國立臺灣師範大學電機工程學系王偉彥黃義盛2014-10-302014-10-302012-07-31http://rportal.lib.ntnu.edu.tw/handle/20.500.12235/32074本計畫中，我們針對某一類主動式移動機器人系統，提出了一個使用混合式直接/間接適應性派屈模糊類神經網路之線上路徑追蹤控制器設計。此新穎的方法包含控制目的的決定、近似器的架構、系統動態的模組化、線上控制演算法的發展與系統穩定性的分析。根據模型資訊與控制資訊的重要性，我們使用一個權重因子去結合直接型與間接型的適應性派屈模糊類神經網路控制器。因此，控制器的設計方法在設計的過程中將更加有彈性。此外，我們使用李普諾夫理論去確認整個系統的穩定性。模擬的結果驗證了我們所提出的方法可以達到適合的追蹤效能。In this project, we proposed a path-tracking controller design for a class of autonomous mobile robot system using hybrid direct/indirect adaptive Petri-fuzzy-neural network. This novel approach consists of control objectives determination, approximator configuration design, system dynamics modeling, online control algorithm development, and system stability analysis. According to the importance and viability of plant knowledge and control knowledge, a weighting factor is utilized to sum together the direct and indirect adaptive FNN controllers. Therefore, the controller design methodology is more flexible during the design process. Moreover, the stability of the whole system can be verified by using Lyapunov theory. Simulation results illustrate that the proposed method can achieve favorable tracking performance.派屈模糊類神經網路主動式移動機器人路徑追蹤控制Petri-fuzzy-neural networkautonomous mobile robotpath-tracking control