Mobile Robot Location Algorithm Based on Improved Particle Filtering

Abstract

To solve the simultaneous localization and mapping (SLAM) problem, many techniques have been proposed, and the Particle Filter (PF) is one of effective ways. However, the PF algorithm needs a large number of samples to approximate the posterior probability density of the system, which makes the algorithm complex. What's more, the judgment of resampling is imperfect. Based on this, an improved PF algorithm which introducing population diversity factor and genetic algorithm into the process of re-sampling is proposed in this paper. The effective sample size and the population diversity factor are used to determine whether to re-sampling. When re-sampling is needed, the genetic algorithm is used to optimize the particle set. The simulation result shows that estimation accuracy of the improved algorithm is better than that of traditional particles filter, not only in accuracy, but also in efficiency.