A comparison between spline-based and phase-domain time-delay estimators.

Abstract

We previously presented a novel time-delay estimator that uses cubic splines to estimate continuous time delays from sampled data. Previous simulation results showed that the spline-based algorithm yields superior performance when compared to the normalized correlation and the sum squared differences (SSD) algorithms implementing a variety of subsample interpolators. In this short communication, we present additional simulation results that compare the spline-based time-delay estimator to well-known phase-domain estimators. Under comparable conditions, spline-based methods exhibit biases approximately one tenth and standard deviation only one hundredth of those observed for the phase-based approaches.