Enhancing Efficiency of Digital Filters via Artificial Intelligence Algorithms

Abstract

The "infinite impulse response" (IIR) digital filter was developed using the hybrid optimization artificial intelligence (AI) algorithm. At AI algorithm D-PSO, which stands for the "particle swarm optimization" dynamic topology, we make use of an approach that is both efficient and simple to memorize. The researchers investigated strategies to maximize the permissible ripples in a passband, the stopband, and the transition band while minimizing the nonlinear mean square error (MSE) between an initial filter response and the required filter response in those bands. Using the approach resulted in a considerable reduction in the square error rates of both the passband and the stopband, according to the findings of an experiment. The IIR digital filter square error values with the D-PSO are superior to those of digital filters using evolutionary AI algorithms such as the dynamic inertia PSO (DI-PSO) or the modified dynamic inertia PSO (MDI-PSO); this allows the IIR digital filter to achieve better results because the square error values depend on the sequence in which the filters are applied. Therefore, compared to digital filters and coefficients created by evolutionary ways, the ones produced by the suggested method are more precise and stable than those produced by evolutionary AI approaches.