Energy-Aware Aircraft Trajectory Generation Using Pseudospectral Methods with Differential Flatness
Andrew B. Mills, Daniel Kim, Eric W. Frew · 2017 IEEE Conference on Control Technology and Applications (CCTA) · 2017
A path optimization strategy for generating energy-aware soaring trajectories for small Unmanned Aerial Systems (sUAS) is presented. Soaring paths are generated using a point-mass dynamics model and deterministic local wind field. The dynamics model, periodic constraints, and an energy-aware cost function are combined into a nonlinear program (NLP) framework to generate energy optimal paths. A dynamics model is derived which is differentially flat. This property, combined with a Legendre pseudospectral approximation facilitates the formulation of a novel and efficient energy-aware path planning optimization algorithm. The algorithm's ability to generate energy optimal trajectories is demonstrated for loiter and travel patterns in a linear wind gradient.