Adaptive Weight-Shift Flight Control of Flexible Wing Aircraft

Abstract

Weight-shift controlled flexible wing ultralight aircraft are renowned for their maneuverability, low-cost operation, and high flight efficiency; yet their potential for broader applications remains untapped. No current flight control systems exist to replace the intensive manual effort and specialized training required for operating this class of aircraft. This developmental gap is attributed to modelling and control challenges presented by the complex and highly nonlinear dynamic interactions from attitude adjustment by weight-shifting, the aeroelastic properties of the flexible wing, and their resulting effects on flight behaviour. These dynamics constrain the applicability of model-based control approaches due to inherent parameter uncertainty, thereby necessitating an adaptive approach to flight control system design. This work presents novel robotic piloting systems that address the automatic weight-shift flight control problem. Fly-by-wire and computer-controlled operation is realized through the implementation of cable-driven parallel manipulators which replicate human-powered weight-shift action. Optimal coordination of the parallel actuators is achieved through a model-free approach, utilizing online adaptive critics reinforcement learning, enabling automatic piloting systems to track pitch-roll trajectories despite uncertain or unknown system dynamics and external disturbances. Through industry partner collaboration, the piloting systems are validated with real-world hardware surrogates, a flight dynamics simulator environment, and optionally-manned experimental flight test. The significant contributions of this thesis include the development of implementation techniques for online actor-critic based adaptive learning suitable for performance-constrained hardware, the design of model-free force control for cable-driven parallel robots, the creation of an open-source weight-shift controlled dynamics simulation model validated with real flight data, and the advancement of fly-by-wire weight-shift flight control technology. The results push the boundaries of weight-shift aircraft flight control technology by providing a means to enhance human pilot safety with automatic control, and pave the way for the development of unmanned and autonomous weight-shift aircraft.