Cobertura de regiões com múltiplos veículos aéreos de asa rotativa

Abstract

Multi-UAV Coverage Path Planning is a fundamental problem in inspection, mapping and environmental monitoring. A recent energy-aware method formulates the problem as a Multiple Set Traveling Salesman Problem and solves it with GRASP and Tabu Search. This work replaces that solver with an energy-aware Adaptive Large Neighborhood Search (ALNS), named Energy-ALNS, integrated into the same planning pipeline and preserving the Boustrophedon Cellular Decomposition, energy model and sweep-angle strategy. Across 16 scenarios and five random seeds, the Energy-ALNS wins in 11 cases, ties in 4 and loses in 1 against the original solver, with a mean gain of 1.90\% in energy makespan (Wh). An operator analysis shows that energy-aware operators, not the reinforcement-learning controller, drive competitiveness. A Gazebo Sim~8 / ROS~2 Humble / PX4 simulation confirmed route executability under real multirotor dynamics.