Efficient Design of Sparse FIR Digital Filter

Abstract

This paper proposes the design of sparse filter, whose impulse response not only contains real valued coefficients but also some zero amplitude coefficients. The advantage of zero valued coefficients results in reduced the hardware cost while implementing the digital filter. The optimal filter coefficients are computed using hybrid grey wolf and cuckoo search optimization (HGWOCS) utilizing least square error function. Here, hard thresholding has been employed to discard the zero valued coefficients. It is found that, the performance of the designed sparse filter outperforms the equivalent non-sparse filter. The sparse FIR filter produces least error deviation from the ideal filter response.