An optimization technique for QMF based on modified particle swarm

In this paper, a new optimization technique for Quadrature Mirror Filter (QMF) based on Modified Particle Swarm (MPSO) is proposed for the design of a two channel linear phase bank in frequency domain. The proposed technique provides a low pass prototype filter characteristics which are close to an ideal QMF design. Using proposed technique, we have evaluated various performance parameters such as pass band error, stop band error, transition band error, PRE, and amplitude variation of QMF. Our simulation results show that 23% reduction in amplitude error, 60% improvement in pass band error, 74% decrease in stop band error, 78% reduction in transition band error, and 24% improvement in PRE can be achieved using proposed algorithm over the other reported techniques in the literature. The design results of the proposed MPSO technique presented in this paper clearly indicate substantial improvements over earlier reported results. This technique can be effectively applied in multi-rate DSP applications.