|Robot Tracking in SLAM with Masreliez-Martin Unscented Kalman Filter
Ming Tang, Zhe Chen, and Fuliang Yin*
International Journal of Control, Automation, and Systems, vol. 18, no. 9, pp.2315-2325, 2020
Abstract : The simultaneous localization and mapping (SLAM) is a significant topic in intelligent robot. In this paper, a robot tracking algorithm in SLAM with Masreliez-Martin unscented Kalman filter (MMUKF) is proposed. A robot dynamic model based on SLAM characteristics is first used as state equation to model the robotic movement, and the measurement equations are deduced by linearizing the motion model. Next, the covariance of process noise is estimated with an adaptive factor to improve tracking performance in the MMUKF. Finally, the MMUKF is employed to estimate the positions of robot and landmarks. The proposed algorithm can complete robot tracking with good accuracy, and obtain reliable state estimation in SLAM. Simulation results reveal the validity of the proposed Algorithm.
Adaptive factor, Masreliez-Martin unscented Kalman filter (MMUKF), robot tracking, simultaneous localization and mapping (SLAM).