent direction points are same level but reverse for that direction points are a1, b1, c1, d1, e1, f1. Firstly handoff agent knows his relative directions but in the beginning handoff agent not knows about tracking agent’s relative directions, which has been displayed in Figure 8. Figure 8: Direction points of Handoff Agent and Tracking Agent. Handoff Agent Tracking Agent RightLeft Up Down Forward Backward c1 b1 a1 d1 e1 f1 ab c d e f
  • 29. 20 © 2017 by Daffodil International University 3.2.2 Various 3D Aspects 3D PC illustrations are regularly alluded to as 3D models. The view of relative directions seems from different 3D aspects. In Figure 9 we show 24 different cases of relative direction aspects of tracking agent perspective. Figure 9: Different cases of tracking agent in 3D aspect. Right a1 Left Up Down Forward Backward b1 c1 d1 e1 f1 Case 1 Case 6 Case 11 Case 16 Case 21 Up Down LeftRight ForwardBackward Case 2 Down Up Left Right Backward Case 3 Forward Up Down Backward RightLeft Forward Case 4 Up Down Forward Backward Left Right Case 5 Up Down Forward Backward Left Right Down Up BackwardLeft Right Case 7 Backward Forward Down Up Left Right Case 8 Forward Backward Up Down Right Left Case 9 Backward Forward Up Down Left Right Case 10 Up Down Forward Backward Left Right Backward Forward Left Right Up Down Forward Case 12 Backward Forward Right Down Up Left Right Forward Backward Case 13 Down Up Backward Forward Left Right Up Down Right Left Case 15Case 14 Up Down Forward Backward Left Left Forward Backward LeftRight Forward Backward Forward Backward Down Up Left Right Up Down Backward Forward Left Right Up Up Down DownRight Case 17 Case 18 Case 19 Case 24Case 22 Case 23 Case 20 Up DownLeft Right Forward Backward Up DownRight Left Forward Backward Backward Forward Up Down Right Left Up Down Forward Backward Left Right c1 c1 c1 c1 c1 c1 c1 c1 c1 c1 c1 c1 c1 c1 c1 f1 c1 c1 c1 c1 c1 c1 c1 c1 b1 b1 b1 b1 b1 b1 b1 b1 b1 b1 b1 b1 b1 b1 b1 b1 b1 b1 b1 b1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 e1 e1 e1 e1 e1 e1e1 e1 e1 e1 e1 e1 e1 e1 e1 e1 e1 e1 e1 e1 e1 e1 f1f1 f1 f1 f1 f1 f1 f1 b1 b1 b1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 d1 d1 d1 f1 d1 d1 d1 d1 d1 d1 d1 d1 d1 d1 d1 d1 d1 d1 d1 d1d1 d1d1
  • 30. 21 © 2017 by Daffodil International University From Figure 8 and Figure 9 we can say that relative direction of tracking agent always changeable but direction points of tracking agent are constant. Here we find total twenty four relative direction movement. So we can say that, Total relative direction movement (M) = Number of relative direction (n) * 4 Or, M = 6*4 = 24 3.2.3 Realistic Inputs At first handoff agent knows his relative directions but handoff agent not knows about tracking agent’s relative directions. Moreover handoff agent compare with tracking agent’s relative directions by tracking agent’s directions points. For MARD algorithm development we use some niumerical value for identify relative directions. Every relative direction point has own value which depends on the relative directions such as 0 for Right, 1 for Left, 2 for Up, 3 for Down, 4 for Forward and 5 for Backward, which have been shown in Table 1. We have used these values because the reason in our proposed algorithm we have used two array agent which contains these direction variables. Relative Direction Direction Point Value (For relative direction) Right 0 Left 1 Up 2 Down 3 Forward 4 Backward 5 Table 1: Relative directions and their numerical values.
  • 31. 22 © 2017 by Daffodil International University 3.3 Mathematical Exploration In this area, we will numerically define the MARD algorithm with every one of the imperatives specified. The main navigating auto in the computer science serves an arrangement of benefit arrange from various wellsprings of the algorithm design. The procedure for mathematical computation of this algorithm is as follow: Ti  ( Hj ) In here T refers tracking agent, where T = {Right, Left, Up, Down, Forward, Backward} and H refers handoff agent, where H = {right, left, up, down, forward, backward}. Besides i is the set of index T’s relative direction, where i = 0, 1, 2, 3, 4, 5 and j is the set of index H’s relative direction, where j = 0, 1, 2, 3, 4, 5. For i, handoff agent H’s j relative direction is assigned into tracking-human T’s i relative direction. 3.4 Structure of Algorithm This investigation part presents propose algorithm for figuring the relative directions between two agents, where how one agent's can learn in another agent's relative directions through in programming point of view. In this segment exhibit our analysis plan, practical information, programming idea and structure of MARD algorithm for application. We very briefly develop MARD algorithm. The process of MARD algorithm for programming or application perspective goal is as per the following: 1) Step1: Creat a HandoffAgent class  An array Direction = [right, left, up, down, forward, backward].  Direction point variables of handoff agent are a, b, c, d, e and f.  Points are containing direction values (see Figure 8 and Table 1). So that a = 0, b = 1, c = 2, d = 3, e = 4 and f = 5.  Direction point variables are involving relative directions. So that, right = a, left = b, c = up, d = down, e = forward and f = backward (see Figure 8 and Table 1). 2) Step2: Creat a TrackingAgent class  Direction = [Right, Left, Up, Down, Forward, Backward].  Direction point variables of tracking agent are a1, b1, c1, d1, e1 and f1. Direction points contain different direction value. These values are 0, 1, 2, 3, 4 or 5 (see Figure 8, Figure 9 and Table 1).
