Improved Artificial Moment Method for Decentralized Local Path Planning of Multirobots

Abstract

An improved artificial moment method is presented for the decentralized local path planning of multirobots. In the method, coordinated and repulsive moments are improved with critical factors, attractive moments are improved with attractive angles, and attractive points are also improved. Furthermore, a modified method for robots final motion vectors is proposed, where if a robot has dangerous companions, then the superposition principle may not be used for its final motion vector. For these improvements, the artificial moment method can be used safely in complicated dynamical environments while high quality solutions are often yielded. Simulation results indicate that the improved method is effective.