Identification of predictor and filter parameters by ARMA methods†
D. J. KRAUSE, Daniel Graupe · International Journal of Control · 1973
The present paper describes procedures for recursive joint identification of the parameters of mixed autoregressive-moving-average models and for the determination of their order. The procedures given are subsequently extended to joint the estimation of the transition and the covariance parameters of single output and multi-output Kalman—Bucy filters in cases where both input and measurement noise sequences exist.