  • 32. 23 © 2017 by Daffodil International University 3) Step3: Creat a Main class, main function and object  handoffAgent is a object of HandoffAgent class  trackingAgent is a object of HandoffAgent class 4) Step4: Creat a Loop  A variable i = 0 to 5.  The condition of the loop is i <= 5.  After every loop value of i will increase (i++).  If complete all possible loops then go to step 6 for end process. 5) Step5: Declare j and find relative directions  if i == a1 then, j = 0; // j = 0 means right //  else if i == b1 then, j = 1; // j = 1 means left //  else if i == c1 then, j = 2; // j = 2 means up //  else if i == d1 then, j = 3; // j = 3 means down //  else if i == e1 then, j = 4; // j = 4 means forward //  else if i == f1 then, j = 5; // j = 5 means backward //  trackingAgent.Direction[ i ] = handoffAgent[ j ];  Print ―Tracking Agent’s‖ + i direction name == ―Hanoff Agent’s‖ + j direction name;  Go to step 4 for complete loop. 6) Step6: End  End execution of algorithm. We present a pseudocode which in view of objects oriented programming (OOP) formation that recognizes relative directions and correspondences between two agents. The pseudocode of the MARD algorithm is given below:
  • 33. 24 © 2017 by Daffodil International University Algorithm 1: Multi-Agent Relative Direction Algorithm 1: HandoffAgent class { 2: Direction = [ right, left, up, down, forward, backward ]; 3: directions points: a; b; c; d; e; f; 4: //points are contains values (see Figure 8 and Table 1)// 5: right = a; left = b; up = c; down = d; forward = e; backward = f1; 6: } 7: 8: TrackingAgent class { 9: Direction = [ Right, Left, Up, Down, Forward, Backward ]; 10: directions points: a1; b1; c1; d1; e1; f1; 11: //points are contains values (see Figure 8, Figure 9 and Table 1)// 12: } 13: 14: Main class { 15: void function main( ) { 16: HandoffAgent handoffAgent = new HandoffAgent(); 17: TrackingAgent trackingAgen = new TrackingAgent(); 18: 19: int j; 20: 21: for i = 0 to 5 22: if i == a1 then, j = 0; 23: else if i == b1 then, j = 1; 24: else if i == c1 then, j = 2; 25: else if i == d1 then, j = 3; 26: else if i == e1 then, j = 4; 27: else if i == f1 then, j = 5; 28: 29: trackingAgent.Direction[ i ] = handoffAgent.Direction[ j ]; 30: 31: end for 32: } 33: }
  • 34. 25 © 2017 by Daffodil International University 3.5 Identify Relative Direction Identify tracking agent and its relative directions are very important aspect. Figure 10 represents a simple concept of tracking agent's relative directions identification. Figure 10: Tracking agent's direction identification. (a) Color location tracking agent’s direction point. (b) Shape are indicates different relative directions. (c) Handoff agent takes an image for identifying tracking agent’s relative directions. We have shown a sample pseudocode of this idea for identify the tracking agent which has been shown in Figure 10. At first handoff agent’s camera takes an image of tracking agent. There are six color locations to find out tracking-human’s direction points. An array Shape contains different shapes which are Circle, Diamond, Triangle, Hexagon, Rectangle and Pentagon. These shapes refers different relative directions for track a tracking agent’s relative directions. Variable k for completing the loop and also has been used for identifying the relative directions among the directional points (a1, b1, c1, d1, e1 and f1) of tracking agent. If specific color is equal to a specific Shape, then direction points of tracking agent will contain a relative direction value which is k (see table 1). The structure of identify tracking agent’s relative direction in programming perspective goal is as per the following: Handoff Agent Camera Image Red Yellow Green Orange Blue Violet d1 c1 e1 f1 a1b1 (a) Right Left Up Down Forward Backward Tracking Agent (b) (c)
  • 35. 26 © 2017 by Daffodil International University Algorithm 2: Direction Identification Algorithm 1: Camera Image of trackingAgent; 2: directions points: a1; b1; c1; d1; e1; f1; 3: location of direction points in image: Green; Blue; Red; Yellow; Orange; Violet; 4: Shape = [Circle, Triangle, Rectangle, Diamond, Pentagon, Hexagon]; 5: for k := 0 to 5 6: if Green == Shape[k] then, a1 = k ; 7: else if Blue == Shape[k] then, b1 = k ; 8: else if Red == Shape[k] then, c1 = k ; 9: else if Yellow == Shape[k] then, d1 = k ; 10: else if Orange == Shape[k] then, e1 = k ; 11: else if Violet == Shape[k] then, f1 = k ; 12: end for 3.6 Machine Learning Aspect Machine learning is an area of computer science and software engineering that gives PCs the capacity to learn without being expressly programmed. Machine learning explores the study and construction of algorithms that can learn from and make predictions on data, such algorithms overcome following strictly static program instructions by making data-driven predictions or decisions, through building a model from sample inputs. The pseudocodes of Algorithm 2 can be used for various purposes of machine learning for identify multi-agent based relative directions. In the next part we represent an alternative algorithm approach of multi-agent based relative direction identification algorithm, where realistic inputs can be worked better at machine learning in future. 3.7 Alternative MARD Algorithm Approach In this section, we give another alternative algorithm approach of MARD algorithm. Firstly, if we look deeply in Figure 9 we can understand that if we find ―Forward‖ and ―Up‖ direction then we can easily estimate Left, Right, Down and Backward directions. In Figure 9 there have total 24 case befor go to the alternative MARD algorithm we want to show the condition of these cases. The condition for 3D cases is summarized in Table 2. In here i = 3, 5 indicates that the value of Up and Forward are depends on i.
