Experimental Design and Control of a Smart Morphing Wing System using a Q-learning Framework

A novel control and testing platform for a smart morphing wing system is introduced to obtain optimal aerodynamic properties. This paper delves into the manufacturing process for said system, from the computer-aided modeling to the assembly, and its corresponding difficulties. The issues associated with the primary rendition of the model are addressed, as well as proposed solutions to these issues. Additionally, a novel formulation is introduced, which utilizes the 3D printed airfoil model for Computational Fluid Dynamics (CFD) analysis and reinforcement learning. In particular, image processing techniques and algorithms are employed to obtain an outline for the various morphed configurations, which are converted into airfoil coordinates and analyzed using a CFD tool, before being imported into a Q-learning algorithm.