Development of global optimal coverage control using multiple aerial robots
This addresses the need for reliable global optimization in practical applications like environmental monitoring, though it is incremental as it builds on existing game-theoretic methods.
The paper tackled the problem of coverage control for mobile sensor networks, which often converges to local optima, by proposing a new algorithm that ensures deterministic convergence to the global optimum with collision avoidance, validated through simulations and experiments with aerial robots.
Coverage control has been widely used for constructing mobile sensor network such as for environmental monitoring, and one of the most commonly used methods is the Lloyd algorithm based on Voronoi partitions. However, when this method is used, the result sometimes converges to a local optimum. To overcome this problem, game theoretic coverage control has been proposed and found to be capable of stochastically deriving the optimal deployment. From a practical point of view, however, it is necessary to make the result converge to the global optimum deterministically. In this paper, we propose a global optimal coverage control along with collision avoidance in continuous space that ensures multiple sensors can deterministically and smoothly move to the global optimal deployment. This approach consists of a cut-in algorithm based on neighborhood importance of measurement and a modified potential method for collision avoidance. The effectiveness of the proposed algorithm has been confirmed through numerous simulations and some experiments using multiple aerial robots.