Interactive two-stage recursive least squares identification algorithms for controlled autoregressive systems

For a stochastic system described by the controlled autoregressive model, this paper gives an interactive two-stage recursive least squares (2S-RLS) algorithm based on the hierarchical identification principle by separating the system parameters of the controlled autoregressive model. Differently from the traditional recursive least squares algorithm, the proposed 2S-RLS algorithm does not require computing the covariance matrix of large sizes and can reduce the computational burden. A numerical example illustrates the efficiency of the proposed 2S-RLS identification algorithm.