Identification in the delta domain: a unified approach using hybrid FAPS algorithm

Abstract

This paper investigates the application of a firefly based hybrid algorithm, namely FAPS, integrating firefly algorithm (FA) with pattern search (PS), to identify linear dynamic systems in presence of static nonlinearities in the discrete-delta domain. The advantage of using delta operator is to provide unification of continuous-time systems with discrete domain results at a high sampling rate. Two popular identification models, viz. hammerstein and wiener are taken up in this work. The parameters of these models as well as the polynomial nonlinearities considered are calculated using FAPS algorithm, through the minimization of mean square error (MSE) occurring between the true and identified model outputs. Simulations illustrate the efficacy of the proposed technique.