The Research Of Mobile Robot Localization Method Based On Bayesian Theory

With the continuous development of information and intelligence, mobile robots are gradually coming into view, improve people’s lives. Mobile robot localization are prerequisites of robot navigation, motion and so on, it is the basic issue of mobile robot research. In this paper, we focuses on the research of mobile robot localization based on bayesian theory, the main research work is as follows:①We establish the mathematical modeling for mobile robot localization in the framework of bayesian theory, focus on analysis the data association and state estimation issues in mobile robot localization.②In order to improve the efficiency of data association for the mobile robot localization algorithm. We propose a novel data association approach called K-JCBB. Firstly, the little correlative measurements are separated into several groups by K-means Clustering. The number of groups is decided by the characteristics of the environment. Secondly, several local correlations are got though using JCBB and ICNN to each group. Finally, the local correlation is attached to find the most joint compatibility one as the best global correlation. K-JCBB can improve the efficiency of data association algorithm significantly in the environments with dense features.③In order to obtain precision localization state estimation of mobile robot localization. We analyze the extended kalman filter algorithm, unscented kalman filter algorithm and the cubature kalman filter algorithm, and utilize them in two typical nonlinear state estimation, the running time and estimation accuracy are analyzed. We utilize cubature kalman filter algorithm for mobile robot localization because it has high precision and low time complexity.④In order to verify the correctness of the above methods. In this paper, we establish motion process modeling and observation process modeling of mobile robot in the framework of bayesian theory, a simulation platform is established on MATLAB based on these modeling. We compare the performance of CKF-SLAM to EKF-SLAM.The main work and achievements are summarized in the last. We further pointed out the direction of future research work.