摘要：AbstractThis paper presents a method of identifying a two-channel Wiener system, where two parallel linear subsystems are connected through output nonlinearities. First a state-space model of the best linear approximation of the two-channel Wiener system is identified by using a subspace identification method, and two parallel linear subsystems are recovered by utilizing a block-diagonal decomposition of the state-space model. For all realizable two-channel systems, the coefficients of the Legendre polynomial expansion of the nonlinearities are estimated by the least-squares method to find the best configuration of the two-channel Wiener system. Numerical results show the applicability of the proposed method.
