Based on the problem that the generating method of random array structure is inefficient, a method is proposed to generate the random target arrays by using coaxial circu- lar array in the polar coordinates in the premise that the array angular resolution of source identification is guaranteed. According to the principle of moving sound source identification, this work deduces the basic non-equidistance coaxial circular rings＇ radius, and generates target random arrays which were suitable for moving sound source identification through array partitioning, condition filtering in the polar coordinates and simulation evaluation. Finally, numerical simulation and moving car sound source identification test have been done. The analytical results show that using this method to generate random array is effective. Compared with the traditional regular arrays, the target random array has more accurate moving sound source identification performance.
A new time-frequency analysis method is proposed in this study using a multi-rate signal decomposition technique for the analysis of non-stationary signals. The method uses a multi-rate filter bank for an improved non-stationary signal decomposition treatment, and uses the Wigner-Ville distribution(WVD) analysis for signal reconstruction. The method presented in this study can effectively resolves the time and frequency resolution issue for non-stationary signal analysis and the cross-term issue typically encountered in time-frequency analysis.The feasibility and accuracy of the proposed method are evaluated and verified in a numerical simulation.
