A humanoid robot path planning method based on virtual force-directed particle swarm optimisation

Abstract

In this paper, a path planning method for a humanoid robot based on virtual force-directed Particle Swarm Optimisation (PSO) is presented. We take each obstacle as a source of repulsive force affecting the humanoid robot, and the target as a source of attractive force, combining PSO with footstep planning method to find a path for the robot. We aim at avoiding the slow convergence of PSO path planning and affect the speed and the process of updating particles of PSO. A method that can guide the particle evolution and accelerate the convergence of the algorithm by the virtual force between the mobile humanoid robot and obstacles or destination is presented. Experiment shows that this method can effectively realise humanoid robot path planning. Compared with the basic particle swarm path planning algorithm, the convergence speed of the virtual force-directed particle swarm path planning method is faster and less time-consuming.