An Optimal Smooth-Path Motion Planning Method for a Car-like Mobile Robot

Abstract

This paper proposes an optimal motion planning method consisting of a genetic algorithm (GA), potential field (PF), and Dubins curve for a Car-like mobile robot to solve the problem of finding the shortest and most feasible path in the global environment. Firstly, the GA finds the shortest path by evaluating, selecting, crossing over, and mutating from the initial population and finally provides the strongest individual evolution. Then the result from the GA is further applied with the PF algorithm to improve the ability of obstacle avoidance in the environment. Finally, the Dubins curve method is combined to smooth the path and helps the Car-like mobile robot solve the nonholonomic constraints problem. The major advantages of this method include finding the shortest path, improving avoidance obstacle ability, and smoothing the output path in an environment effectively. The simulation of the proposed method is executed on MATLAB to verify the ability to solve motion planning problems for a Car-like mobile robot.