Acoustic localization of distributed coherent and incoherent sources using SEMWAN with subarray smoothing

Abstract

Accurate estimation of the location and level of remote acoustic sources from recorded acoustic signals is attractive in many applications. In wind tunnel tests, where noise sources are commonly distributed and incoherent, high resolution array signal processing techniques like the spectral estimation method with additive noise (SEMWAN) have been useful for source localization when background noise measurements are available. However, for many continuously distributed systems, such as a simple vibrating plate, the assumption of incoherent sources is incorrect, and techniques like SEMWAN may yield spurious results. In this presentation, results are reported for the use of SEMWAN alongside a subarray smoothing technique to formulate the coherent source localization problem as an incoherent source localization problem. Simulations comparing localization performance for distributed coherent and incoherent sources are shown. Results from a proof-of-concept experiment using multiple sources and a 15-element linear receiver array are also evaluated against simulation. Performance comparisons are made between SEMWAN, MUSIC, and conventional beamforming techniques in addition to showing the effects of subarray smoothing. [Sponsored by NAVSEA through the NEEC and by the U.S. DoD through an NDSEG Fellowship.]