Vision-Based Safety System for Industrial Collaboration Robot

Ho, Jun Hao (2023) Vision-Based Safety System for Industrial Collaboration Robot. Final Year Project (Bachelor), Tunku Abdul Rahman University of Management and Technology.

Text Full Text_HO JUN HAO.pdf Restricted to Registered users only Download (16MB)

Abstract

Collaborative Robots (COBOT) enables human and robot to work together in the same workspace. With high demands of manufacturing industries to produce many products, more robots to increase efficiency of manufacturing will also be in high demand as well. Hence many safety system is introduced to allow human to work safely around COBOTs. This project aims to integrate Vision System, by using the 3D Kinect V2 camera, to provide better safety system as it can scans the environment of the workspace far better than conventional sensors. Multiple papers has been reviewd, including theory and concepts, with previous works that included Vision-based system as well. Panda Robot, a 7 D.O.F. robot, developed by Franka Emika is used to simulate Pick and Place process in this research paper. The method propose for the vision-based safety system is to control the velocity of the COBOT according to distance of human to the pick and place workspace, and to have collision avoidance if the human is too close and blocks the path of the COBOT. The method is simulated in RViz with the integration of MoveIt! and as mentioned, Kinect V2 camera, with the use of Octomap Representation to sense human and act as obstacle for the Panda Robot. Robot trajectory is based on the RRTConnect search algorithm, that also used in collision avoidance.Various experiments were conducted to test the proposed method, where the motion time was recorded for the COBOT to move from a point to another, increase when human is near and blocks the path. As the robot would move at slower pace and, or to avoid the human or any object obstacle to reach its destination would increase in the total travelling time. As a conclusion, the integration of vision system and 3D camera to detect object or human obstacle into the COBOT, the safety of human in working together with the COBOT is greatly increased.