Efficiency of Stochastic Algorithm for Different Target Functions in the Task of Estimating Radio Pulse Time Shift

A comparative analysis of efficiency of different target functions for stochastic algorithms in estimating time shift between unfiltered radio pulses received on spatially spaced receivers is carried out. As target functions, we investigated measures of signals similarity that are heterogeneous in their properties. It includes mean square difference, correlation coefficient, and mutual information. Note that two approaches were used to estimate mutual information. In the first case, the entropy of samples was calculated from histograms, and in the second one, using Parzen windows. The developed algorithms were tested on simulated and real signals. Algorithms showed high efficiency.