Volterra 泛函级数在非线性系统辨识中的应用

利用Volterra泛函级数描述非线性系统，将非线性系统辨识问题转化为标准的最小二乘问题，并通过QR分解进行求解，在对输入矩阵P进行Householder变换过程中，提出利用输入向量对输出向量的影响因子进行P阵的列选择，同时完成模型辨识，有效地克服了Volterra泛函数辨识中维数灾难问题，数字仿真表明了该方法的有效性。

Application of Volterra series in the identification of nonlinear systems

By applying the Volterra series to describe the nonlinear system, the identification problem of nonlinear system is transformed into a normal least mean squares problem, which can be solved by QR decomposition. In the process of decomposing the input matrix by the Householder transformation by using of an effect factor between the input vector and the output vector, the selection of row vectors in the input matrix is processed, as well as the structure identification of system is presented, which overcomes the dimension disaster in the identification of Volterra series. The algorithm is validated by the simulation results.