Data-Driven Learning of System Dynamics & Representation

In this area, our research focus on using data driven model to model and simulate dynamic systems, fluid dynamics, multi-physics, using on the shelf methods and developing new frameworks,  including Koopman operators, Physics informed machine learning, functional analysis, and  Riemannian geometry, to represent system dynamics in temporal-spatial domain.

Physics-Informed Machine Learning

Under this umbralla, we research on learning, solving differential equations to obtain run-time physical information for system identification and control. Beyond the system identification problem that has been previous investigated using PINN, we are also exploring methods to learn unknown physics for new physics discovery.

Smart Structures with Sensing and Actuating

In this area, we focus on creating, integrating smart sensors and structures that can generate data and information, and provide necessary cyber-physcial interface for data driven engineering.