Acoustic source localization based on compressed sensing and auditory bionics for hearing aids
Ruiyu Liang,Li Zhao,Ji Xi,Xuewu Zhang,Qingyun Wang
DOI: https://doi.org/10.19650/j.cnki.cjsi.2011.06.029
2011-01-01
Abstract:Compared with single microphone, microphone array helps to access acoustic location parameters between the microphone pairs more easily. However, when the space between the microphones in the microphone pairs is too small, it has become very difficult to acquire adequate positional parameters. Aiming at this problem, this paper proposes a signal model, which contains source information and is constructed using differential microphone signals based on the localization ability of Ormia ochracea. Moreover, to decrease computation complexity, we adopt multi-channel compressed sensing (CS) to obtain acoustic source data. Then, according to the constructed CS model, adaptive projection subgradient method is used for signal reconstruction. Finally, acoustic source location is obtained by estimating the energy in the reconstructed signal. Theoretical analysis and simulation results conclude that the proposed approach exhibits a number of advantages over other source localization techniques, which include increased resolution, improved robustness to noise, less computation task, and etc.