LAUSR

Abstract This paper investigates multi-role coverage control for multi-color mass games, whose objective is to represent a given multi-color image by n-color mobile agents, i.e., each agent plays the role of each color pixel. First, the original image is quantized to an n-color digital image, where each color is regarded as a target for the same color agents. Then, the global objective function is formulated in terms of the coverage area, that is, to let each color agent maximize each color function by occupying the target pixels. In this novel setting, conventional Voronoi-based gradient approaches tend to become stuck in local optima. To address this issue, we present a Voronoi-based cut-in approach so that the agents are led away from local to global optimal positions without affecting collision avoidance performance. Finally, the effectiveness of this approach is demonstrated through a numerical simulation.