Robot Tracking in SLAM with Masreliez-Martin Unscented Kalman Filter
- PDF / 1,400,154 Bytes
- 11 Pages / 594.77 x 793.026 pts Page_size
- 47 Downloads / 203 Views
ISSN:1598-6446 eISSN:2005-4092 http://www.springer.com/12555
Robot Tracking in SLAM with Masreliez-Martin Unscented Kalman Filter Ming Tang, Zhe Chen, and Fuliang Yin* 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. Keywords: Adaptive factor, Masreliez-Martin unscented Kalman filter (MMUKF), robot tracking, simultaneous localization and mapping (SLAM).
1.
INTRODUCTION
Robot localization and tracking have played an important role in many application scenarios, such as automatic drive, augmented reality and virtual reality [1–4]. Traditional localization algorithms may become increasingly inaccurate over time in the absence of any global position information, which makes the position estimation of robot unreliable [5, 6]. Alternatively, some robot tracking algorithms in simultaneous localization and mapping (SLAM) have been developed [7]. The robot tracking algorithms in SLAM can be employed in the unknown environment [8, 9], where the robot can track itself and create a continuous environmental map [3]. In some environments, the spurious observations due to noise could even conceal those of the true robot, causing poor location performance. In order to solve the robot positioning inaccuracy problem, the Kalman filter (KF) based SLAM algorithm is used to track robot, where robot position and map are uncertainty [10]. In the KF, the system state equation is used to predict the state value and the observation model is applied to update state, which completes the robot tracking. The KF is the optimal estimator when the models of state and observation are linear. However, the KF may emerge poor tracking effect or even
divergence when the observation model is nonlinear. In order to deal with the problems of the Kalman filter, Smith et al. [11] used the extended Kalman filter (EKF) for robot tracking in SLAM to estimate robot pose. The nonlinear system is transformed into linear system by the Taylor series expansion. The EKF based SLAM algorithm can produce good tracking effect, and it has been employed in some scenarios [12, 13]. Moutarlier and Chatila [12] proposed the basic framework of SLAM system by the EKF with a laser range finder to complete robot tracking. Davison [13] exploited the EKF for the first monocular vision SLAM system to estimate robot pose. Nevertheless, the EKF c
Data Loading...
