Recursive Estimation Method for Bilinear Systems by Using the Hierarchical Identification Principle

This paper considers the parameter identification of bilinear systems with unknown states, which are disturbed by an autoregressive moving average noise. The hierarchial identification principle is employed to derive new algorithms for interactively estimating the states and parameters via a bilinear state observer. However, the general bilinear state-space model involves many parameters, which causes heavy computational burden. Motivated by this fact, we propose a hierarchical generalized extended least squares (HGELS) algorithm by decomposing the original system into a series of subsystems with small dimensions for enhancing computational efficiency. The performance of the proposed algorithms is illustrated through a numerical example.