Path Planning of Mobile Robots Based on Improved Potential Field Algorithm

In view of the problems existing in the traditional potential field algorithm (PFA), such as unrecognized path, local minimal trap and oscillation, a potential field algorithm combining multi-behavior strategy and obstacles variable affected range, and the path planning of robots suitable for the complex obstacle environment is proposed. The obstacles affected range can be changed to eliminate the common necessary conditions of the above problems, so as to avoid unrecognized path, oscillation caused by multiple obstacles and local minimal trap caused by multiple obstacles in advance. Based on the new classification method for step-by-step and oscillation, the multi-behavior strategy is designed with the exact starting and ending conditions. The behaviors are switched by predicting the common expression of problems and the connection of starting and ending conditions, thus avoiding the local minimum trap caused by single obstacle and the oscillation caused by single obstacle in advance. The simulated results based on MATLAB verify the effectiveness and stability of the proposed method in the complex battlefield obstacle environment, and the proposed method has the feasibility for path planning compared with potential field algorithm, dynamic window approach, A-star algorithm and rapid-exploration random tree algorithm. © 2020, Editorial Board of Acta Armamentarii. All right reserved.