Data-driven Adaptive Optimal Control and Its Applications
Abstract: In many industrial applications, constructing an accurate model from physical laws is a hard and time-consuming undertaking, which makes traditional model-based control approaches impractical. With the recent advance of information science and technologies, scientists and engineers are actively seeking efficient ways to develop data-driven intelligent control systems that are highly robust, adaptive, scalable to uncertain or unknown environments. Adaptive dynamic programming (ADP) is a practically sound data-driven, non-model-based approach for control design in complex systems. In this talk, I will introduce a novel framework for adaptive optimal control by ADP. This framework can be employed to address different control problems, including output regulation, cooperative control and output-feedback control of linear and nonlinear dynamical systems. I will also present its applications to intelligent transportation systems, especially connected vehicles and autonomous vehicles. The future research challenges and opportunities in this area will be discussed as well.
Posted in Faculty Research, News, Research