Research on Indoor UWB Positioning Based on Expectation Maximization in NLOS Environment

Abstract

SummaryUltra‐wide band (UWB) technology is especially suitable for indoor positioning because of its strong anti‐interference ability, it can achieve centimeter‐level positioning accuracy in the line of sight (LOS) environment, but it is challenging to construct a robust and high‐precision algorithm in indoor complex and changeable environments. Therefore, it establishes an improved standard Kalman filter (SKF) based on expectation maximization (SKF‐EM), which models the prediction error covariance and measurement noise using the maximum likelihood criterion, estimates the filter parameters online through the expectation maximum theory, and updates the Kalman filter gain according to the Kalman filter based on dropped measurements method (SKF‐DMM). Experimental results show that: in case of serious occlusion, the maximum positioning error of SKF‐EM algorithm is 1.03 m, 39% lower than that of SKF and 60.2% lower than that of UWB.