基于Box-Jenkins模型的迭代辨识与控制器设计方法研究

Iterative Identification and Controller Design Method Based on Box-Jenkins Models

针对存在有色噪声干扰的Box-Jenkins系统,采用数学优化的方法,提出了该模型的新型迭代辨识方法。与递推广义增广最小二乘算法相比,论文提出的迭代算法反复利用了系统的所有可测数据信息,因而具有更高的参数估计精度和更快的收敛速度。最后分析了Box-Jenkins模型的低阶迭代控制器设计步骤及方法,通过仿真说明该迭代算法的有效性。

For Box-Jenkins systems with colored noises,the mathematical optimization methods are used to develop a new-type iterative algorithm.Compared with the recursive extended generalized least squares algorithm,the proposed iterative algorithms make sufficient use of all measurable data and thus can produce more accurate parameter estimates and have faster convergence rate.Finally,low-order iterative controller design steps and algorithms are analyzed.Simulation shows that proposed iterative algorithms are reasonable.