Hybrid optimization design of FIR filter

Combined polynomial approximation with particle swarm optimization (PSO), a hybrid optimization design method is presented for improving the finite impulse response (FIR) filter performance which determined by the interpolation point in the transition zone. The method is based on the technique of frequency sampling; and the attenuation of stop-band is taken as an objective function. First, polynomial approximation is used to estimate the search range and the initial position of particles. Then, the improved PSO algorithm is used to find the optimal interpolation point in the transition zone. Design examples such as Low-pass FIR, band-pass FIR, band-stop FIR and bell-shaped IIR filter verified the effectiveness and superiority of the method.