A Study On Mobile Robot Simultaneous Localization And Mapping Based Kalman Filter And Particle Filter

The mobile robotics, a new technology that is on the leading edge, incorporates ControlScience, Computer Science, Artificial Intelligence (AI), and Navigation Science and so on.Simultaneous Localization and Mapping (SLAM) plays an important role in the mobilerobotics. SLAM is defined that in an unknown environment, the mobile robot equipped withsensors will rely on that to acquire the environmental information and its own poseinformation to locate its position and build maps.The research contents in this paper are as follows:How to solve the problem of the low accuracy of the mobile robot’s localization and itsmap building becomes a complicated puzzle of the SLAM based on Extended Kalman Filter(EKF-SLAM). In order to deal with the above problem, this paper has put forward thealgorithm of EKF-SLAM based on Distributed State Fusion technology which can beexplained from two parts.(1) The multi-sensor information fusion technology has beendiscussed in detail. The optical fusion theory of the linear minimum variance by weighing thematrixes and the scalars separately is included.(2) This paper integrates the multi-sensorinformation fusion technology into the EKF-SLAM, and proposes the algorithm of theEKF-SLAM related to the Distributed State Fusion. In addition, the algorithm has beenrealized by scalar weighting fusion. Moreover, the results of simulation in MATLAB haveproved that the algorithm provided by this paper has improved the accuracy of the mobilerobot’s localization and map building.This paper then completely expounds the methods of the particle filter (PF) and theinformation filter (IF). Here, the PF-SLAM has three defects.(1) The particle degenerationcaused by sampling will influence the accuracy and the consistency of the SLAM.(2) Onlythe contempory information is used to the process of a single iteration in the absence of thehistory information which will also influence the accuracy and the consistency.(3) How tosignificantly reduce the particle number and the computation complexity, and improve theefficiency of the algorithm becomes an urgent problem. In order to solve the above drawbacks,this paper has adopted PF-SLAM based on the Sparse Extended Information Filter (SEIF-PF SLAM). The simulation experiments in MATLAB have indicated that the SEIF-PF SLAMhas enhanced the accuracy of the localization and the map building of the mobile robot,lowered the computation complexity of the PF, improved the efficiency, and meanwhile meetsthe demands of consistency.