  • 36. 27 © 2017 by Daffodil International University Case Tracking agent direction Equal to Hanoff Agent direction And Tracking agent direction Equal to Hanoff Agent direction 1 Up == up && Forward == forward 2 Up == up && Forward == backward 3 Up == down && Forward == forward 4 Up == down && Forward == backward 5 Up == left && Forward == up 6 Up == left && Forward == down 7 Up == right && Forward == up 8 Up == right && Forward == down 9 Up == forward && Forward == up 10 Up == forward && Forward == down 11 Up == backward && Forward == up 12 Up == backward && Forward == down 13 Up == right && Forward == forward 14 Up == right && Forward == backward 15 Up == left && Forward == forward 16 Up == left && Forward == backward 17 Up == down && Forward == left 18 Up == down && Forward == right 19 Up == up && Forward == left 20 Up == up && Forward == right 21 Up == backward && Forward == right 22 Up == backward && Forward == left 23 Up == forward && Forward == right 24 Up == forward && Forward == left Table 2: Condition of Cases. There are total 12 orders, two particular orders for one individual case. It means that, Total order = total number of direction 6*2 = 12. These orders show the left and right direction of these cases. Table 3 shows the orders.
  • 37. 28 © 2017 by Daffodil International University Order Tracking agent direction Assignment (=) Handoff agent direction 1 Right = Right 2 Left = Left 3 Right = Left 4 Left = Right 5 Right = Backward 6 Left = Forward 7 Right = Forward 8 Left = Backward 9 Right = Down 10 Left = Up 11 Right = Up 12 Left = Down Table 3: Order of Cases. Now, we have proposed an alternative MARD algorithm approach that is based on only two directions (Up and Forward). The algorithm has twenty four cases for learning all possible directions of tracking agent which based on three dimension space. The process of alternative MARD algorithm approach for programming or application perspective goal is as per the following: 1) Step1: Creat a HandoffAgent class  Direction = [ right, left, up, down, forward, backward ].  Direction point variables of handoff agent are a, b, c, d, e and f.  Points are containing direction values (see Figure 8 and Table 1). So that a = 0, b = 1, c = 2, d = 3, e = 4 and f = 5.  Direction point variables are involving relative directions. So that, right = a, left = b, c = up, d = down, e = forward and f = backward (see Figure 8 and Table 1). 2) Step2: Creat a TrackingAgent class  Direction = [Right, Left, Up, Down, Forward, Backward].  Direction point variables of tracking agent are a1, b1, c1, d1, e1 and f1. Direction points contain different direction value. These values are 0, 1, 2, 3, 4 or 5 (see Figure 8, Figure 9 and Table 1).
  • 38. 29 © 2017 by Daffodil International University 3) Step2: Creat a Main class, main function and object  handoffAgent is a object of HandoffAgent class  trackingAgent is a object of HandoffAgent class 4) Step4: Creat a Loop  A variable i = 2 and 4.  Loop condition is i < 5.  The limitation of loop is i = i + 2.  If complete all possible loops thent go to step 7. 5) Step5: Declare j and Find Up and Forward Direction  if i == a1 then, j = 0; // j = 0 means right //  else if i == b1 then, j = 1; // j = 1 means left //  else if i == c1 then, j = 2; // j = 2 means up //  else if i == d1 then, j = 3; // j = 3 means down //  else if i == e1 then, j = 4; // j = 4 means forward //  else if i == f1 then, j = 5; // j = 5 means backward //  trackingAgent.Direction[ i ] = handoffAgent[ j ]; // for learn Up and Forward direction //  trackingAgent[ i+1] = handoffAgent[ j +1 ]; // for learn Down and Backward direction //  Print ―Tracking Agent’s‖ + i direction name == ―Hanoff Agent’s‖ + j direction name;  Go to step 3 for complete all loops. 6) Step6: Find Right and Left Direction  if Case 1 then, Order 1 and Order 2.  else if Case 2 then, Order 3 and Order 4.  else if Case 3 then, Order 3 and Order 4.  else if Case 4 then, Order 1 and Order 2.  else if Case 5 then, Order 5 and Order 6.  else if Case 6 then, Order 7 and Order 8.  else if Case 7 then, Order 7 and Order 8.  else if Case 8 then, Order 5 and Order 6.  else if Case 9 then, Order 3 and Order 4.  else if Case 10 then, Order 1 and Order 2.  else if Case 11 then, Order 1 and Order 2.  else if Case 12 then, Order 3 and Order 4.  else if Case 13 then, Order 9 and Order 10.  else if Case 14 then, Order 11 and Order 12.  else if Case 15 then, Order 11 and Order 12.
