On the simulation of a class of time delay estimation algorithms

The time delay between signals received at two (or more) sensors has proven to be a useful parameter in passive sonar for estimating the location of an acoustic source. This paper presents the results of a simulation comparing the smoothed coherence transform and maximum likelihood estimation methods to the basic cross correlation technique for time delay estimation. Band-limited random signals which are corrupted by white noise and received at two sensors are considered at various signal-to-noise ratios. The variance of the time delay estimates are compared to the minimum variance obtainable in theory as given by the Cramér-Rao lower bound.