Decomposition based recursive least squares parameter estimation for input nonlinear equation-error systems

This paper developed a decomposition based recursive least square algorithm for estimating the parameters of an input nonlinear equation-error system, where the nonlinear system was parameterized as a bilinear-parameter system and decomposed it into two subsystems whose parameters were cross-estimated by the least squares methods. The proposed algorithm estimated much less parameters than the over-parameterization identification methods. Simulation results confirm the effectiveness of the proposed algorithm.