A Data-Driven Linear Model for Flexible Vehicles Based on Koopman Operator

This paper is concerned with the linear modeling method for a flexible air-breathing hypersonic vehicle. It is to develop a data-driven linear modeling approach to capture nonlinear properties of the FAHV based on the Koopman operator. The nonlinear dynamics of the FAHV is lifted into a lifted function space, and its evolution is then approximated in a finite-dimensional linear form under the frame of Extended Dynamic Mode Decomposition. In this linear model, the FAHV’s immeasurable flexibility is implicit in the lifted states mapped from measurable rigid body states. Numerical simulations verify the proposed linear model has high accuracy in predicting FAHVs’ dynamics. Particularly, in description of the flexibility effect, it exhibits a performance superior to existing models such as the nonlinear rigid model and the local linearized model. It is promise to apply various classic methods in control of FAHVs with this linear state space model in the basis.