Volterra级数模型的非线性压缩测量辨识算法

针对非线性系统的辨识问题,提出了非线性压缩测量辨识算法,且推导出了一种符合压缩感知测量准则的测量模型。相比递归最小二乘法,该方法极大地减少了所需的测量数,使得高阶Volterra级数辨识成为可能。此外,还分析了实际应用中的各项因素对辨识准确性的影响,如信号稀疏度、测量噪声、测量矩阵形式等。

Nonlinear compressed measurement identification based on Volterra series

For the identification problem of nonlinear systems, the accuracy and stability of the nonlinear compression measurement identification algorithm were proved in the simulation experiment, and the complete signal was obtained accurately only by using constant multiple measurement times of the signal sparsity. Compared with the least square method, the proposed algorithm has greatly reduced the needed measurements, therefore, it is possible for the identification of high-order Volterra series. Furthermore, the influence of all factors on the accuracy of system identification was analyzed, such as signal sparsity, measurement noise, measurement matrix form, etc.