报告题目：Adaptive Approximation-Based Control of Non-Canonical Nonlinear Systems
主讲人：Gang Tao (美国弗吉尼亚大学教授、长江学者讲座教授、千人计划、IEEE Fellow)
时间：2017年12月29日 (星期五) 上午10:00
个人简介：Professor Gang Tao received his B.S. degree from University of Science and Technology of China in 1982, M.S. and Ph.D. degrees from University of Southern California, USA, during 1984-1989. He worked in the areas of adaptive control, with particular interests in adaptive control of systems with multiple inputs and multiple outputs and with nonsmooth nonlinearities and actuator failures, instability and robustness of adaptive control systems, and in passivity characterizations of control systems. He has authored or coauthored 7 books, and more than 400 technical papers and book chapters. His current research interests include adaptive control of systems with uncertain actuator failures and nonlinearities, with structural damage and sensor uncertainties and failures, adaptive approximation-based control, and resilient aircraft and spacecraft flight control. He is currently a Professor at University of Virginia, USA. Since 2009, he has been a Chang Jiang Scholar visiting professor and a 1000-plan professor at Nanjing University of Aeronautics and Astronautics, China. He is a fellow of IEEE.
报告内容：Most results of adaptive approximation-based control (using neural networks and fuzzy systems) are for nonlinear systems with certain canonical-form structures. Such canonical-form systems have explicit relative degrees, which largely helps the adaptive control design. Non-canonical form systems do not have such explicit relative degrees, nor do their approximations which are also in non-canonical forms. Adaptive control of such more practical non-canonical form systems with neural network and fuzzy system approximations is to be solved.
This talk describes a new study on adaptive control of non-canonical form neural network or fuzzy function approximation nonlinear systems. A desirable adaptive control design has to reparametrize such systems, using an inherent and implicit relative degree formulation, a concept yet to be fully studied for various approximation models. This talk gives a unified presentation of some our recent results on adaptive control of neural network and fuzzy system based non-canonical form uncertain nonlinear systems, including the system relative degrees, normal forms, control design, adaptation, and performance analysis.