Control and Obstacle Avoidance for an UAV Carrying a Load in Forestal Environments

Unmanned aerial vehicles (UAV) provide a wide range of applications, amongst which is load transportation. Aiming at a real application, such as delivery of medical kits for victims in disasters, it is important to include an obstacle avoidance technique in the navigation task, to avoid any obstacle that might appear along the way. Potential fields are a technique that considers objects as identical magnetic poles, such that one of them cannot get close to the other. Also, it is not necessary to create a potential map, since the vehicle can avoid the obstacles as they appear along the UAV navigation, characterizing a local decision making, not a global path-planning. In this work a UAV-cargo system is considered, navigating towards a desired position where the load should be delivered. Euler-Lagrange equations models the UAV-cargo system and its stability is guaranteed by a non-linear controller that rejects load oscillations. In such a case, the load is a modeled disturbance. To deal with obstacle avoidance, a forest-like environment is considered, and the UAV should navigate through the tree trunks till reaching the target position. Finally numerical simulations are presented to demonstrate the system performance.