Novel Maximum Likelihood Estimation of Clock Skew in One-Way Broadcast Time Synchronization

TL;DR

This paper introduces a novel maximum likelihood estimation method for clock skew in one-way broadcast time synchronization, improving accuracy by minimizing delay-induced errors with only two synchronization periods.

Contribution

It presents a new MLE approach based on Gaussian delay assumptions that is independent of clock offset and outperforms existing methods in accuracy.

Findings

Proposed MLE achieves lower estimation error than existing methods.

Requires only two synchronization periods for accurate estimation.

Experimental results validate the superior performance of the proposed approach.

Abstract

Clock skew compensation is essential for accurate time synchronization in wireless networks. However, contemporary clock skew estimation is based on inaccurate transmission time measurement, which makes credible estimation challenging. Based on one-way broadcast synchronization, this study presents a novel maximum likelihood estimation (MLE) with an innovative implementation to minimize the clock skew estimation error caused by delay. A multiple one-way broadcast model is developed for observation and set collection, while the distribution of actual delay is discussed. A clock skew MLE based on Gaussian delay is then proposed and its Cramer-Rao lower bound provided. The developed MLE is independent of the clock offset estimation, and requires only two synchronization periods to generate an accurate estimation. Extensive experimental results indicate that the performance of the proposed MLE is better than the methods popularly used in existing time synchronization protocols.