中文摘要：

为了辨识一类非线性Hammerstein-Wiener系统，基于递推贝叶斯算法和奇异值分解，提出了一种两阶段在线辨识算法。该算法首先利用递推贝叶斯算法估计乘积项参数，然后利用奇异值分解得到待估计参数。仿真结果表明，所提算法可以以较小的计算量获得精度较高的参数估计值。

英文摘要：

To identify a class of nonlinear Hammerstein-Wiener system, this paper proposed a two-stage online estimation al- gorithm based on recursive Bayesian algorithm and singular value decomposition. The algorithm estimated the products of the parameters firstly, and then obtained the parameters by singular value decomposition. Simulation reveals that the proposed al- gorithm can get high-accuracy estimates with less computation burden.