Using Genetic Algorithms and Bézier Curves for Automatic Path Optimization of a 6-DOF Robot

Abstract

This paper describes a novel method for automatically planning point-to-point motion paths for a robot with six degrees of freedom (6-DOF). A linear motion path between two points on such robots may be impractical due to joint angle constraints or exceeding the manipulator's operational range. The proposed method employs a genetic algorithm to generate suitable motion paths based on the second-, third-, and fourth-orders of Bézier curves. The control points of Bézier curves are determined using a genetic algorithm, which can adjust the fitness function as the end-effector moves closer to the obstacle. As a result, the algorithm can adjust its motion path planning in response to obstacles. The motion paths are generated with the goal of optimizing the robot's inverse kinematic configuration. The results show that using a genetic algorithm and Bézier curves can produce motion paths with smooth transitions, minimal changes in joint angles, and no sudden jerks within the robot's operational area in both obstacle-free and obstacle avoidance scenarios. This solution may be useful for intelligent robots with automated path-planning capabilities in unknown environments.