Robot Path Planning Based on Improved Particle Swarm Optimization

Abstract

Path planning is one of the research hotspots of today's robotics. This paper makes some improvements in allusion to the issue that particle swarm has poor local search ability and sinks into local optimum easily in robot path planning. In this method, the inflection point factor is introduced into the fitness function to cut down the number of turns and accelerate the convergence of the algorithm, and a new adaptive and dynamically adjusted inertia weight is adopted to adjust the local and global search capabilities of particles, so as to avert sinking into the suboptimization prematurely and improve the search quality. Experimental comparison results indicate that the algorithm can improve the quality of particle search in the early and late stage, so that an optimal path can be planned more quickly and effectively.