Comments on "On Clock Synchronization Algorithms for Wireless Sensor Networks Under Unknown Delay"

TL;DR

This paper critiques a previous clock skew estimation method for wireless sensor networks, emphasizing the importance of noise correlation and providing corrected bounds and optimal parameters for improved accuracy.

Contribution

It identifies errors in the previous estimator's derivation due to noise correlation and offers corrected bounds and parameter selection guidelines.

Findings

Corrected the range of gap for estimator validity

Provided revised performance bounds considering noise correlation

Updated optimal gap selection for accurate clock skew estimation

Abstract

The generalization of the maximum-likelihood-like estimator for clock skew by Leng and Wu in the above paper is erroneous because the correlation of the noise components in the model is not taken into account in the derivation of the maximum likelihood estimator, its performance bound, and the optimal selection of the gap between two subtracting time stamps. This comment investigates the issue of noise correlation in the model and provides the range of the gap for which the maximum likelihood estimator and its performance bound are valid and corrects the optimal selection of the gap based on the provided range.