Indoor Localization for Mobile Robots Using Odometry and Vision System: A Pseudo Measurement Approach

摘要

This paper studies the problem of indoor localization for mobile robots using odometry and vision system. As the odometry adopts a higher sampling rate than the camera, a pseudo measurement approach is proposed to recover the visual measurements such that the position of the mobile robot can be estimated by the visual measurements with the same sampling rate of the odometry. To improve the robustness with respect to measurement variations, a self-tuning parameter is introduced in the pseudo measurement and the normalized least mean square algorithm is used to adjust the self-tuning parameter adaptively. By formulating the localization problem in the framework of Bayesian estimation, the unscented Kalman filter is utilized to handle the nonlinear filtering problem. Experimental results are provided to illustrate the effectiveness of the proposed approach.