Mobile Robot SLAM Algorithm Based On Particle Filter

With the development of autonomous mobile robots,autonomous navigation technology has become a hot topic in the field of robotics.In order to realize the autonomous navigation of mobile robot,the simultaneous localization and mapping(SLAM)of mobile robot in unknown environment is the key to solve the problem.Firstly,the paper describes the research background,significance and development of SLAM algorithm.Through the Matlab simulation experiment analysis and comparison of the two classical SLAM algorithm: Based on the extended kalman filter and particle filter.Secondly,the Fast SLAM algorithm is prone to accumulation of linearization error,the SLAM Algorithm is based on strong tracking unscented kalman filter(STUFast SLAM),using strong tracking kalman unscented filter to estimate the posterior probability density function by replacing extended kalman filter.It can improve the particle sampling precision and the robustness of the algorithm.The simulation results show the effectiveness of the algorithm.Then,on the basis of STUKF-SLAM algorithm,two improved Fast SLAM algorithms:SLAM algorithm based on differential evolution and improved resampling are proposed to solve the problem of particle diversity.These two algorithms are improved based on resampling,the first algorithm uses differential evolution algorithm to replace the resampling process.it reduces the particle depletion trend and improves the robustness of the algorithm;In the second algorithm,a new set of particles is obtained by combining big weight particles with small weight particles after resampling,so that the information of the small weight particles is preserved and the diversity of the particles is increased.The validity of the algorithm is verified by simulation and victoria park experiments.Finally,on the experimental platform of QBot2,we apply the improved algorithm in the SLAM experiment of indoor robot and verify the validity of the algorithm.