目标函数能自动在线调整的自校正控制器

本文论述了一种目标函数能自动在线调整的新型自校正控制器。这种自校正控制器既保留了最小方差控制器能达到渐近最优性能指际、算法简单、在线计算量小的优点,又采用了极点配置设计方法,以确保闭环系统稳定性和要求的动态响应。因此它兼有最小方差和极点配置两者的优点,使确定目标函数权多项式这个难题迎刃而解。 该自校正控制器首先通过广义最小二乘辨识,获得对象的参数a_i和b_i,然后由极点配置求得目标函数的权多项式P,Q,R,并建立一个算法,使P,Q,R随对象参数变化而自动地在线调整。数字仿真表明,这种控制器特别适用于具有大惯性和慢时变的对象,在工业控制中具有广泛的应用前景。

Self-Tuning Controllers with Objective Function Adjustable On Line Automatically

This paper discusses a new type of self-tuning controller with its objective function adjustable on-line automatically. This self-tuning controller keeps the advantages of the minimum-variance controller, i.e.it has successive approximation to the optimal performance, simple algorithm and small on-line calculating quantity. Moreover, the design method of pole assignment is adopted to guarantee the stability of the closed loop system and the required dynamic response. The self-tuning controller, therefore, has both advantages of minimum-variance and pole assignment, and if also solves the difficult problem of determining the weight polynomial of the objective function.The object parameters ai and bi of the self-tuning controller are obtained through the generalized least square identification, and then the weight polynomials P,Q,R of the objective function are obtained through pole assignment. An algorithm is also established to adjust P,Q,R automatically on-line with the change of the object parameters. Digital simulation suggests that the controller fits particutarly to the object having big inertia and slow time variation. The controller has wide applications in the industrial control.