An UWB Positioning Algorithm Based on Clock Compensation and Filtering

Single-Sided Two-Way Ranging (SS-TWR) is a classic method used for UWB ranging, but it suffers from errors caused by clock asynchrony and measurement noise. Therefore, this paper proposes a solution based on clock compensation and filtering. Firstly, a Markov model is employed to compensate for clock offset-induced errors by adjusting the Time of Flight (ToF) on a frame-by-frame basis. Then, a Kalman particle joint filtering method is used to eliminate measurement noise. Next, a random forest model is built using estimated distance values from simulation experiments. The model is used to predict new estimated distances based on measured values of SS-TWR and correct the coordinate values of Time of Arrival (TOA) positioning. Finally, on one hand, the performance of the ranging algorithm was evaluated by establishing a dynamic distance measurement simulation scenario for road vehicles. The simulation results show that the average error of this ranging method is reduced by about 4.85cm compared to SS-TWR. On the other hand, an indoor positioning environment was set up using four UWB modules to evaluate the performance of the positioning system in a real-world scenario. Analysis of the experimental results shows that the average positioning error of this positioning system is reduced by about 34% compared to TOA positioning based on SS-TWR.