【作者】 吕英从；

【导师】 王平；

【作者基本信息】 河海大学 ， 控制理论与控制工程， 2004， 硕士

【摘要】 模糊控制器的解析结构对模糊控制系统的设计有着十分重要的指导作用，模糊控制系统中有关论域变换、模糊推理等基本环节的设计，对控制效果有着十分明显的影响。 针对模糊控制系统中存在控制死区的问题(尤其是采用梯形隶属函数)，本文针对典型的二维模糊控制器，分别采用Sum-Min、Max-Min、Max-Product推理方法对其解析结构作了分析和比较，研究了采用不同隶属函数时的结构特点，揭示了其和PID控制器的内在联系，为模糊控制系统的设计提供了一定的理论依据。 对Mamdani系统，分析了隶属函数分别采用梯形和高斯型时一维模糊系统与函数插值的等价性和逼近误差。数字仿真结果表明，引入插值算法的模糊控制器可以消除量化死区和稳态时的小幅震颤，和常规PID控制器相比，对系统参数和结构的变化有更好的适应性。更多还原

【Abstract】 Structure analysis of typical fuzzy controller can offer important guidance for designing fuzzy controller. The fundamental aspects of fuzzy controller systems, such as interval transforms and inference methods, have obvious influence to the control performance.Dead-zone often exists in fuzzy control system especially when trapezoid membership functions were used. Based on Mamdani’s sum-min, max-min, max-product inference operator, the structure characteristics of the three classes of fuzzy controllers are investigated and compared. The different characteristics of fuzzy controllers using different membership functions are also studied. Finally, the internal relation between PID controller and fuzzy controllers are also discussed. These provide some theoretic bases for designing fuzzy control systems.The equivalence between the Mamdani type fuzzy systems using trapezoid or gauss membership functions and the piecewise interpolation functions is demonstrated. And the estimation of the approximation error is given. Computer simulation results make clear that fuzzy controller using interpolation arithmetic has better control performance than the traditional one because it can eliminate quantization dead-zone and adjust steady-state errors completely. Compared with PID controller, it suits for the systems with changeful parameters and structures much better than PID controller.更多还原