Identification of Fractional Linear Dynamical Systems with Autocorrelated Errors in Variables by Generalized Instrumental Variables

This paper addresses the problem of identifying stochastic linear fractional discrete-time systems from noisy input/output data. The input noise is supposed to be fractional noise, while the output noise is assumed colored or fractional noise. The paper also proposes an estimator based on the generalized instrumental variables estimator. We have compared the proposed estimator with other estimators with the help of Monte Carlo simulations. The results of the simulated examples indicate that the proposed estimator provides good accuracy.