SlideShare ist ein Scribd-Unternehmen logo
1 von 2
Downloaden Sie, um offline zu lesen
Overview. Article is about proposing new robot configuration by introducing new plaining
algorithm. The scope of the paper research is the process of sequential part assembly in multi-
robot environment. The main focus is robot collaboration and proposition of optimal robot
configuration mechanism to overcome common problems.
Paper focuses on finding robot base and arm configuration for optimal task performance while
overcoming existing constraints. The main constraints is: robots cannot intersect with one another
(having sequence of operations, it has to predict future states). When moving to the next sequential
operation robot have two possibilities: 1) change grasp position on the part; 2) move part to next
operation while maintaining same grip position. While humans can perform both of this task
simultaneously, robots have to be programed to do one of them. This presents problem for team
of robots how to decide and which task to perform
Article introduces multi-robot grasp planning as Constraint Satisfaction Problem (CSP) with each
grasp positon in each operation being a variable. This variable imposes two types of constraints:
collision and transfer. In order to be able to find optimal solution to robot grasp planning paper
presents solution algorithm, while assuming that robot should preform regrasp between each of
operations. This enables to solve problem as set of smaller CSP, which can be solved separately
for each assemble operation. Advantage of the suggested algorithm being anytime planner can
be stopped in any time of computation if cost of removing excessive regrasp operation is larger
than cost of preforming this operations. As input this algorithm takes sequence of relative positions
of assembly parts.
The problem which CSP helps to solve is – find grasping configuration for all robots required by
the assemblies in all operations. The goal is to find solution which presents minimal number of
robot regrasp operations. The manufacturing process is done as set of sequential assembly
operations, each operation being either change grasp or move part to next operation without
changing the grip. To solve CSP problem paper presents two solutions: 1) backtracking search
(worst case scenario – exponential in number of variables); 2) local neighborhood computation
(solve conflicts in local neighborhood of the graph). The proposed algorithm starts from solving
the easiest problem first: solve constraint graph with no transfer constrains, then gradually add
new ones and solves them. The flow of the algorithm is 1) solve problems for each of connected
components using Back Tracing search; 2) set collection of solutions from sub operations (step1)
as current best solution; 3) start incremental addition of new transfer constraints, reducing number
of regrasps (this procedure tries to solve problems as fast as possible, iterating over all valid
combinations, prioritizing this solutions, in order to try to solve them first, this reduces time as we
conduct search only in local neighborhood of graph );
Results of conducting experiment shows that proposed algorithm works efficiently and better than
direct – brute-force computation. It generates first “best” solution in aprox. 4 seconds, in most
cases optimal result required one regrasp operations.
Scientific Proposal.
It is worse examining robots cooperation and communication systems. As an example we can use
client-server architecture, where robots that assemble product are clients, and we have
Computational cluster with server running on it. The communication between server and client is
wireless, this requires examination possible interferences to the network through various factory
production elements. Going further in this them we can implement Swarm intelligence (derived
from UAV) to robots, and have robots exchange information in real time between each other. This
approached require development of design making process how and when robot can make a
decision about his next set of action, while cooperating with others. This will help to solve
problem, when robots find themselves trapped inside structure it is assembling.
Also in real world factory setting there are usually additional constrains: limited space for robot
navigation conveyor belts or moving parts. Robots can be unable to place parts for regrasp
operations or they may encounter new obstacles while moving part to next production chain. This
setting requires additional improvements to the CSP algorithm.
As mentioned in paper, in future we must take into account robots that is used to insert fasteners.
They can be added as new variable to the equation, which can have several states: wait, fasten, and
move. From design point of view this robots can either be stationary or they can move. An
example of stationary robot with arm that have only one degree of freedom vertical or horizontal,
to make it easier for other robots to bring parts for fastening operation. Also this robot can have
tracking laser to improve tracking. All this presents new challenges to the CSP algorithm.
Besides we can add number of robots to the equation, in order to calculate optimal number of
robots for the task. If robot can have to states either hold or insert fasteners, we can determine,
which one of the group should be doing first task, which ones second. Of course, we have to limit
number or robots, have redefined maximum, minimum number for specific task, but equation will
calculate optimal number. New algorithm can start with precomputed assumption, quickly check
it for current state, if it satisfying the condition we use it. If not it should be able to accept new
constantans on the go and build new graph of states.

