Design of an Fir Low Pass Filter Using Bare Bones Particle Swarm Optimization

Design of an FIR filter as per expected response is a field of wide interest. Now a day evolutionary algorithm are in lime light for filter design problem. Among the various algorithms bare bones particle swarm optimization (BBPSO) has capability to perform operation in multidimensional space with less number of control parameter, this make the algorithm simple and it is applied by simple computer code. In this work design of linear phase low pass FIR filter is presented. The BBPSO is modified form of canonical PSO by removing the velocity term of the canonical PSO and replaced by Gaussian sampling strategy. To investigate the behavior of BBPSO based filter design approach, experiments are performed on standard benchmark test functions and filter design problem. Two cases of filter design with filter order 20 and 30 have been realized using BBPSO based solution approach. This work considers a fitness function based on the mean squared error between the actual and the ideal filter response. In order to compare the performance, the PSO algorithm is also simulated. A comparison of simulation results reveals the optimization efficacy of the BBPSO algorithm over the PSO optimization techniques for the solution of the multimodal, non-differentiable, highly non-linear FIR filter design problems.