Find in Library

In this study, an adaptive online data-driven tracking controller for a highly flexible aircraft (HFA) was developed. This control design innovatively combines integral reinforcement learning (IRL) and the optimal control theory to ensure asymptotic tracking performance, even when system dynamics information is difficult to obtain. As online data collection may cause partially observable control problems, this study also incorporates a class of state parameterization method in the proposed controller in order to deal with partial observability. Finally, the proposed controller is demonstrated via a simulation of the longitudinal dynamics of a HFA model.