Bayesian Measurement Noise Estimation in Ultra Wide Band Systems for Clock Synchronization

Abstract

Establishing clock synchronization between devices using wireless links is a common problem in various application areas, like IEEE 1588 wireless time synchronization, or in time-difference of arrival (TDoA) systems. While the estimation of clock parameters are widely solved by linear filtering methods, in dense indoor environments, the estimation of the measurement noise is an important task. The paper proposes a method to show that the estimation of the measurement noise in ultra wide band (UWB) systems makes it possible to reach more accurate frequency drift estimation, especially in case of harsh multipath propagation. The proposed algorithm is validated on real measurements, and is compared to the state-of-the-art method based on Kalman-filtering.