Self-Positioning for Mobile Robot Indoor Navigation Based on Wheel Odometry, Inertia Measurement Unit and Ultra Wideband

GPS signals are often unavailable in indoor scenes where mobile robots often perform tasks. Without the help of global positioning signals, the self-positioning of indoor mobile robots becomes very difficult. In this paper, a self-positioning method for indoor mobile robots is proposed, which combines wheel odometer, inertial navigation unit (IMU) and ultrawideband (UWB). Firstly, through the analysis of each sensor participating in the fusion positioning, the positioning model of each sensor is determined. Secondly, the multi-sensor data fusion method based on extended Kalman filter is proposed so as to improve the overall positioning accuracy and robustness. Besides, in the indoor experimental environment, the feasibility of this method is verified by experiments on mobile robots. Finally, the experimental results show that, compared with the odometer and UWB positioning methods, the proposed method not only improves the positioning accuracy significantly, but also reduces the motion noise in pose estimation.