Deconvolution of sparse underwater acoustic multipath channel with a large time-delay spread

Abstract

The deconvolution of multipath underwater acoustic channel with a large time-delay spread is investigated. The channel deconvolution involves estimating the multipath time-delays and attenuation factors from a noisy received signal consisting of multiple overlapped signals. Similar to conventional deconvolution methods, the proposed method estimates channel impulse response based on least-squares criterion. However, the proposed method harnesses the sparse structure of an underwater acoustic channel, and [script-l](1)-norm of the channel impulse response is adopted as the cost function to be minimized. In addition, the available a priori knowledge of support constraint and attenuation factor constraint are imposed and channel deconvolution problem is converted to a convex optimization problem. Instead of employing the existing standard algorithms, which require huge storage space and high computational complexity, a simple iterative algorithm for solving the optimization problem with fast convergence rate and low complexity is developed. The computational complexity of the proposed algorithm is O(N log(2)(N)) per iteration with N being the length of the received signal. Simulation results confirm that the proposed method provides better performance in terms of temporal resolution and robustness to noise compared with other extant multipath channel deconvolution techniques.