Research on UWB/MEMS IMU tight combination adaptive robust Kalman filtering localization algorithm

In the ultra-wide band (UWB) and micro-electro-mechanical system (MEMS) inertial measurement unit (IMU) tight-combination filtering system, the existence of large non-line-of-sight (NLOS) errors in the UWB ranging signals and the anomalies in the filtering system are prone to cause serious discrepancies in the probability distributions of the noise parameter and the real noise, which leads to dispersion or even collapse of the entire filtering system, and seriously affects the performance of the tightly-combined system. The UWB/MEMS IMU tight combination adaptive robust extended Kalman filter (AREKF) method is proposed to give full play to the short-time high-precision characteristics of the INS to detect anomalies in the UWB observations, and utilize the innovation sequences to adaptively adjust the covariance matrices of the observation noise and the prediction state error, to reduce the impact of the anomalous observation or NLOS error, and effectively prevent the filter from over-convergence, dispersion, and collapse phenomena. Dynamic pedestrian experimental results show that the UWB/MEMS IMU tight combination AREKF method has RMS better than 0.3 m in the N and E directions and 1.3 m in the U direction, which effectively improves the accuracy, stability and reliability of the positioning system compared with the UWB WLS, the UWB EKF, the UWB/MEMS IMU loose combination EKF and the UWB/MEMS IMU tight combination EKF.