Wind Energy Based Path Planning for a Small Gliding Unmanned Aerial Vehicle

This paper presents a framework for energy-based path planning for a small xed-wing gliding aircraft. The dynamic equations of motion for a gliding aircraft are derived for a three degree of freedom (3DOF) longitudinal model and a six degree of freedom (6DOF) model. The mechanisms of atmospheric energy gain are explored and formulated as a path planning problem. An energy-based reward function is presented which attempts to maximise energy extracted from the atmosphere. This is tested using a Rapidly Exploring Random Tree-like (RRT-like) path planner which explores reachable states with a limited set of control inputs. The planner is tested with full knowledge of the wind eld for a thermal bubble and a horizontal shear layer. It is capable of generating energy-gain trajectories using both static and dynamic soaring. The method is also tested for a ight through a thermal using only local wind information with a limited planning horizon.