  • 39. 30 © 2017 by Daffodil International University  else if Case 16 then, Order 9 and Order 10.  else if Case 17 then, Order 5 and Order 6.  else if Case 18 then, Order 7and Order 8.  else if Case 19 then, Order 7 and Order 8.  else if Case 20 then, Order 5 and Order 6.  else if Case 21 then, Order 9 and Order 10.  else if Case 22 then, Order 11 and Order 12.  else if Case 23 then, Order 11 and Order 12.  else if Case 24 then, Order 9 and Order 10. 7) Step7: End Process  After complete all steps then end process. The learning approach of the relative direction for three dimension modeling purpose is look complicated but is concept is easy to identify relative direction when handoff agent can learn tracking agent's all relative directions by using Up and Forward direction of tracking agent. The process of Alternative MARD algorithm for programming or application perspective goal is as per the following: Algorithm 3: Alternative MARD Algorithm 1: HandoffAgent class { 2: Direction = [ right, left, up, down, forward, backward ]; 3: directions points: a; b; c; d; e; f; 4: //points are contains values (see Figure 8 and Table 1)// 5: right = a; left = b; up = c; down = d; forward = e; backward = f1; 6: } 7: TrackingAgent class { 8: Direction = [ Right, Left, Up, Down, Forward, Backward ]; 9: directions points: a1; b1; c1; d1; e1; f1; 10: //points are contains values (see Figure 8, Figure 9 and Table 1)// 11: } 12: Main class { 13: void function main( ) { 14: HandoffAgent handoffAgent = new HandoffAgent(); 15: TrackingAgent trackingAgen = new TrackingAgent(); 16: int j; 17: for i = 2 and 4 18: if i == a1 then, j = 0; 19: else if i == b1 then, j = 1; 20: else if i == c1 then, j = 2;
  • 40. 31 © 2017 by Daffodil International University 21: else if i == d1 then, j = 3; 22: else if i == e1 then, j = 4; 23: else if i == f1 then, j = 5; 24: trackingAgent[ i ] = handoffAgent[ j ]; 25: // for learn Up and Forward direction // 26: trackingAgent[ i+1] = handoffAgent[ j +1 ]; 27: // for learn Down and Backward direction // 28: end for 29: 30: if Case 1 then, Order 1 and Order 2. 31: else if Case 2 then, Order 3 and Order 4. 32: else if Case 3 then, Order 3 and Order 4. 33: else if Case 4 then, Order 1 and Order 2. 34: else if Case 5 then, Order 5 and Order 6. 35: else if Case 6 then, Order 7 and Order 8. 36: else if Case 7 then, Order 7 and Order 8. 37: else if Case 8 then, Order 5 and Order 6. 38: else if Case 9 then, Order 3 and Order 4. 39: else if Case 10 then, Order 1 and Order 2. 40: else if Case 11 then, Order 1 and Order 2. 41: else if Case 12 then, Order 3 and Order 4. 42: else if Case 13 then, Order 9 and Order 10. 43: else if Case 14 then, Order 11 and Order 12. 44: else if Case 15 then, Order 11 and Order 12. 45: else if Case 16 then, Order 9 and Order 10. 46: else if Case 17 then, Order 5 and Order 6. 47: else if Case 18 then, Order 7and Order 8. 48: else if Case 19 then, Order 7 and Order 8. 49: else if Case 20 then, Order 5 and Order 6. 50: else if Case 21 then, Order 9 and Order 10. 51: else if Case 22 then, Order 11 and Order 12. 52: else if Case 23 then, Order 11 and Order 12. 53: else if Case 24 then, Order 9 and Order 10. 3.8 Multi-Agent Route Direction A route direction is ordinarily part into a few sections that are then verbalized. These verbalized bearings can be guidelines to make a specific move, for example, "move" or "turn", or portrayals of the earth like "There is a market to one left side". The request of the headings ought to mirror
  • 41. 32 © 2017 by Daffodil International University the straight request in which the course is crossed. An investigation contrasting landmark point based guidelines and relative directions on people on foot in a genuine city condition, measuring both goal and subjective achievement and find that at some choice focuses, relative direction indication work better [4]. In Figure 11 a mapping image illustrate that there have two agents who are stand in different area. Handoff agent stayed in area 1 and tracking agent stayed in area 2. Tracking want to find ―School‖. Handoff agent knows about the school location. So hanoff agent helps tracking agent to find school location by relative directions. MARD algorithm can solve this type of issue. That MARD algorithm can assume an uncommon part in the correspondence of route direction has been exhibited in a few investigations. Multi-agent based computation is a way to recognize critical focuses along the course where turning moves should be made or could be taken and in addition to find the start and the finish of the route. MARD algorithm additionally has a part in the clear impact of route directions, and to affirm that the devotee has accurately executed a turn. A image for multi-agent based route directions where use relative directions is shown in Figure 11. Figure 11: Multi-agent based route direction in mapping. Area 1 Area 2 School Up Down Foward Backward Right Left RightLeft Up Down Backward Foward Tracking Agent Handoff Agent
  • 42. 33 © 2017 by Daffodil International University Chapter 4 Methodology This part concentrates on how process of choosing algorithm engineering method. This part gives an outline of the measurable investigation which was performed on this algorithmic research. This additionally incorporates what computational restrictions were available and how this affected the outcomes. Methodology is the efficient, theoretical examination of the strategies connected to a field of study. It involves the theoretical investigation of the methods and standards related with a branch of information. 4.1 Algorithm Engineering Method Algorithm engineering (AE) is a usual methodology for algorithmic research. It centers on the analysis, design, examination, execution, enhancement, implementation and test assessment of algorithm. In our paper we used Algorithm engineering methodology [22, 24, 25, 26]. First, we introduce the relative direction concept on multi-agent best system structure. At that point we designed the MARD algorithm. The primary part of this algorithmic research is set realistic models and set real inputes which helps test the algorithm on various relative direction cases. We use some numerical values for identify relative directions, which helps to demonstrate a realistic model for develop the MARD algorithm. In this manner, we designed the algorithm. Before design MARD algorithm we create "Tracking agent" and "Handoff agent", after that we are keen on an effective algorithm. First algorithm design depends on proper inputs, mathematical exploration and human perspective concept. Second algorithm structure construct on image processing aspects. At last alternative algorithm approach has been designed on machine learning aspects and two direction based identification.
  • 43. 34 © 2017 by Daffodil International University This algorithm was analyzed by pseudocode and theoretical knowledge of mathematics. It concerns the result of the algorithm configuration stage in the picked programming dialect. Experiment of algorithm by programming in application level which shows different results and various preferences. All execution processes were measured and promote analysis. Since there may be varieties in execution and a component of randomness in algorithm implementation, various tests were performed on each relative direction. Figure 12: Algorithm Engineering methodology cycle. An essential objective of AE in our research is additionally to accelerate the exchange of algorithmic information into applications. In this experiment period of algorithmic research picking the correct issue cases for testing is continually testing until get correct output. Applications Real Inputs Realistic Model Design ExperimentsAnalysis Implementation
  • 44. 35 © 2017 by Daffodil International University 4.2 Experiment Control Algorithm Engineering is constantly determined by certifiable applications. The application situation decides the equipment which must be demonstrated generally sensibly. The consequences of an experimentation stage may then later on request an update of the demonstrating stage, in light of the fact that the picked models are not appropriate. Once in a while an examination of the picked model would already be able to uncover its deficiency. Apart from delimiting the negative effects of round-off by controlling the accumulation of numerical errors, numerical analysis also helps us in assessing their actual magnitude at runtime. In doing so it allows us to check at runtime whether computed solutions are reliable or not. This is an important part of the basis of reliable computing.
