Research on UWB fusion location algorithm

Abstract

In order to solve the problem that ultra-wideband (ultra-wideband, UWB) positioning is greatly affected by the environment in the indoor environment of mobile robot, an indoor positioning algorithm based on UWB, odometer and inertial measurement unit is proposed. Taking advantage of the high positioning accuracy of UWB in indoor environment and the high accuracy of odometer positioning in a small range, the adaptive parameter equation of error is constructed by using linear regression function. The weight coefficient is obtained by combining the position data solved by UWB and the previous moment fusion algorithm to solve the position difference and the mileage displacement formula, and the weight value of angle increment is given according to the state of the system. Through the calculation, the displacement increment matrix of each measurement of the system is obtained, and the final linear optimal value is obtained by substituting the extended Kalman filter algorithm. Compared with the single UWB positioning method, the results show that the system can maintain good accuracy and robustness, and the mean error of the whole system is less than 8cm, which provides a solution for the single UWB positioning of mobile robot affected by NLOS error.