Research of Escaping Local Minima Strategy for Artificial Potential Field

Abstract

Artificial Potential Field method has been used in many practical purposes for its simplicity and validity. However,it has obvious shortcomings with the incidental trap in local minima and oscillations in the presence of obstacles or in narrow passages. How to avoid and solve these problems has been an active topic in the Artificial Potential Field research. A transformative Gaussian function with improving the expression of the potential environment was proposed. By analyzing the cause of oscillation and the character of the robot path in a local minimum,the PSO was integrated into the process of detecting the potential. Also a new method ELM was proposed to escape the local minima. Simulation results indicate the efficiency of the method in avoiding the oscillation and decreasing the probability of trapping in local minima.