Timestamp Free Synchronization with Clock Skew Estimation in Wireless Sensor Networks

Time synchronization is crucial for applications such as object tracking and system analysis in Wireless Sensor Networks (WSNs). The existence of clock skew will cause clocks of different nodes to drift at different rates. In this paper, a novel time synchronization scheme combining timestamp-free and one-way communication is proposed, which provides improved clock skew estimation. The proposed scheme has no additional communication overhead and mitigates the power consumption. Based on the proposed scheme, the Window Maximum Likelihood Estimator (WMLE) for clock skew is proposed, which utilizes a minimized subset of observations to estimate clock skew. Cramer-Rao Lower Bound (CRLB) of the WMLE under Gaussian random delay model is derived. The efficiency of the proposed estimator has been validated by simulation results.