Weitere ähnliche Inhalte

Ähnlich wie Robot Planing Article Overview

Iaetsd protecting privacy preserving for cost effective adaptive actions
Iaetsd protecting  privacy preserving for cost effective adaptive actionsIaetsd protecting  privacy preserving for cost effective adaptive actions
Iaetsd protecting privacy preserving for cost effective adaptive actionsIaetsd Iaetsd
 
An Integrated Prototyping Environment For Programmable Automation
An Integrated Prototyping Environment For Programmable AutomationAn Integrated Prototyping Environment For Programmable Automation
An Integrated Prototyping Environment For Programmable AutomationMeshDynamics
 
High-Speed Neural Network Controller for Autonomous Robot Navigation using FPGA
High-Speed Neural Network Controller for Autonomous Robot Navigation using FPGAHigh-Speed Neural Network Controller for Autonomous Robot Navigation using FPGA
High-Speed Neural Network Controller for Autonomous Robot Navigation using FPGAiosrjce
 
GRAPH MATCHING ALGORITHM FOR TASK ASSIGNMENT PROBLEM
GRAPH MATCHING ALGORITHM FOR TASK ASSIGNMENT PROBLEMGRAPH MATCHING ALGORITHM FOR TASK ASSIGNMENT PROBLEM
GRAPH MATCHING ALGORITHM FOR TASK ASSIGNMENT PROBLEMIJCSEA Journal
 
Bra a bidirectional routing abstraction for asymmetric mobile ad hoc networks...
Bra a bidirectional routing abstraction for asymmetric mobile ad hoc networks...Bra a bidirectional routing abstraction for asymmetric mobile ad hoc networks...
Bra a bidirectional routing abstraction for asymmetric mobile ad hoc networks...Mumbai Academisc
 
Producer consumer-problems
Producer consumer-problemsProducer consumer-problems
Producer consumer-problemsRichard Ashworth
 
Optimized Robot Path Planning Using Parallel Genetic Algorithm Based on Visib...
Optimized Robot Path Planning Using Parallel Genetic Algorithm Based on Visib...Optimized Robot Path Planning Using Parallel Genetic Algorithm Based on Visib...
Optimized Robot Path Planning Using Parallel Genetic Algorithm Based on Visib...IJERA Editor
 
Reinforcement learning-ebook-part1
Reinforcement learning-ebook-part1Reinforcement learning-ebook-part1
Reinforcement learning-ebook-part1Rajmeet Singh
 
Reinforcement Learning / E-Book / Part 1
Reinforcement Learning / E-Book / Part 1Reinforcement Learning / E-Book / Part 1
Reinforcement Learning / E-Book / Part 1Hitesh Mohapatra
 
Optimizing Mobile Robot Path Planning and Navigation by Use of Differential E...
Optimizing Mobile Robot Path Planning and Navigation by Use of Differential E...Optimizing Mobile Robot Path Planning and Navigation by Use of Differential E...
Optimizing Mobile Robot Path Planning and Navigation by Use of Differential E...IOSR Journals
 
Trajectory Planning Through Polynomial Equation
Trajectory Planning Through Polynomial EquationTrajectory Planning Through Polynomial Equation
Trajectory Planning Through Polynomial Equationgummaavinash7
 
EuroAD 2021: ChainRules.jl
EuroAD 2021: ChainRules.jl EuroAD 2021: ChainRules.jl
EuroAD 2021: ChainRules.jl Lyndon White
 
A Path Planning Technique For Autonomous Mobile Robot Using Free-Configuratio...
A Path Planning Technique For Autonomous Mobile Robot Using Free-Configuratio...A Path Planning Technique For Autonomous Mobile Robot Using Free-Configuratio...
A Path Planning Technique For Autonomous Mobile Robot Using Free-Configuratio...CSCJournals
 
Comparison Between Genetic Fuzzy Methodology and Q-Learning for Collaborative...
Comparison Between Genetic Fuzzy Methodology and Q-Learning for Collaborative...Comparison Between Genetic Fuzzy Methodology and Q-Learning for Collaborative...
Comparison Between Genetic Fuzzy Methodology and Q-Learning for Collaborative...gerogepatton
 
Comparison Between Genetic Fuzzy Methodology and Q-Learning for Collaborative...
Comparison Between Genetic Fuzzy Methodology and Q-Learning for Collaborative...Comparison Between Genetic Fuzzy Methodology and Q-Learning for Collaborative...
Comparison Between Genetic Fuzzy Methodology and Q-Learning for Collaborative...ijaia
 

