IIR filter design using time and frequency responses by genetic algorithm for system identification

The design of adaptive infinite-impulse-response (IIR) filters has attracted increasing attention in recent years that is a beneficial subject in signal processing, control and system identification. In this paper the Genetic Algorithm is used to estimate parameters of high order IIR filters and to show more GA capability for this issue, time-frequency fitness function is employed. In simulations, the results of implementing GA with three different types of fitness functions are compared. These results show that the fitness function of time-frequency increases rapidly convergence of genetic algorithm, as it could be estimated an appropriate model in acceptable time.