Visual Localization of Workpiece based on Evolutionary Algorithm and its Application in Industrial Robot

Abstract

Machine vision is one of the important means for industrial robots to obtain position parameters and operation information. Its introduction into robot positioning assembly system has greatly improved the intelligent level of robots and the automation level of the system. The industrial robot system equipped with machine vision technology can make robots break away from simple teaching or offline programming, significantly improving industrial production efficiency and reducing labor costs. Visual positioning of workpiece is the process of determining the position of workpiece relative to robot. It can be combined with many different types of robots and industrial automation systems, including picking robots, conveyor systems, assembly lines and stacking systems. The most common application of workpiece visual positioning is to use it as a part of automatic picking system. In this case, the workpiece must be able to move freely through the predefined path when the robot arm picks up. In this paper, MATLAB software is used to build the robot model, and the optimization program of its trajectory planning is written. In addition to using the improved differential evolution algorithm to optimize the robot trajectory, this paper also uses genetic algorithm and traditional differential evolution algorithm to optimize the problem. By comparing the optimization results of three optimization methods, it is proved that the improved differential evolution algorithm is an optimization method with excellent performance, high efficiency, parallelism, robustness and other advantages.