Simultaneous multiple bands time delay and frequency component estimation based on expectation–maximization

Abstract

Acoustic sensor systems often operate in complex environments containing multiple sound and noise sources. This paper proposes a new multi-band multi-source time delay estimation and frequency analysis algorithm based on the Expectation–Maximization (EM) method. The algorithm improves the multi-source signal superposition model and introduces the EM method to iteratively find parameters for each frequency band. The algorithm also provides multiple time delay estimates while simultaneously estimating the intensity distribution of the frequency bands for each of the corresponding time delays. The distributions are valuable for subsequent tasks such as target classification and identification. The performance of the algorithm is assessed by simulations and field experiments.