Gradient-based iterative parameter estimation for Box-Jenkins systems with finite measurement data

A gradient-based iterative (GI) identification algorithm is developed for Box-Jenkins systems (or models) with finite measurement input-output data. Compared with the pseudo-linear regression stochastic gradient approach, the proposed algorithm updates the parameter estimation using all the available data at each iterative computation (at each iteration), and thus can produce highly accurate parameter estimation. An example is included.