Simultaneous localization and mapping method for a planet rover based on a Gaussian filter

Abstract

This paper addresses the issues of the application of the simultaneous localization and mapping (SLAM) for solving the problems of navigation and trajectory planning of a planet rover when moving along an unknown surface of the celestial body of the solar system. To solve the high-precision navigation problem of a planetary rover in the Mars exploration environment, processing of large amounts of information from various measuring instruments in real-time is required. When solving practical navigation problems in SLAM method, an extended Kalman filter (EKF) or particle filter (PF) is often used to process information. The particle filter, which is used in the Gmapping SLAM method, is better suited for the analysis of nonlinear systems; however, its computational volume is much larger than that of the extended Kalman filter. Since extended Kalman filter and particle filter have disadvantages for the navigation system, it is proposed to use the Gaussian filter and its modifications. The Gaussian smoothing filter algorithm based on the distributed computing scheme (DIS RTP) is significantly superior to the extended Kalman filter algorithms and particle filter algorithms in terms of computational speed. A new DIS RTP-Gmapping SLAM method is proposed to solve the problem of navigation and mapping of planetary rover. This algorithm combines the efficiency and accuracy of the DIS RTP algorithm and the good performance of the Gmapping SLAM method when working in a complex environment. A comparison of the proposed method with Gmapping SLAM method was carried out, which showed that with an equal number of selected particles, DIS RTP-Gmapping SLAM method demonstrates higher accuracy.This paper addresses the issues of the application of the simultaneous localization and mapping (SLAM) for solving the problems of navigation and trajectory planning of a planet rover when moving along an unknown surface of the celestial body of the solar system. To solve the high-precision navigation problem of a planetary rover in the Mars exploration environment, processing of large amounts of information from various measuring instruments in real-time is required. When solving practical navigation problems in SLAM method, an extended Kalman filter (EKF) or particle filter (PF) is often used to process information. The particle filter, which is used in the Gmapping SLAM method, is better suited for the analysis of nonlinear systems; however, its computational volume is much larger than that of the extended Kalman filter. Since extended Kalman filter and particle filter have disadvantages for the navigation system, it is proposed to use the Gaussian filter and its modifications. The Gaussian smoothing f...