Identification of MIMO Hammerstein-Wiener System

A new approach to identification of multi-input multi-output (MIMO) Hammerstein-Wiener system is presented. The output nonlinear block consists of several single-input single-output (SISO) blocks, one of which is dead zone and saturation nonlinearity. The hinging hyperplane (HH) model expresses the character. The MIMO input nonlinear block is described by multi layer feed forward neural networks. The transfer function matrix indicates the MIMO linear dynamics block. According to the prior structure knowledge, the identification problem is transformed to constrained optimization using prediction error method (PEM). The interior-point method (IPM) is adopted to solve the nonlinear programming. Finally, the simulation examples illustrate the performance and validate the effectiveness of the proposed algorithm.