Relative Localization Algorithm for UWB/IMU Loose Combination Unmanned Ground Vehicles Based on EKF Model Compensation

Aiming at the problems of high localization error and weak adaptability to environmental conditions of traditional unmanned ground vehicle (UGV) relative localization algorithms, a loose combined ultra-wideband and inertial measurement unit (UWB/IMU) localization algorithm based on the compensation of extended Kalman filter (EKF) model is proposed. Firstly, the effects of ranging noise and Angle of Arrival (AOA) noise on localization accuracy are fully considered, and an iterative constraint-weighted Time Difference of Arrival/Angle of Arrival (TDDAI ADA) joint localization model is constructed, so that the system can maintain high localization stability under different noise conditions. Then, the EKF model is used to fix and compensate for the errors generated during localization, so as to further improve the relative localization accuracy. Finally, the effectiveness of the proposed algorithm is verified by algorithm simulation experiments and actual scenario analysis experiments. The results show that compared with the traditional relative localization algorithm, the proposed algorithm reduces the root-mean-square error (RMSE) of localization in X-axis and Y-axis directions by 48.3% and 54.5%, respectively, in the case of unmanned ground vehicles in dynamic motion.