Performance of swarm based optimization techniques for designing digital FIR filter: A comparative study

Abstract

In this paper, a linear phase FIR filter is designed through recently proposed nature inspired optimization algorithm known as Cuckoo search (CS). A comparative study of Cuckoo search (CS), particle swarm optimization (PSO) and artificial bee colony (ABC) nature inspired optimization methods in the field of linear phase FIR filter design is also presented. For this purpose, an improved L1 weighted error function is formulated in frequency domain, and minimized through CS, PSO and ABC respectively. The error or objective function has a controlling parameter wt which controls the amount of ripple in the desired band of frequency. The performance of FIR filter is examined through three key parameters; Maximum Pass Band Ripple (MPR), Maximum Stopband Ripple (MSR) and Stopband Attenuation (As). Comparative study and the simulation results reveal that the designed filter with CS gives better performance in terms of Maximum Stopband Ripple (MSR), and Stopband Attenuation (As) for low order filter design, and for higher order it also gives better performance in term of Maximum Passband Ripple (MPR). Superiority of the proposed technique is also shown through comparison with other recently proposed methods.