A Novel Finite-Wordlength Particle Swarm Optimization Technique for FRM IIR Digital Filters

A novel technique is presented for finite-wordlength (FW) particle swarm optimization (PSO) of BIBO stable FRM digital filters incorporating bilinear-LDI IIR interpolation subfilters. A novel LUT scheme is developed to ensure that the FWPSO automatically searches over permissible FW multiplier coefficient values only in the course of optimization. The salient feature of the proposed LUT scheme is that unlike the conventional PSO, there is no need to limit the search space in the course of optimization to prevent going over the boundaries of the search space. This is achieved by introducing barren layers in the LUTs. The usefulness of the proposed FWPSO is illustrated through its application to the design and simultaneous magnitude and group-delay optimization of a lowpass IIR-based FRM digital filter.