论文部分内容阅读
对于未知时延的多输入单输出(MISO)系统,借助分离性原理,推导出迭代的可分离的非线性最小二乘(SNLS)辨识方法.为降低收敛于局部最小的可能性,利用全局优化理论,推导了全局可分离的非线性最小二乘(GSNLS)辨识方法;为消除强观测噪声所引起的参数估计的偏差,将GSNLS方法调整为一新颖的全局可分离的非线性多新息递推最小二乘(GSNMIRLS)辨识方法,仿真实验验证了算法的有效性.
For the MISO system with unknown delay, an iterative and separable nonlinear least squares (SNLS) method is derived by means of the separation principle.In order to reduce the possibility of convergence to a local minimum, a global optimization Theory, a globally separable nonlinear least squares (GSNLS) method is derived. In order to eliminate the bias of parameter estimation caused by strong observational noise, the GSNLS method is adjusted to a novel global separable nonlinear multi-modal Push the least squares (GSNMIRLS) identification method, the simulation experiments verify the effectiveness of the algorithm.