Research on Indoor Wireless Positioning Precision Optimization Based on UWB

Abstract

The ultra-wide band (UWB) indoor positioning precision often has large deviations due to environmental influences. To reduce noise error and improve the UWB indoor positioning precision, this paper divides the indoor positioning into static positioning and mobile positioning, and proposes different optimization algorithms for the two positioning modes. An improved self-organizing feature mapping neural network clustering algorithm is used for static positioning. After training, the layout of the neural network is established, and each weight vector is located at the center of the input vector cluster. Experiments indicate that the fully trained neural network can effectively filter noise, and reduce the mean square error significantly within 3.0×10−3. The positioning precision is 32.39% and 17.24% higher than those of the K-mean filtering algorithm and the Kalman filtering algorithm. For mobile positioning, the optimized neural network clustering algorithm is integrated with the unscented Kalman filter (UKF) to smooth the positioning data and reduce non-line-of-sight (NLOS) error. Experiments prove that this method can effectively reduce the errors caused by the NLOS state change, and estimate the distance with a high precision for ultra-wide band positioning and tracking.