论文部分内容阅读
对于动态系统进行控制的前提是对于系统的状态进行估计和预报。直接做这项工作往往比较困难,因为系统的数学模型中含有未知时变参数。对于未知的时变参数进行估计已有许多方法。本文采用古典函数逼近方法,即用多项式函数来估计和预报时变参数。这种方法简便易行,如果能取得大量的观测数据,多项式函数的阶数又取得较高,就即保证了时变参数的预报精度,又减少了工作量。如果采用样条函数来逼近,效果会更佳。对于线性及非线性系统本文均做了讨论,并用模拟例子说明了这种方法是实用的。
The premise of controlling the dynamic system is to estimate and forecast the state of the system. It is often difficult to do this work directly, because the system’s mathematical model contains unknown time-varying parameters. There are many ways to estimate unknown time-varying parameters. In this paper, the classical function approximation method, which uses polynomial functions to estimate and predict time-varying parameters. This method is simple and easy. If a large number of observational data can be obtained and the order of the polynomial function gets higher, the forecasting accuracy of time-varying parameters is guaranteed and the workload is reduced. If using spline function to approximate, the effect will be better. This paper discusses both linear and nonlinear systems and shows that the method is practical with simulation examples.