Three-Dimensional Localization of Known Objects for Robot Arm Application based on a Particle Swarm Optimized Low End Stereo Vision System

Abstract

This paper presents an automated system for 3D localization of known objects to be reached out to by a robot arm. The developed system mainly consists of a customized vision system, PC, robot arm controller, serial communication connections and robot which is a five-axis articulated robot arm. The customized vision system feeds the scene images ahead of the robot arm to the PC; a computer vision algorithm on the PC receives the fed images. The computer vision algorithm processes and extracts the needed data for the system. The extracted data is then sent to the robot arm controller via an RS232 serial connection. The robot arm controller then controls the robot kinematics and joints velocities based on the desired localization of the robot arm end-effector and motion velocity. Due to the use of customized stereo vision with low-end cameras, a particle swarm optimization (PSO) algorithm is employed to minimize the estimated object coordinates error. The successfulness of the localization is then demonstrated through the successful reach-out of the robot arm to the targeted object.