Estimation of nonlinear systems without a priori structural information and its application in Hammerstein model identification with dynamic nonlinearities

In the paper, we discuss identification of a nonlinear system without structural information and propose two methods, the kernel method and the orthonormal basis method. The convergence results are established for both methods without a priori structural information. We then apply the results to identification of Hammerstein models with an unknown dynamic nonlinearity. It is also shown that identification of the linear part in Hammerstein models is possible with no knowledge of the dynamic nonlinearity.