Optimal design of FIR high pass filter based on L1 error approximation using real coded genetic algorithm

Abstract

In this paper, an optimal design of linear phase digital finite impulse response (FIR) highpass (HP) filter using the L1-norm based real-coded genetic algorithm (RCGA) is investigated. A novel fitness function based on L1 norm is adopted to enhance the design accuracy. Optimized filter coefficients are obtained by defining the filter objective function in L1 sense using RCGA. Simulation analysis unveils that the performance of the RCGA adopting this fitness function is better in terms of signal attenuation ability of the filter, flatter passband and the convergence rate. Observations are made on the percentage improvement of this algorithm over the gradient-based L1 optimization approach on various factors by a large amount. It is concluded that RCGA leads to the best solution under specified parameters for the FIR filter design on account of slight unnoticeable higher transition width.