Approximate Maximum Likelihood Estimation of Time Delay in Deterministic and Stochastic Signal Plus Noise Models

摘要

The problem of estimating signal delay in signal-plus-noise models is considered. A technique is developed for approximating likelihood in such a way that a dynamic programming algorithm may be used to efficiently and recursively estimte delay. The approximation provides an alternative to the implementation of a growing bank of matched filters for exhaustive maximization of likelihood. Simulation results indicate acceptable performance at output signal to noise ratios in excess of 15 dB for deterministic signals and in excess of 20 dB for stochastic signals.