Fuzzy Model Based Tracking Control with H-infinity Performance

• Main Authors: Wei-Chi Lin, 林韋志
• Other Authors: Kuang-Yow Lian
• Format: Others
• Language: en_US
• Published: 2001
• Online Access: http://ndltd.ncl.edu.tw/handle/99843381112734941171

碩士 === 中原大學 === 電機工程研究所 === 89 === abstract There has recently been a rapid growing interest in using T-S fuzzy model to approximate nonlinear systems. The T-S fuzzy dynamic model, which originates from Takagi and Sugeno, is described by fuzzy IF-THEN rules in which the consequent parts represent local linear models. Once a fuzzy representation of a nonlinear system is described by if-then rules, the control problem then becomes to find a local linear/nonlinear compensator to achieve the desired objective. The controller and observer design of the T-S fuzzy system is carried out via the so called parallel distributed compensation (PDC) approach. The stability analysis and controller synthesis are then systematically formulated into linear matrix inequalities (LMIs). The LMI problem can be solved very efficiently by convex optimization techniques. Tracking control design is also an important issue for practical applications, such as robotic tracking control, missile tracking control, and attitude tracking control of aircraft. In this thesis, we will study the tracking control problem of a nonlinear system by the T-S fuzzy model. Based on the nonlinear $H_infty $ criterion, the controller is designed to cope with the tracking control problem for a nonlinear continuous-time systems. Then, we know that measuring full states can be difficult and costly. Moreover, sensors are often subject to noise. An observer is designed to estimate the immeasurable states. It has long been known that descriptor systems have a tighter representations for a wider class of systems in comparison to traditional state-space representation. Recently this concept has been extended to T-S fuzzy model systems. On the other hand, if the fuzzy descriptor system is used, the number fuzzy rules will be decreased. This rule reduction is an important issue for LMI-based control synthesis since larger number of LMI rules may lead to infeasible problems. Finally, a real time control of parallel parking is presented in this thesis. We utilize the robust tracking controller to realize parallel parking. Based on the DS1102 and SIMULINK, we can easily use build-up blocks to establish our experimental environment. Our experimental result verifies the theoretical derivation in this thesis.