【作者】 乔国栋；

【导师】 王江；

【作者基本信息】 天津大学 ， 控制理论与控制工程， 2005， 硕士

【摘要】 传统的基于T-S模型的自适应模糊控制器采用分段线性逼近的思想,每条规则的输出是输入变量的线性组合函数,需要在线调节的后件参数较多,仿真速度很慢,实现较困难。本文提出一种新型的变论域模糊控制器,按照误差的变化,在线调节伸缩因子,连续改变论域的大小,从而动态生成模糊规则,提高了控制器的精度。本文通过设计监督控制器,保证系统的稳定性,获取很好地控制性能。若在变论域模糊控制下,系统不稳定,监督控制将被激活,以确保系统的稳定性;相反,若在模糊控制下,系统已经稳定,监督控制将不会被激活。由于监督控制的存在,闭环系统的全部变量一致有界,从而加快收敛速度,提高了暂态和稳态性能。本文将变论域模糊控制,自适应技术和H_∞最优控制理论相结合对一类模型未知的非线性动态系统进行控制,不仅扩充了H_∞最优控制的智能性,而且还提高了变论域模糊控制的鲁棒性。通过Lyapunov函数推导整个闭环系统的稳定性。所提算法通过选取权重因子r,可将逼近误差和外部扰动对跟踪误差的影响减小到任意给定的标准ρ。本文以混沌系统为例对所提算法进行验证,证明了提出的有效性。更多还原

【Abstract】 The essence thought of the conventional T-S adaptive fuzzy controller ispiecewise linear approximation. The output in every rule is the linear combination ofinput variable. Since all of the consequent parameters needed to be on-line adjusted,the simulation velocity is slow and the realization is difficult. The article proposes anew variable universe fuzzy controller, which can continently adjust the size ofuniverse according to the change of error by on-line tuning the contraction andexpansion factor. Fuzzy rule are dynamically tuned. The control precise is improved.In the paper, a supervise controller is designed to guarantee system stability. Theresulting system is expected to show better performance. If the variable universefuzzy controller works well, the supervisory controller is idle. If the system trends tounstable, the supervisory controller begins to work. All signals involved are bounded;the performance in transient and steady is improved.The variable universe fuzzy controller is combined with H_∞ controller, whichprovide robotic control with intelligence and improve fuzzy control robotic. Thestability is given by Lyapunov function. The proposed algorithm can attenuate theeffect of approximate error and external disturbance on tracking error to given lever ρby selecting weight factor r .In article, the simulation results of chaotic systems are given to confirm theproposed control algorithm is reasonable to practical application.更多还原