 | Prof Hugh HT Liu 劉泓濤教授 多倫多大學航空航太研究所教授兼空中機器人中心主任 |
Biography 簡歷
Hugh H.T. LIU is a Professor of the University of Toronto Institute for Aerospace Studies, where he has been on the faculty since 2000. His research interests in the area of aircraft systems and control include autonomous unmanned systems, cooperative and formation control, fault tolerant control, active control on advanced aircraft systems, as well as integrated modeling and simulations. He currently serves as the Associate Editor of AIAA Journal of Guidance, Control and Dynamics. He is also an Associate Editor of the Canadian Aeronautics and Space Journal. Dr. Liu is a fellow of Canadian Academy of Engineering, a fellow of Engineering Institute of Canada, an Associate Fellow of AIAA The Canadian Aeronautics and Space Institute (CASI), and a Fellow of Canadian Society of Mechanical Engineers, and a registered Professional Engineer in Ontario, Canada. In 2021, Professor Hugh H.T. Liu was awarded the CASI McCurdy Award, presented for his outstanding achievement in the science and creative aspects of engineering relating to aeronautics and space research.
劉泓濤，1991年本科畢業于上海交通大學，1994年碩士北京航空航太大學，1998年多倫多大學獲得博士學位。現任職于多倫多大學航空航太研究所教授，並擔任多大空中機器人中心主任。劉泓濤教授是國際知名學者，研究飛行器系統與控制。他是加拿大工程研究院(Engineering Institute of Canada)院士，加拿大機械工程師學會會士，加拿大航空宇航學會副會士，美國航空航太學會副會士，並擔任若干國際學術期刊編委。2021年劉泓濤教授獲得加拿大航空宇航學會McCurdy Award，表彰他在該領域的卓越成就。2022年劉泓濤教授榮獲加拿大工程院院士。
Title 題目 | UAV Flight Control in the Era of Deep Learning for Advanced Air Mobility 基於深度學習的無人機自主控制：低空經濟的挑戰和機遇 |
Abstract 撮要 | The rapid technological advancement in UAVs (drones), robotics and AI have created potential opportunities for emerging applications. In the meantime, the new challenges are encountered. How much shall we trust the intelligence, especially in safety-critical aviation domain? In this talk, we touch on this topic from a learning-based intelligent flight control perspective to provide our viewpoints. Based on a brief review of AI in UAV control, the discussions focus on two fronts: 1) safe-learning and self-supervised learning; 2) deep learning with hard constraints. The aim is to develop a model-based, learning-enhanced flight control strategy to take advantage of fast progress of AI algorithms yet maintain the safety and validation of control design. We will illustrate our recent study with a case study for advanced air mobility: slung payload control with variable-length cable. 無人機（UAV）、機器人技術與人工智能（AI）的快速發展，正在推動諸多新興應用的出現。然而，這一技術浪潮也帶來了新的挑戰，尤其是在對智能系統信任程度的把握方面——特別是在對安全性要求極高的航空領域。本報告將從基於學習的智慧飛行控制角度，探討這一核心問題。在簡要回顧AI在無人機控制中的應用現狀後，我們將重點圍繞兩個方向展開討論：1. 安全學習（Safe Learning）與自監督學習（Self-Supervised Learning）；2. 帶有硬約束的深度學習（Deep Learning with Hard Constraints）。我們的目標是構建一種基於模型、融合學習的飛行控制策略，在充分利用AI演算法飛速進步的同時，保障控制系統的安全性、可靠性與可驗證性。我們將通過一個**先進空中機動應用（Advanced Air Mobility）**案例來展示該方法的實際效果：基於可變長度纜繩的吊掛載荷控制問題。該案例將展現模型驅動控制與AI技術融合下的協同控制與安全保障能力。 |