Implementation of multi swarm PSO algorithm for ripples reduction in digital fir low pass filter

*Uday Kumar P and Kaladhara Sarma GRC

Research Scholar in Electronics and Communication Engineering, Shri Venkateshwara University, Gajraula, Amroha - 244236, U.P., India

*Corresponding Author E-mail:

Received June 10, 2013; Accepted June 25, 2013


Engineers seek to design a digital FIR filter in such a way preferably to avoid the effect of errors in Signal Processing. There are several methods to design a FIR filter but a problem exists to determine the sample value in the transition zone. Obviously, the optimization techniques such as Ant colony algorithm, Bees algorithm, Firefly algorithm, and Genetic Algorithm method cannot guarantee that the interpolator is the optimal sampling point. This method is complex in structure which takes longer time in operation and suffers from local optimal solutions. In this paper multiple swarm with multiple elements of PSO algorithm is used to determine the frequency response of Digital FIR low pass filter, which provides  optimal filter coefficients  with fast convergence speed such that error function is minimized, when compared the errors obtained from  windowing technique. PSO algorithm is one of the efficient ways to improve the stop band attenuation such that the samples are interpolated near the discontinuity and reduces the errors. The performance of this PSO algorithm is compared with the conventional window techniques has been verified via computer simulations using MATLAB 7.0.      

Keywords: PSO, MSE, filter ripples, multiple swarm and window techniques.