In view of the problem of non-line of sight(NLOS)and random error in ultra-wide band(UWB)wireless positioning system in complex environment,a UWB/IMU combined positioning algorithm based on improved Sage-Husa Kalman filter(SHKF)is proposed.First,a boosting tree based on probability density is designed,and the NLOS signals is identified by introducing the probability distribution density of UWB/IMU collected feature data into the loss function of the boosting tree.Then,an improved SHKF algorithm is designed to define an adaptive factor according to the changing trend of innovation,and adjust the strategy of correcting the error of innovation in real time to adjust the influence of historical noise on the current positioning,so as to improve the stability and accuracy of UWB/IMU combined positioning.The experimental results show that the NLOS signal identification accuracy of the proposed method is up to 99.12%,and the root mean square error of positioning is reduced to 4.30 cm,which improves the positioning accuracy of UWB/IMU integrated system in complex environment.
non-line of sightSage-Husa Kalman filterUWB/IMU combined positioningboosting tree
万方数据
维普
