A Method of Multi-Channel Non-Uniform Subband Adaptive Noise Cancellation

Abstract

Unmanned Underwater Vehicle (UUV) self-generated noise is the primary factor that affects the performance of passive flank array in the detection of weak signal, which is a complex mechanism and physics process, demonstrated broadband, colored spectral property in frequency and spatial domain. In practice, adaptive noise cancellation (ANC) is usually used. Since the colored distribution of self-generated noise, the key issue to improve the capabilities of ANC is how to acquire the reference signals and use subband ANC in the system. As uniform sub-bandwidth and critical subsampling adopted, the aliasing in-band limit the improvement of convergence performance. In the paper, a method of multi-channel non-uniform subband ANC is proposed, in which the multi-channel reference signal is obtained through accurate delay and difference the outputs of hydrophones, ANC can be implemented in non-uniform subband according the distribution of power spectral density of self-generated noise. In avoiding the aliasing in-band, DFT based subband decomposition method was used. The results of experiment show that the proposed method has better performance than traditional ones.