1. Autonomous Decision Making RoboArm
Salvatore Siciliano ‘16, Madeleine Boudreau ‘16, Thai Le ‘16, Patrick Hoagland ‘16
Advisor: Dr. Joseph Palladino
Abstract
The goal of the RoboArm Senior Project was to autonomously complete a Trinity “T” shaped puzzle. The
project combines mechanical and electrical components with decision making logic, which analyze the
colors present in a video feedback of the robot’s environment. The robot uses LabVIEW graphical
programming by National Instruments along with a Logitech webcam to capture an image of a block to
determine its color, then again to determine the presence of a block in a given position on the puzzle. The
mechanical components were designed to allow the robot to access all 45 positions on the puzzle using
five degrees of freedom. These degrees of freedom were controlled by four servo motors and one stepper
motor using pulse-width modulation signals and DC power sources. Although the RoboArm can
successfully complete the puzzle, using hard-coded positions decreased the accuracy of each placement.
This could be improved by adding encoders. Overall, the project demonstrates how incorporating decision
making based on feedback signals can improve quality and safety in industrial robot-aided production.
References & Acknowledgements
Acknowledgements:
- Trinity College Engineering Department
- Professor Taikang Ning
- Professor John Mertens
- Technician: Andrew Musulin
References:
1. NI myRIO-1900 [Digital image]. (n.d.). Retrieved April 28, 2016, from http://www.ni.com/myrio/
2 Savox Servo Motor [Digital image]. (n.d.). Retrieved April 28, 2016, from
http://www.amazon.com/Savox-SB-2270SG-Monster-Brushless- Standard/d p/B005OKVQK4
3. Superior Electric SLO-SYN Stepper Motor [Digital image]. (n.d.). Retrieved April 28, 2016, from
http://www.mcsupplyco.com/uploads/productlayouts/ProductDetailDisplay.asp?DirectLink=true&Produc
tID=1496
4. SainSmart Microstep Driver ST-M5045 [Digital image]. (n.d.). Retrieved April 28, 2016, from
http://www.amazon.com/SainSmart-Micro-Stepping-Stepper-Driver-Bi-
polar/dp/B00DFSF9GE/ref=sr_1_cc_1?s=aps&ie=UTF8&qid=1462147458&sr=1-1-
catcorr&keywords=sainsmart CNC stepper
Movement ConceptDesign Components
• Hardware Components:
Processor: NI myRIO-1900
Servo Motors:
o 1x Savox SB-2270SG
o 1x TowerPro MG996R
o 1x TowerPro MG995
o 2x HiTEC HS-225BB
Stepper Motor: Superior Electric SLO-SYN KML061F05
Stepper Motor Driver: SainSmart Microstep Driver ST-M5045
Belt System
Camera: Logitech C270
DC Power Supplies: 5V, 7V and 30V
• Software: LabVIEW by National Instruments
Graphical Programming Language
Figure 2. NI myRIO-1900 Microcontroller [1]
Figure 3. Savox SB-2270SG
Servo Motor [2]
Figure 5. SainSmart Microstep
Driver ST-M5045 [4]
Figure 4. Superior Electric SLO-SYN
KML061F05 [3]
Figure 6. 45 Block Positions in the Puzzle
The RoboArm’s movement is dependent on the color
recognition feedback signal from the Logitech C270 camera
mounted on the front of the arm. The general movement
procedure is described below:
1. Approach pick-up position for the blocks.
2. At the pick-up location, the camera will send a feedback
signal of the block, and the arm will process that signal to
detect the color, blue or yellow, of the block to be placed.
3. Once the color is determined, the arm will then travel to
Design
the first position for a block of that color. Position 1 for blue, position 11 for yellow as seen in Figure 6.
4. Once it has arrived, the camera will send another feedback signal of its environment, and the robot will
process if there is a block of that color already placed in that position.
5. If a block has already been placed in that position, the RoboArm will then travel to the next position for
that color, check for the presence of a block again, and proceed until it finds an empty position.
6. Once an empty position is found, the RoboArm will drop the block and return to the pick-up position.
This process is repeated until the full puzzle is complete. This movement concept allows the RoboArm to
pick up blocks of different colors in any order and make decisions based on a feedback signal to fill the
puzzle appropriately.
Problem Definition
The goal of this project was to design a robotic arm that would be able to sort blocks colored yellow and
blue, pick up the blocks, and place them accordingly in the final Trinity “T” design in Figure 5.
The “T” is 10 inches long by 10 inches wide, and is comprised of 45 different compartments for the robot
to place blocks into. The blocks themselves are ¾” cubes.
The RoboArm’s success is based off of three metrics: the ability to correctly identify the color of the blocks,
the ability to pick up and move the blocks once identified, and finally the ability to correctly position the
blocks in the appropriate places in order to build the final design. If the RoboArm is able to fulfill all three
of these metrics, the project will be considered a complete success. If two of these metrics are fulfilled,
then RoboArm will be semi-successful. Finally if only one of these metrics is fulfilled, then RoboArm will
be deemed unsuccessful and needing further work.
Results and Discussion
Figure 7 (left) : The final design for the RoboArm is comprised of five
degrees of freedom. The first degree of freedom is a stepper motor that
moves the RoboArm’s base horizontally along the bottom edge of the
“T”. Above the base are three members whose motion is controlled with
four servo motors. Attached at the end of the three members there is a
gripper which is controlled by a fifth servo motor and a Logitech C270
HD Webcam.
Figure 8 (right) : Full RoboArm Design Assembly including completed
puzzle, RoboArm, Stepper Motor and Band, microstep motor driver, and
MyRIO
The RoboArm was able to successfully complete the Trinity “T” design autonomously. The use of hard
coded positions for the blocks increased the potential for inaccuracies and scanning each position to
determine a vacant location slowed down overall efficiency. Although this design and movement
concept are not the most efficient designs possible, they were designed to comply with the original
problem statement and highlight the benefits of adding feedback signals and decision making to
improve quality and safety in industrial robot-aided production. Increasing the degrees of freedom and
implementing a more complex image processing algorithm for a live camera feed would further improve
the accuracy and efficiency of this robotic arm.