Ähnlich wie Robot Planing Article Overview (20)

Iaetsd protecting privacy preserving for cost effective adaptive actions
Iaetsd protecting  privacy preserving for cost effective adaptive actionsIaetsd protecting  privacy preserving for cost effective adaptive actions
Iaetsd protecting privacy preserving for cost effective adaptive actions
 
Termpaper ai
Termpaper aiTermpaper ai
Termpaper ai
 
H011114758
H011114758H011114758
H011114758
 
An Integrated Prototyping Environment For Programmable Automation
An Integrated Prototyping Environment For Programmable AutomationAn Integrated Prototyping Environment For Programmable Automation
An Integrated Prototyping Environment For Programmable Automation
 
High-Speed Neural Network Controller for Autonomous Robot Navigation using FPGA
High-Speed Neural Network Controller for Autonomous Robot Navigation using FPGAHigh-Speed Neural Network Controller for Autonomous Robot Navigation using FPGA
High-Speed Neural Network Controller for Autonomous Robot Navigation using FPGA
 
GRAPH MATCHING ALGORITHM FOR TASK ASSIGNMENT PROBLEM
GRAPH MATCHING ALGORITHM FOR TASK ASSIGNMENT PROBLEMGRAPH MATCHING ALGORITHM FOR TASK ASSIGNMENT PROBLEM
GRAPH MATCHING ALGORITHM FOR TASK ASSIGNMENT PROBLEM
 
Robotic arm tool
Robotic arm toolRobotic arm tool
Robotic arm tool
 
Bra a bidirectional routing abstraction for asymmetric mobile ad hoc networks...
Bra a bidirectional routing abstraction for asymmetric mobile ad hoc networks...Bra a bidirectional routing abstraction for asymmetric mobile ad hoc networks...
Bra a bidirectional routing abstraction for asymmetric mobile ad hoc networks...
 
FrackingPaper
FrackingPaperFrackingPaper
FrackingPaper
 
compiler design
compiler designcompiler design
compiler design
 
Producer consumer-problems
Producer consumer-problemsProducer consumer-problems
Producer consumer-problems
 
Optimized Robot Path Planning Using Parallel Genetic Algorithm Based on Visib...
Optimized Robot Path Planning Using Parallel Genetic Algorithm Based on Visib...Optimized Robot Path Planning Using Parallel Genetic Algorithm Based on Visib...
Optimized Robot Path Planning Using Parallel Genetic Algorithm Based on Visib...
 
Reinforcement learning-ebook-part1
Reinforcement learning-ebook-part1Reinforcement learning-ebook-part1
Reinforcement learning-ebook-part1
 
Reinforcement Learning / E-Book / Part 1
Reinforcement Learning / E-Book / Part 1Reinforcement Learning / E-Book / Part 1
Reinforcement Learning / E-Book / Part 1
 
Optimizing Mobile Robot Path Planning and Navigation by Use of Differential E...
Optimizing Mobile Robot Path Planning and Navigation by Use of Differential E...Optimizing Mobile Robot Path Planning and Navigation by Use of Differential E...
Optimizing Mobile Robot Path Planning and Navigation by Use of Differential E...
 
Trajectory Planning Through Polynomial Equation
Trajectory Planning Through Polynomial EquationTrajectory Planning Through Polynomial Equation
Trajectory Planning Through Polynomial Equation
 
EuroAD 2021: ChainRules.jl
EuroAD 2021: ChainRules.jl EuroAD 2021: ChainRules.jl
EuroAD 2021: ChainRules.jl
 
A Path Planning Technique For Autonomous Mobile Robot Using Free-Configuratio...
A Path Planning Technique For Autonomous Mobile Robot Using Free-Configuratio...A Path Planning Technique For Autonomous Mobile Robot Using Free-Configuratio...
A Path Planning Technique For Autonomous Mobile Robot Using Free-Configuratio...
 
Comparison Between Genetic Fuzzy Methodology and Q-Learning for Collaborative...
Comparison Between Genetic Fuzzy Methodology and Q-Learning for Collaborative...Comparison Between Genetic Fuzzy Methodology and Q-Learning for Collaborative...
Comparison Between Genetic Fuzzy Methodology and Q-Learning for Collaborative...
 
