Predictive Cost Adaptive Control of Three-Dimensional CFD-Simulated Flight

This paper applies predictive cost adaptive control (PCAC) to CFD-simulated flight of a three-dimensional wing. PCAC is an adaptive control technique that uses no prior model, no offline learning, and no probing inputs. Based on model-predictive control, PCAC uses recursive least-squares with variable-rate forgetting for multiple-input, multiple-output system identification. The present paper applies PCAC to a discontinuous Galerkin finite-element simulation of a wing in free flight. The motion of the wing is dictated by the aerodynamic forces and moments computed from the unsteady CFD simulation. The rigid-body dynamics of the wing are implemented with up to six degrees of freedom, although each degree of freedom can be frozen to simplify the simulation for testing. For actuation, PCAC uses two independent elevons on the flying wing, actuated smoothly using an arbitrary Lagrangian-Eulerian deformation. Results demonstrate the ability of PCAC to follow the commanded position changes by actuating the elevons in flight with multiple degrees of freedom.