  • 45. 36 © 2017 by Daffodil International University Chapter 5 Results and Analysis In this chapter various outcomes are given together an exchange about how the outcomes could be translated. This section is dedicated to exhibiting how algorithm performs. This segment demonstrates how algorithm performs in respect to the each other and talks about various part of correlation. In here we investigate the possibility of trouble rating and the idea of relative directions being inalienably troublesome. One of the principle tasks of this part is to characterize test cases for MARD algorithm to get setups of algorithm when apply to sequencing relative direction learning issue. 5.1 Accuracy Result The accuracy of MARD algorithm involves determining how accurately learning the relative directions, and we measure it by counting the number of relative directions and different directional movement case. First, we apply the algorithm to the test three dimensional movement cases and mark the output on the directions. Then, we count all of direction cases. Table 5 shows the results of the relative directions of the proposed algorithm. We find total 24 cases for relative direction movement which is shown in Figure 9 and in Table 4 we have displayed these different results of different tracking agent 3D perspective. Case No. Tracking Agent Handoff Agent Direction Point Direction Point value Relative Direction Directio n Point Relative Direction Direction Point Value 1 a1 0 Right a right 0 b1 1 Left b left 1 c1 2 Up c up 2 d1 3 Down d down 3 e1 4 Forward e forward 4 f1 5 Backward f backward 5
  • 46. 37 © 2017 by Daffodil International University 2 a1 1 Left a right 0 b1 0 Right b left 1 c1 2 Up c up 2 d1 3 Down d down 3 e1 5 Backward e forward 4 f1 4 Forward f backward 5 3 a1 1 Left a right 0 b1 0 Right b left 1 c1 3 Down c up 2 d1 2 Up d down 3 e1 4 Forward e forward 4 f1 5 Backward f backward 5 4 a1 0 Right a right 0 b1 1 Left b left 1 c1 3 Down c up 2 d1 2 Up d down 3 e1 5 Backward e forward 4 f1 4 Forward f backward 5 5 a1 3 Down a right 0 b1 2 Up b left 1 c1 4 Forward c up 2 d1 5 Backward d down 3 e1 1 Left e forward 4 f1 0 Right f backward 5 6 a1 3 Down a right 0 b1 2 Up b left 1 c1 5 Backward c up 2 d1 4 Forward d down 3 e1 0 Right e forward 4 f1 1 Left f backward 5 7 a1 2 Up a right 0 b1 3 Down b left 1 c1 4 Forward c up 2 d1 5 Backward d down 3 e1 0 Right e forward 4 f1 1 Left f backward 5 8 a1 2 Up a right 0 b1 3 Down b left 1 c1 5 Backward c up 2 d1 4 Forward d down 3 e1 1 Left e forward 4 f1 0 Right f backward 5 9 a1 1 Left a right 0 b1 0 Right b left 1 c1 4 Forward c up 2 d1 5 Backward d down 3 e1 2 Up e forward 4 f1 3 Down f backward 5
  • 47. 38 © 2017 by Daffodil International University 10 a1 0 Right a right 0 b1 1 Left b left 1 c1 5 Backward c up 2 d1 4 Forward d down 3 e1 2 Up e forward 4 f1 3 Down f backward 5 11 a1 0 Right a right 0 b1 1 Left b left 1 c1 4 Forward c up 2 d1 5 Backward d down 3 e1 3 Down e forward 4 f1 2 Up f backward 5 12 a1 1 Left a right 0 b1 0 Right b left 1 c1 5 Backward c up 2 d1 4 Forward d down 3 e1 3 Down e forward 4 f1 2 Up f backward 5 13 a1 2 Up a right 0 b1 3 Down b left 1 c1 1 Left c up 2 d1 0 Right d down 3 e1 4 Forward e forward 4 f1 5 Backward f backward 5 14 a1 2 Up a right 0 b1 3 Down b left 1 c1 0 Right c up 2 d1 1 Left d down 3 e1 5 Backward e forward 4 f1 4 Forward f backward 5 15 a1 3 Down a right 0 b1 2 Up b left 1 c1 0 Right c up 2 d1 1 Left d down 3 e1 4 Forward e forward 4 f1 5 Backward f backward 5 16 a1 3 Down a right 0 b1 2 Up b left 1 c1 1 Left c up 2 d1 0 Right d down 3 e1 5 Backward e forward 4 f1 4 Forward f backward 5 17 a1 5 Backward a right 0 b1 4 Forward b left 1 c1 3 Down c up 2 d1 2 Up d down 3 e1 1 Left e forward 4 f1 0 Right f backward 5
  • 48. 39 © 2017 by Daffodil International University 18 a1 4 Forward a right 0 b1 5 Backward b left 1 c1 3 Down c up 2 d1 2 Up d down 3 e1 0 Right e forward 4 f1 1 Left f backward 5 19 a1 5 Backward a right 0 b1 4 Forward b left 1 c1 2 Up c up 2 d1 3 Down d down 3 e1 0 Right e forward 4 f1 1 Left f backward 5 20 a1 4 Forward a right 0 b1 5 Backward b left 1 c1 2 Up c up 2 d1 3 Down d down 3 e1 1 Left e forward 4 f1 0 Right f backward 5 21 a1 4 Forward a right 0 b1 5 Backward b left 1 c1 1 Left c up 2 d1 0 Right d down 3 e1 3 Down e forward 4 f1 2 Up f backward 5 22 a1 5 Backward a right 0 b1 4 Forward b left 1 c1 0 Right c up 2 d1 1 Left d down 3 e1 3 Down e forward 4 f1 2 Up f backward 5 23 a1 4 Forward a right 0 b1 5 Backward b left 1 c1 0 Right c up 2 d1 1 Left d down 3 e1 2 Up e forward 4 f1 3 Down f backward 5 24 a1 5 Backward a right 0 b1 4 Forward b left 1 c1 1 Left c up 2 d1 0 Right d down 3 e1 2 Up e forward 4 f1 3 Down f backward 5 Table 4: Accuracy result of tracking agent’s direction which compare with handoff agent’s directions In Table 4, we can see that there have no changes in handoff agent’s directions because handoff agent’s directions are constant. We can see a change of directions and direction points in tracking
