Concurrent Learning Adaptive Model Predictive Control

A concurrent learning adaptive-optimal control architecture for aerospace systems with fast dynamics is presented. Exponential convergence properties of concurrent learning adaptive controllers are leveraged to guarantee a verifiable learning rate while guaranteeing stability in presence of significant modeling uncertainty. The architecture switches to online-learned model based Model Predictive Control after an online automatic switch gauges the confidence in parameter estimates. Feedback linearization is used to reduce a nonlinear system to an idealized linear system for which an optimal feasible solution can be found online. It is shown that the states of the adaptively feedback linearized system stay bounded around those of the idealized linear system, and sufficient conditions for asymptotic convergence of the states are presented. Theoretical results and numerical simulations on a wing-rock problem with fast dynamics establish the effectiveness of the architecture.