Dual channel speech enhancement using particle swarm optimization

Adaptive processing for canceling noise is a powerful technology for signal processing that can completely remove background noise. In general, various adaptive filter algorithms are used, many of which can lack the stability to handle the convergence rate, the number of filter coefficient variations, and error accuracy within tolerances. Unlike traditional methods, to accomplish these desirable characteristics as well as to efficiently cancel noise, in this paper, the cancelation of noise is formulated as a problem of coefficient optimization, where the particle swarm optimization (PSO) is employed. The PSO is structured to minimize the error by using a very short segment of the corrupted speech. In contrast to the recent and conventional adaptive noise cancellation methods, the simulation results indicate that the proposed algorithm has better capability of noise cancelation. The results show great improvement in signal to noise ratio (SNR) of 96.07 dB and 124.54 dB for finite impulse response (FIR) and infinite impulse response (IIR) adaptive filters respectively.