Blind Identification of Nonlinear MIMO System Using Differential Evolution Techniques and Performance Analysis of Its Variants

The present work deals with the nonlinear multiple input multiple output (MIMO) system identification exploring the use of evolutionary computing techniques such as Differential Evolution. The conventionally used standard derivative based identification schemes does not work satisfactorily for nonlinear MIMO systems, which is due to premature settling of weights but the proposed update algorithm works better preventing the premature settling of the model parameters. Simultaneously, the performance comparison of different variants of DE has been demonstrated which reveals the best mutant of DE family that can be implemented into prescribed identification process through the real world applications.