The Application Of Nonlinear Filtering In The Mobile Robot Slam

With the development of mobile robot in research and application, Simultaneous localization and map building (SLAM) has attracted immense attention in mobile robotics field. The SLAM problem has attracted a lot of researchers with a broad range of interests and applications.For the nonlinear property of the process model and the observation model, This paper focuses on researching the application of nonlinear filter algorithm in the SLAM. Several improved methods and novel solutions are presented in order to improve consistency and computational efficiency, and additionally extend SLAM application domains. The main content of this dissertation include the following aspects:Firstly, a systemic execution of the algorithm based on the Extended Kalman Filter (EKF) is presented, and successfully tested in simulation. The results show that the estimation precision and robustness are improved by an appropriate increase in the number of environmental features and to speed up the update frequency of the robot pose state and environmental map.Secondly, Unscented Kalman Filter(UKF) is applied to SLAM algorithm to overcome the disadvantage of the traditional EKF SLAM algorithm. In the EKF algorithm, a nonlinear system is linearized to a linear system, which may lead to the decline of accuracy and even instability of the filter. SUT-EKF SLAM algorithm is given according to the environmental features of the state vector is linear. The algorithm applies the scaled unscented transformation(SUT) only to the vehicle states by using the EKF both in the prediction of the map features and the update of the complete states vector. The simulation results indicate that UKF SLAM and SUT-EKF SLAM can reduce the EKF linearization error effectively, and the second method is more efficient in computation.Finally, The application of Particle Filter is studied in the SLAM. In order to overcome the disadvantage of the EKF, two modified Fast SLAM is presented. Firstly, a UPF-Fast SLAM method that integrats the frame work of Fast SLAM is proposed,in which EKF method is substituted by UKF to estimate the robot mobile path. Secondly, IEKF is used to replace the EKF to estimate the mobile robots path and map. These methods can improve the estimation precision of the robot pose state, the results show the efficiency of the two methods.