Indoor Cooperative Localization Algorithm Based on UWB/PDR Data Fusion

In recent years, there is an increasing demand for indoor localization, and multi-source data fusion algorithms have gradually become a research hotspot. Ultra-wideband (UWB) technology is widely used for its high accuracy, but the effect decreases significantly in non-line-of-sight (NLOS) environments. Pedestrian Dead Reckoning (PDR) technology estimates step length and direction from sensor data, which has high localization accuracy in a short period of time, but the problem of error accumulation is serious with the increase of time. To address these problems, this paper proposes a nested Extended Kalman Filter (N-EKF) that fuses the ranging information of UWB with the step length and heading angle values of PDR to obtain the pedestrian position coordinates. The experimental results show that the fusion algorithm reduces the average error by 75% and 52.2% compared to the UWB and PDR localization algorithms, which effectively alleviates the excessive UWB ranging error in weak NLOS environments.