基於四旋翼無人機經常操作於系統參數(質量、慣量、阻尼係數等)不確定情況下，所謂的沉浸式適應性回步控制設計已蔚為主流解決方案之一。相較於傳統的適應性控制設計，此類設計常展現較佳的暫態響應與強健性。然因此設計當同時出現非參數不確定性時，預測參數收斂性無法確保，明顯限制了其實際應用性。有鑑於此，本計畫主要目的即在建構一四旋翼機之沉浸式強健適應性控制設計以改善此缺點。首先將不確定耦合消去項及所謂σ-修正項納入其中參數更新律，接著證明其所謂的預測誤差最終將收斂於一預設小範圍內，明顯擴大了其實際應用性。最後，在控制設計完成後，本計畫接著將其實踐於一以Arduino為控制核心之四旋翼平台，以驗證其有效性。
Due to the frequent appearance of parametric uncertainty (mass, moment of inertial, viscous coefficient, etc.), the so-called immersion and invariance (I&I) based adaptive backstepping method has become the major approach for the control of quadrotors nowadays. Comparing with the conventional schemes, such an approach exhibits better transients and robustness. Owing to its modular features, it does not guarantee the convergence of the parameter errors in the presence of both the parametric and non-parametric uncertainties simultaneously, which restricts its practical applicability. In this regard, this project aims to construct an I&I based robust adaptive control scheme for quadrotors. In particular, the so-called σ-modification term in the update algorithm is developed for conquering the aforementioned drawbacks. Next, it was proven that the prediction errors of the I&I based schemes will ultimately converge to the residual set around zero with the radius about the size of the upper bound of the non-parametric uncertainty, which apparently enhances its applicability. Finally, the proposed design is implemented on an Arduino-based quadrotor platform to demonstrate its validity.
