Laser Localization of Autonomous Vehicle Using an Improved Particle Filter

This article presents an improved Rao-Blackwellized particle filter to overcome particles degeneracy phenomenon and acquire the better localization precision of the autonomous vehicle. The joint posteriori probability density is given that being correlative with the position and pose of the autonomous vehicle and the mark characters of the map. The algorithm utilizes a Markov chain Monte Carlo method with the sampling particle of the target to the resample mechanism of the Rao-Blackwellized particle filter. Simulation results show that the improved algorithm is valid.