A vision-based Flight Guidance and Navigation System for autonomous cross-country soaring UAVs

In this work, a vision-based Flight Guidance and Navigation System is developed for providing cross-country soaring functionality to a small UAV glider. The overall system is based on existing major contributions to autonomous soaring while still missing components have been designed. A cumulus-cloud mapping algorithm was extended to not only estimate the position of cumuli but also their drift velocity and their updraft strength both factors with a significant impact on cross-country soaring performances. A guidance law was developed to navigate the vehicle during final glide from the last visited cloud to the destination given flight performance, mission and wind constraints. A path planning algorithm is proposed to then guide the vehicle in minimum time to its destination. Simulations were conducted in a nonlinear 6-DOF simulator for a small UAV glider to demonstrate the benefits of the system.