Fractional order system identification based on genetic algorithms

System identification deals with estimating the plant parameters under control using input-output measuring data. Most of practical plants have fractional order dynamic properties which are based on integration and differentiation of noninteger order. In this work the structure and the parameters of fractional order unknown transfer function are estimated using input-output data. Integer order Least Squares identification is used first to confirm the structure (order) of the unknown transfer function. Then, Genetic Algorithms (GAs) is followed to find the most accurate fractional order estimate that represents the system. Illustrative examples are presented in which fractional order transfer functions are identified in a way that faithfully estimates the dynamics of the unknown plants