Bipartite guidance, navigation and control architecture for autonomous aerial inspections under safety constraints

In this work the autonomous flight of a drone for inspection of sensitive environments is considered. Continuous monitoring, the possibility of override and the minimisation of the on-board computational load are prioritized. The drone is programmed with a Lyapunov vector guidance and nonlinear control to fly a trajectory passed, leg after leg, by a remote ground station. GPS is the main navigation tool used. Computational duties are split between the ground station and the drone’s on board computer, with the latter dealing with the most time critical tasks. This bipartite autonomous system marries recent advancements in autonomous flight with the need for safe and reliable robotic systems to be used for tasks such as inspection or structural health monitoring in industrial environments. A test case and inspection data from a test over flat lead roof structure are presented.