  • 49. 40 © 2017 by Daffodil International University agent part because these results represented by the perpective of hanoff agent’s directions and direction points. 5.2 Comparison To get a thought of how every algorithm performs it is reasonable to plot tackling times in a histogram. Another way for showing the execution is to sort the unraveling times and plots confuse record as opposed to explaining time. Both of these are of intrigue however since they can uncover distinctive things about the algorithms execution. 5.2.1 Comparison with UAV Relative Attitude Estimation Algorithm UAV relative attitude estimation algorithm that consolidates direction of arrival gauges with polarimetric multi-radio wire channel appraisals to figure the relative attitude between two unmanned aerial vehicles [1]. This approach sustainable only for radio frequency based environment but our propose MARD algorithm can estimate relative direction which approach can any computer programming environment. From Figure 13 and Table 5 we can learn that, there are 24 different cases of a tracking agent’s relative direction aspects. Our MARD algorithm can solve these 24 different cases but UAV relative attitude estimation algorithm can not solve all cases because this algorithm can not solve reverse and front cases or reverse and front direction such as, case 2, 3, 4, 6, 8, 10, 12, 14, 16, 17 and 18 which is showed in Table 5 and Figure 13 show a bar graph where this comparison percentages are shown. Algorithm Total Case Solved Failed Case MARD Algorithm 24 Null UAV Relative Attitude Estimation Algorithm 13 Case 2, 3, 4, 6, 8,10, 12, 14, 16, 17 and 18 Table 5: Comparison of case solution between two algorithms. Figure 14 represented the use of area between two algorithms in computer science environment. UAV Relative Attitude Estimation Algorithm can be used only in in UAV or unmanned aerial
  • 50. 41 © 2017 by Daffodil International University vehicle area in computer science. Our MARD algorithm approach can be used in any section of computer science. Figure 13: Comparison graph of case solution between two algorithms. Figure 14: Use of area between two algorithms. UAV Section Any Computer Sicence Section
  • 51. 42 © 2017 by Daffodil International University 5.2.2 Comparison with Others Algorithms So far fewer researches have been done on relative direction. Most of the research work about single agent based relative direction but only one research work about multi-agent based relative direction which published in recent year [1]. So it is very hard to present comparision among those algorithm and approach. Table 6 represents some multi-agent based and single agent based relative direction approaches. Before viewing Table 6, first we need to know about Multi-agent approach and Single agent approach.  Single agent Approach: In single agent approach an agent is to look for informative knowledge into the aggregate conduct of agents which don't really need to be intelligent.  Multi-agent Approach: Multi-agent based approach refers a multi-agent system which composed of multiple intelligent agents inside a domain. The main advantages about multi-agent approach is that when more than one agent are perform multiple interacting in one computer environment. In multi-agent opinion isolated agents who cooperate together to a target but in single agent approach could not be accomplished it because single agent acting alone. Algorithm Multi-agent Approach Single Agent Approach MARD Algorithm Multi-agent based Approach UAV Relative Attitude Estimation Algorithm Multi-agent based approach Relative Direction Change Algorithm Single agent based approach Complexity of Reasoning with Relative Directions Algorithm Single agent based approach
  • 52. 43 © 2017 by Daffodil International University Energy-Efficient Routing Algorithm Based on Relative Identification Single agent based approach Relative Direction of Oriented Points Algorithm Single agent based approach An ant colony optimisation algorithm for the 2D and 3D Single agent based approach Relative Direction Algorithm as a Binary Relation Single agent based approach Table 6: Multi-agent based and single agent based algorithm.
  • 53. 44 © 2017 by Daffodil International University Chapter 6 Discussion This experiment depended on a greatly rich thought. By setting up a algorithm to keep running out of sight of the relative direction learning purpose, we could accomplish uncertain measures of watching time. This enabled us to endeavor a more profound hunt than would have been conceivable in a period apportioned circumstance. With an exceptionally restricted spending we assembled and introduced a beneficiary, spectrometer and control programming, all of which have performed honorably. 6.1 Summary The objective of this proposal is to look at execution of multi-agent relative direction algorithm for the getting the hang of learning direction in a multi-agent scheme environment. Our proposed approach which utilizes diverse sorts of relative directions and two agent. This work parts into two four principle parts. The initial segment is tied in with examining qualities of relative direction in displaying a straightforward algorithm. The algorithm fills in as a stage for getting the hang of steering algorithm. The second piece of this work focuses on scientific investigation, and how they unravel MARD algorithm. In the third part, this proposition has tweaked and execute depth view aspects algorithm keeping in mind the end goal to get answer the principle objective toward the start of this work. In the last part, to make it conceivable to look at algorithm. This examination shows a modern algorithm for processing the relative directions between two agents. MARD algorithm examination differentiating direction develops rules and relative direction guidelines regarding applications computer environment, measuring both objective and subjective accomplishment. The MARD algorithm can dealing with and perceiving relative directions, its executed work strategy or computation puts aside greater open door for result. Its completed computation has few phases for recognizing relative directions.
