FastSLAM Method Based on Gaussian Particle Swarm Optimization

Abstract

In simultaneous location and mapping of mobile robots, the traditional FastSLAM method has the problem of sampling particles degeneration and shortage and the problem of needing a large number of particles for positioning accuracy. To solve these problems, this paper studies a FastSLAM method based on gaussian particle swarm optimization(GPSO). The method uses the gaussian particle swarm optimization algorithm in the process of particle filter to update the sampling particles, and then increases the diversity of the sampling particle, relieves the poor sampling particle degradation problems. The other hand, the method makes the particles set distribution in the region of the true state of the high likelihood, and increases the convergence of particles as well as the accuracy of robot localization and map building, at the same time can reduce the particle filter for the required number of particles. The simulation result shows that the method is feasible and effective.