The structure and performance of estimators for real-time estimation of randomly varying time delay

Structures for time-delay estimation are described which are deduced using the maximum likelihood technique. The performance of estimators, designed for time-invariant delays, is discussed for the case of signals with a time-variable delay. A two-step algorithm for the estimation of a randomly varying time delay between two stochastic signals is developed. In the first step, the maximum likelihood estimate is computed over an observation interval small enough to consider the delay to be approximately constant. Due to the short averaging interval, the probability of ambiguous peaks in the estimation function and the variance of the estimate are greatly increased. Therefore, in the second step, a decision-directed, adaptive postfiltering algorithm is presented that effectively suppresses "outliers" and reduces the variance.