Performance analysis of multiple sparse source localization with triangular pyramid microphone array in noise environment

Direction of arrival (DOA) estimation for multiple speakers using a small microphone array is one challenging research topic and has vast promising applications. One efficient solution is transforming the problem of estimating DOAs of multiple sources simultaneously to estimate DOA of one source at each time-frequency (TF) point by using the sparsity attribute of speech signal. The TPMA-RIPD algorithm proposed in [1] is one multiple sources DOA estimation algorithm following this route, which has shown high DOA estimation accuracy and robustness to additive white noise from the experimental results. In this paper, the robustness of the TPMA-RIPD algorithm is analyzed. The TPMA-RIPD algorithm under the real setup experiments has been evaluated, which further validates the good accuracy and robustness of the proposed algorithm. ? 2012 IEEE.EI