Indoor Wheeled Robot Positioning Algorithm Based on Extended Kalman Filter

Abstract

The indoor wheeled robot is widely used in research, industrial manufacturing, and service industries. For the positioning process of indoor wheeled mobile robots, the data from a single sensor is not reliable and accurate. The traditional solution to this problem is to use the extended Kalman filter (EKF) method, which suffers from linearization error and accumulation error. To tackle these problems, we propose Linear transformation error elimination extended Kalman filter(TEKF) to fuse multiple sensors. Firstly, the data of the sensors of the odometer, Inertial measurement unit(IMU) and lidar are collected and preprocessed, and a complementary filtering method is proposed to obtain the angular velocity. Secondly, the second-order Taylor series expansion is performed on the state and the observation equation, which overcomes the linearization error and improves the accuracy of data fusion. Finally, the backtracking processing method is adopted to eliminate the accumulated error and enhance the environmental adaptability. The experimental results of the real indoor wheeled robot shows that TEKF can effectively improve the accuracy of data fusion and ensure that the indoor wheeled robot can be more accurately positioned.