一类非线性系统的鲁棒拟人智能控制

针对一类非线性不确定系统,结合二级倒立摆控制系统设计作为实例,提出一种鲁棒拟人智能控制器设计方法.首先,基于被控对象的物理模型设计拟人定性控制律,采用优化搜索算法得到标称系统的定量控制律;然后在拟人控制律基础上增加一个鲁棒补偿器,用以抑制模型不确定性部分对系统稳定的不利影响、提高系统的鲁棒性.补偿器的设计根据Lyapunov稳定性理论进行,但设计过程中避免了构造具体的Lyapunov函数这一难题.对二级倒立摆系统的稳定控制问题进行了实验验证,实验结果进一步证明了设计方案的有效性.

Robust human-imitating intelligent control for a class of nonlinear systems

A robust human-imitating intelligent control was proposed for nonlinear systems. Firstly, a human-imitating qualitative control law was developed through the analyses of the physical model of the plant. The qualitative control law was quantified by some method of parameter-optimization based on the nominal model of the plant. Then, on the basis of the developed human-imitating controller, a robust compensator was proposed to overcome the bad influences resulted from modeling errors, unmodeled dynamics and disturb- ances. The compensator was derived by Lyapunov stability theory, thus the stability of the closed-loop system was guaranteed. However, the specified Lyapunov function usually essential in a common Lyapunov-based approach was not needed during the design of the robust compensator, as a result, the difficulties in designing the Lyapunov function for a nonlinear system was conquered. Applying the proposed scheme to control a doub- le inverted pendulum, the results of experiment verify the validity of this design.