Abstract

This paper presents a new technology that enables one to locate multiple sound sources with very a large dynamic range simultaneously, including very low frequency and negative signal-to-noise ration sound sources in a non-ideal environment, where there are random background noise and unknown interfering signals. In particular, spatial resolution of source localization is frequency independent. In other words, spatial resolution remains very high at very low as well as at very high frequencies. The underlying principle of this new technology is a hybrid methodology that includes a passive SODAR (nic etection nd anging), advanced signal processing and least-squares minimization. Using this technology, engineers will be able to visualize sound sources in both real time and post processing in an adversary test environment. Live videos of sound sources localization inside a crowd machine shop are shown, where there are unknown background noise, unspecified sound reflections and reverberation, and interfering signals.