  • 54. 45 © 2017 by Daffodil International University 6.2 Conclusion The research gives an introduction to Multi-Agent Relative Direction (MARD) algorithm and the diverse approaches to manage making profitable solvers. It also displays some hypothetical establishment about this computation which is of interest while discussing and picking estimation. Finally the algorithm that will be considered in this suggestion is shown. This segment portrays the sequencing relative directions in any multi-agent based issue. Features of an algorithm are analyzed with a particular true objective to arrange the issue into sub-issues like protest discovering issue, structure issue. This thesis moreover gives an introduction of the unmanned elevated vehicles and discusses the similitudes and differences between past research and this theory. This investigation experiment presents bring in a algorithm for figuring the relative direcrions between two agenrs, where how one agent can realize in another agent's relative directions through in PC or programming point of view. In this thesis exhibit our trial outline, practical info, scientific idea and structure of MARD algorithm. We introduce a novel subjective delineation and depiction this MARD algorithm can handle these issues for managing and seeing relative directions, its executed work technique or calculation sets aside more prominent open entryway for result. General we have seen that the MARD Algorithm is better than other said ideas. It is the best Algorithm in finding the relative directions and the second-best of quickest runtime. Versatility of the MARD Algorithm is likewise great. The target of this thesis is to take a vision at execution of MARD algorithm for the getting the hang of learning directions in a multi-agent conspire condition. Our proposed approach which uses different sorts of relative headings and two operators. This work parts into two four guideline parts. The underlying portion is tied in with looking at characteristics of relative heading in showing a direct algorithm. The algorithm fills in as a phase for getting the hang of guiding algorithm. The second bit of this work concentrates on logical examination, and how they disentangle MARD algorithm. This suggestion has changed and execute profundity see viewpoints algorithm remembering the true objective to get answer the rule objective toward the begin of this work. To make it possible to take a algorithm. This examination demonstrates an advanced algorithm for learninf the relative side between. MARD algorithm examination separating heading creates tenets and relative directions rules with respect to applications PC condition, measuring both target and subjective achievement. The MARD algorithm can managing and seeing relative bearings, its executed work system or algorithm sets aside more prominent open entryway for result. Its finished calculation has few stages for perceiving relative directions.
  • 55. 46 © 2017 by Daffodil International University 6.3 Future Work Future work incorporates examining the conduct of this algorithm in connection to the last dissemination of execution times. The extensive difference and stochastic conduct no doubt requests an investigation with access to a lot of computational power. It is additionally fascinating to think about the impact of various temperature plunge strategies utilized as a part of relative direction, with restarting being a reasonable contrasting option to unendingly diminishing temperatures. Because of significance development of MARD algorithm on the point of view of ―Three Dimension (3D)‖, ―Machine Learning‖ and ―Route Direction‖, this is fundamental for better understanding what it takes to find relative directions depiction theoretic hindrances. The farthest point part issue winds up being solidly related to the set up request of finding relative direction on the measurement and mapping. In future we will optimistic do a research about machine learning and route direction perceptive of multi-agent based relative direction learning computation.
  • 56. 47 © 2017 by Daffodil International University Bibliography [1] Attiya Mahmood, Jon W. Wallace, Michael A. Jensen, "Radio Frequency UAV Attitude Estimation Using Direction of Arrival and Polarization," in 2017 11th European Conference on Antennas and Propagation (EUCAP), Paris, France, 2017, pp. 1857-1859. [2] Sebastian Hahn, Joseph Bethge, Jürgen Döllner, "Relative Direction Change - A Topology-based Metric for Layout Stability in Treemaps," in 12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, vol. 3, Porto, Portugal, 2017, pp. 88- 95. [3] Jonghoon Seo, Seungho Chae, Jinwook Shim, Dongchul Kim, Cheolho Cheong and Tack-Don Han, "Fast Contour-Tracing Algorithm Based on a Pixel-Following Method for Image Sensors," Sensors, vol. 16, no. 3, p. 353, 9 March 2016. [4] Jana Götze, Johan Boye, "―Turn left‖ vs. ―Walk towards the café‖: When relative directions work better than landmarks," in AGILE 2015: Geographic Information Science as an Enabler of Smarter Cities and Communities, Lisboa, Portugal, 2015, pp. 253-267. [5] Jae Hee Lee, "The Complexity of Reasoning with Relative Directions," in Frontiers in Artificial Intelligence and Applications, 21st European Conference on Artificial Intelligence (ECAI 2014), vol. 263, Prague, Czech Republic, 2014, pp. 507–512. [6] Jae Hee Lee, Jochen Renz and Diedrich Wolter, "StarVars—Effective Reasoning about Relative Directions," in Proceedings of the Twenty-Third International Joint Conference on Artificial Intelligence (IJCAI 2013), Beijing, China, 2013, pp. 976–982. [7] Chien-Erh Weng and Tsung-Wen Lai, "An Energy-Efficient Routing Algorithm Based on Relative Identification and Direction for Wireless Sensor Networks," Wireless Personal Communications, vol. 69, no. 1, pp. 253–268, March 2013. [8] Till Mossakowski and Reinhard Moratz, "Qualitative Reasoning about Relative Direction of Oriented Points," Artificial Intelligence, vol. 180–181, pp. 34-45, April 2012. [9] Toru Ishikawa and Mika Kiyomoto, "Turn to the Left or to the West: Verbal Navigational Directions in Relative and Absolute Frames of Reference," in 5th International Conference, Geographic Information Science 2008, Park City, UT, USA, 2008, pp. 119-132.
