針對輪型機器人在系統機動及慣性參數未知下位置穩定問題，雖已有強健控制及 適應性控制設計被提出，但前者常需較高之控制能量(因此可能產生高增益之負面效 應如震盪)，而後者卻展現較差的暫態響應。為此本計畫擬以上一計畫之切換式軌跡 追蹤設計為基礎，提出一切換式指數穩定設計，其切換擬以控制能量或狀態誤差之綜 合函數為依據，在暫態初期適應性控制因較差之參數預測造成較差之響應或控制能量 大於強健控制時作控制權切換，以期達到較節能甚至較佳之暫態響應。此外在系統動 力方程輸入矩陣為正定條件下，我們在上一年計畫完成一種可有效避免適應性控制奇 異點的發生及同時保證漸進穩定的混模控制設計，並已獲得IET CTA 國際期刊接受。 本計畫除進一步納入系統擾動因素外，並放寬輸入矩陣只為非奇異(nonsingular)的 一般性情況，其應用性自然更廣。最後在控制設計完成時，擬將所提設計實踐在一輪 型機器人上，以檢驗其實際可行性。
For the stabilization of wheeled mobile robots with kinematic and/or inertial parametric uncertainties, various schemes based on either the robust or the adaptive tools have been proposed in the literature. However, the robust control usually consumes larger energies (and hence may ignite unwanted high-gain control phenomena such as the chattering) while the adaptive approach exhibits poorer transient performances. For conquering such drawbacks, an exponentially stable switching stabilizer will be developed in this project. In the early stage of transient while the adaptive control consumes larger energy or exhibits poorer performances due to poor parametric estimation, the control authority will be switched to the robust control for better performance or efficiency. Next in the dynamic design stage, we intend to extend the results in our last NSC project, which have been accepted for publication in IET CTA, to more realistic cases with disturbance and whose input matrix being just nonsingular rather than being positive-definite. Finally, the theoretical design will be applied to a real wheeled mobile robot to verify its practical usefulness.
