Multiple-input single-output nonlinear system identification using Bezier- Bernstein polynomials with noise cancellation

This article deals with an identification method for the fractional multiple-input single-output ‎model. It is considered the Hammerstein model to separate dynamic linear and static nonlinear ‎behaviors. Which Bezier-Bernstein polynomials are used to approximate the nonlinear functions ‎and the fractional order transfer function is applied to estimate the linear part. A hybrid ‎identification method based on a modified evolutionary algorithm and a recursive classic method ‎is presented. As an advantage, this method can also correctly identify the system in the presence ‎of output noise. A photovoltaic experimental system and a numerical example are used to ‎illustrate the efficiency and performance of the proposed scheme.‎