  • 57. 48 © 2017 by Daffodil International University [10] Jan Seibert and Brian L. McGlynn, "A new triangular multiple flow direction algorithm for computing upslope areas from gridded digital elevation models," Water Resources Journal, vol. 43, no. 4, p. W04501, April 2007. [11] Alena Shmygelska and Holger H Hoos, "An ant colony optimisation algorithm for the 2D and 3D hydrophobic polar protein folding problem," BMC Bioinformatics, vol. 6, no. 1, p. 30, 14 February 2005. [12] Reinhard Moratz, "Representing Relative Direction as a Binary Relation of Oriented Points," in ECAI 2006, 17th European Conference on Artificial Intelligence, Riva del Garda, Italy, 2006, pp. Pages 407-411. [13] Jeounghoon Kim and Hugh R.Wilson, "Dependence of Plaid Motion Coherence on Component Grating Directions," Vision Research, vol. 33, no. 17, pp. 2479-2489, December 1993. [14] William S. Albert, Ronald A. Rensink, Jack M. Beusmans, "Learning relative directions between landmarks in a desktop virtual environment," Spatial Cognition and Computation, vol. 1, no. 2, pp. 131–144, June 1999. [15] R M Crowder. (1998, January) University of Southampton. [Online]. HYPERLINK "http://www.southampton.ac.uk/~rmc1/robotics/arkinematics.htm" [16] Mike Foster. (2016, December) Graphicarto.com. [Online]. HYPERLINK "http://www.graphicarto.com/directional-cartography-maps-and-relative-direction/" [17] ArcGIS Desktop. [Online]. HYPERLINK "http://desktop.arcgis.com/en/arcmap/latest/extensions/schematics/relative-main-line-layout- algorithm-properties-page.htm" [18] Seyedali Mirjalili, "Dragonfly algorithm: a new meta-heuristic optimization technique for solving single-objective, discrete, and multi-objective problems," Neural Computing and Applications, vol. 27, no. 4, pp. 1053–1073, May 2016. [19] Zhi Li and and Kris Hauser Jianqiao Li, "A Study of Bidirectionally Telepresent Tele-action During Robot-Mediated Handover," in 2017 IEEE International Conference on Robotics and Automation (ICRA), Singapore, 2017, pp. 2890-2896. [20] Richard W. Hedley, Yiwei Huang and Kung Yao, "Direction-of-arrival estimation of animal vocalizations for monitoring animal behavior and improving estimates of abundance," Avian Conservation & Ecology, vol. 12, no. 1, Article 6, pp. 116-128, June 2017. [21] Dunwen Wei, "Discrete-State-Based Vision Navigation Control Algorithm for One Bipedal Robot," Mathematical Problems in Engineering, vol. 2015, Article ID 168645, p. 12, May 2015. [22] Andrew V. Goldberg, Giuseppe F. Italiano, David S. Johnson and Dorothea Wagner, "Algorithm Engineering (Dagstuhl Seminar 13391)," Dagstuhl Reports, vol. 3, no. 9, pp. 169--189, 2014.
  • 58. 49 © 2017 by Daffodil International University [23] Marc Goerigk and Anita Schöbel, "Algorithm Engineering in Robust Optimization," in Algorithm Engineering. Lecture Notes in Computer Science, Lasse Kliemann and Peter Sanders, Ed. Cham: Springer, 2016, vol. 9220, pp. 245-279. [24] Markus Chimani and Karsten Klein, "Algorithm Engineering: Concepts and Practice," in Experimental Methods for the Analysis of Optimization Algorithms, Thomas Bartz-Beielstein, Marco Chiarandini, Luís Paquete and Mike Preuss, Ed. Berlin, Heidelberg, Germany: Springer, 2010, pp. 131-158. [25] Matthias Müller-Hannemann and Stefan Schirra, Ed., Algorithm Engineering. Berlin, Heidelberg, Germany: Springer, 2010, vol. 5971. [26] Peter Sanders, "Algorithm Engineering –An Attempt at a Definition," in Efficient Algorithms. Lecture Notes in Computer Science, Susanne Albers, Helmut Alt and Stefan Näher, Ed. Berlin, Heidelberg, Germany: Springer, 2009, vol. 5760, pp. 321-340. [27] Yoav Shoham and Kevin Leyton-Brown, Multiagent Systems: Algorithmic, Game-Theoretic, and Logical Foundations, 1st ed. New York, USA: Cambridge University Press, December 2008. [28] Mevludin Glavic, "Agents and Multi-Agent Systems: A Short Introduction for Power Engineers," University of Liege, Electrical Engineering and Computer Science Department, Liege, Belgium, Technical Report May 2006. [29] Hjelmervik H, Westerhausen , Hirnstein M, Specht K and Hausmann M, "The neural correlates of sex differences in left–right confusion," NeuroImage, vol. 113, pp. 196–206, June 2015. [30] Brandt J and Mackavey W, "Left-right confusion and the perception of bilateral symmetry," International Journal of Neuroscience, vol. 12, no. 2, pp. 87-94, 1981. [31] Hannay HJ, Ciaccia PJ, Kerr JW and Barrett D, "Self-report of right-left confusion in college men and women," Perceptual and Motor Skills, vol. 70, no. 2, pp. 451-457, April 1990. [32] Hikari Yamashita, "Self-rated right–left confusability and performance on the Money Road-Map Test," Psychological Research, vol. 77, no. 5, pp. 575–582, September 2013. [33] E.B. Lum, A. Stompel and Kwan-Liu Ma, "Using Motion to Illustrate Static 3D Shape - Kinetic Visualization," IEEE Transactions on Visualization and Computer Graphics (TVCG), vol. 9, no. 2, pp. 115 - 126, April-June 2003. [34] C.R. Karanam and Y. Mostofi, "3D Through-Wall Imagingwith Unmanned Aerial Vehicles Using WiFi," in the proceedings of the 16th ACM/IEEE International Conference on Information Processing in Sensor Networks (IPSN), Pittsburgh, Pennsylvania, USA, April 2017, pp. 131-142. [35] S. Rayhan Kabir, Shaikh Muhammad Allayear, Mirza Mohtashim Alam and Md. Tahsir Ahmed Munna, ―A Computational Technique for Intelligent Computers to learn and identify the Human’s Relative Directions,‖ proceeding in International Conference on Intelligent Sustainable Systems (ICISS 2017), India, 2017, pp. 1037-1040.