Morphing Unmanned Air Vehicle Intelligent Shape and Flight Control
This paper develops and demonstrates a complete methodology for the control of a morphing unmanned air vehicle. The shape learning is done with a modified episodic Reinforcement Learning algorithm, which employs an adaptive grid to improve the search performance and accuracy of learning the optimal shape change policy. The shape control, which uses Reinforcement Learning, and the trajectory tracking flight control, which uses Structured AdaptiveModel Inversion Control, are combined in a technique called Adaptive-Reinforcement Learning Control. Optimality is addressed by cost functions representing optimal shapes corresponding to specified operating conditions. The shape change dynamics are represented by Shape Memory Alloy material hysteresis input-output mappings which were determined experimentally. The nonlinear, six degree-of-freedom unmanned air vehicle dynamical model simulation is integrated with a constant strength source doublet panel method CFD code. This combined 3-D vehicle model code is capable of simulating the dynamic response and the forces and moments which are generated due to commanded multiple large scale shape changes, consisting of thickness, sweep angle, dihedral angle, and chord length. The methodology is demonstrated by a numerical examples of the morphing air vehicle simultaneously tracking a specified trajectory and autonomously morphing over a set of shapes corresponding to flight conditions along the trajectory. Results presented in the paper demonstrate that this methodology is effective for accurately tracking a flight trajectory in the presence of parametric uncertainties and initial error conditions.