Some improvements of fuzzy logic controllers with applications
Student thesis: Master's Thesis
Related Research Unit(s)
As one of the well-known intelligent control methods, fuzzy logic control has prominent advantage to be model-free in its design and applications. Although relatively mature fuzzy control schemes already exist in the literature, there are still rooms for improvement and there are still technical problems to be solved. This thesis aims to carry forward some of these remaining tasks in the field. More precisely, the main objectives of this thesis research are to design a new fuzzy logic control scheme and apply it to some control problems to obtain better results, search for more effective fuzzy PID control methods for achieving better performances, and design a fuzzy PI observer for noise suppression of some controlled synchronization systems. Firstly, a new defuzzification method of a fuzzy controller, truly based on logical inference, is designed and applied to a model-airplane landing problem, a truck parking problem, an inverted pendulum system, a double inverted pendulum system, and a torsion pendulum control problem. In the new method, the defuzzification weight of a commonly used fuzzy controller is replaced by the implication logic of the fuzzy rules implemented in the controller, taking advantage of the inference logic over the conventional fuzzy set-based approach. Simulation results show that the new fuzzy controller can achieve shorter settling time than the conventional fuzzy controllers for those control tasks that rely more on the inference logic implemented in the rules. Secondly, a fuzzy-plus-saturated PID controller is designed and applied to some lower-order systems. This new controller uses a saturated function to pre-regulate the tracking error signal before the signal is fed into the fuzzy PID controller. By adding a couple of more degrees of freedom into the controller, still in a very simple form of design, better control results can be obtained. As the fuzzy PID controller itself has simple rules, the whole fuzzy-plus-saturated PID controller is less complex than most existing fuzzy PID controllers, but can achieve better control performances. Finally, a fuzzy PI observer-based chaos synchronization scheme is designed. By using a fuzzy PI observer to process the communication signal from the chaotic drive system before it is transmitted to the response system, the noise effects existing in the transmission channel can be significantly reduced. Simulation results show that the fuzzy PI observer works better than the conventional integral observer for noise suppression.
- Intelligent control systems, Fuzzy systems, Fuzzy logic