SWARM UAV COVERAGE IN NON-CONVEX REGIONS: A CONVEX DECOMPOSITION AND NMPC APPROACH

Unmanned Aerial Vehicles (UAVs), generally referred to as drones, are widely studied and used. After an initial use exclusively in the military sector, drones have expanded their use, becoming part of any application in the civil sector.In this scenario, autonomous exploration of swarms of UAVs and the coverage of large areas present a challenge, especially in the case of areas with particular and unconventional geometric shapes. Considering these issues, the present work aimed to develop a UAV swarm-based exploration system with a specific focus on the coverage and exploration of geometrically concave areas. More specifically, the project compares two distinct coverage strategies, both evaluated using decentralized non-linear predictive control (NMPC).\\ The first method is the 'raw' one, i.e., it involves the swarm exploring the concave area without making any changes. The second, which is exactly the approach proposed in the thesis project, is based on the decomposition of the region of interest into convex subpolygons using a dedicated algorithm. The simulations were conducted on five concave areas, each with different geometric complexities, and for swarm configurations of up to six drones. The results show that in most simulations, the proposed approach is more efficient than its concave counterpart, with a constantly higher quality of coverage of the area to be explored and a more uniform and smooth distribution of swarm trajectories. In conclusion, with this work a new approach to the problem of coverage planning has been investigated through a method that can be considered as a first study step towards a wider and more in-depth investigation work.