Comparison Between Genetic Fuzzy Methodology and Q-Learning for Collaborative...
Comparison Between Genetic Fuzzy Methodology and Q-Learning for Collaborative...Comparison Between Genetic Fuzzy Methodology and Q-Learning for Collaborative...
Comparison Between Genetic Fuzzy Methodology and Q-Learning for Collaborative...
 

Robot Planing Article Overview

  • 1. Overview. Article is about proposing new robot configuration by introducing new plaining algorithm. The scope of the paper research is the process of sequential part assembly in multi- robot environment. The main focus is robot collaboration and proposition of optimal robot configuration mechanism to overcome common problems. Paper focuses on finding robot base and arm configuration for optimal task performance while overcoming existing constraints. The main constraints is: robots cannot intersect with one another (having sequence of operations, it has to predict future states). When moving to the next sequential operation robot have two possibilities: 1) change grasp position on the part; 2) move part to next operation while maintaining same grip position. While humans can perform both of this task simultaneously, robots have to be programed to do one of them. This presents problem for team of robots how to decide and which task to perform Article introduces multi-robot grasp planning as Constraint Satisfaction Problem (CSP) with each grasp positon in each operation being a variable. This variable imposes two types of constraints: collision and transfer. In order to be able to find optimal solution to robot grasp planning paper presents solution algorithm, while assuming that robot should preform regrasp between each of operations. This enables to solve problem as set of smaller CSP, which can be solved separately for each assemble operation. Advantage of the suggested algorithm being anytime planner can be stopped in any time of computation if cost of removing excessive regrasp operation is larger than cost of preforming this operations. As input this algorithm takes sequence of relative positions of assembly parts. The problem which CSP helps to solve is – find grasping configuration for all robots required by the assemblies in all operations. The goal is to find solution which presents minimal number of robot regrasp operations. The manufacturing process is done as set of sequential assembly operations, each operation being either change grasp or move part to next operation without changing the grip. To solve CSP problem paper presents two solutions: 1) backtracking search (worst case scenario – exponential in number of variables); 2) local neighborhood computation (solve conflicts in local neighborhood of the graph). The proposed algorithm starts from solving the easiest problem first: solve constraint graph with no transfer constrains, then gradually add new ones and solves them. The flow of the algorithm is 1) solve problems for each of connected components using Back Tracing search; 2) set collection of solutions from sub operations (step1) as current best solution; 3) start incremental addition of new transfer constraints, reducing number of regrasps (this procedure tries to solve problems as fast as possible, iterating over all valid combinations, prioritizing this solutions, in order to try to solve them first, this reduces time as we conduct search only in local neighborhood of graph ); Results of conducting experiment shows that proposed algorithm works efficiently and better than direct – brute-force computation. It generates first “best” solution in aprox. 4 seconds, in most cases optimal result required one regrasp operations.
  • 2. Scientific Proposal. It is worse examining robots cooperation and communication systems. As an example we can use client-server architecture, where robots that assemble product are clients, and we have Computational cluster with server running on it. The communication between server and client is wireless, this requires examination possible interferences to the network through various factory production elements. Going further in this them we can implement Swarm intelligence (derived from UAV) to robots, and have robots exchange information in real time between each other. This approached require development of design making process how and when robot can make a decision about his next set of action, while cooperating with others. This will help to solve problem, when robots find themselves trapped inside structure it is assembling. Also in real world factory setting there are usually additional constrains: limited space for robot navigation conveyor belts or moving parts. Robots can be unable to place parts for regrasp operations or they may encounter new obstacles while moving part to next production chain. This setting requires additional improvements to the CSP algorithm. As mentioned in paper, in future we must take into account robots that is used to insert fasteners. They can be added as new variable to the equation, which can have several states: wait, fasten, and move. From design point of view this robots can either be stationary or they can move. An example of stationary robot with arm that have only one degree of freedom vertical or horizontal, to make it easier for other robots to bring parts for fastening operation. Also this robot can have tracking laser to improve tracking. All this presents new challenges to the CSP algorithm. Besides we can add number of robots to the equation, in order to calculate optimal number of robots for the task. If robot can have to states either hold or insert fasteners, we can determine, which one of the group should be doing first task, which ones second. Of course, we have to limit number or robots, have redefined maximum, minimum number for specific task, but equation will calculate optimal number. New algorithm can start with precomputed assumption, quickly check it for current state, if it satisfying the condition we use it. If not it should be able to accept new constantans on the go and build new graph of states.