A Resilient Multi-Area Self-Organizing Coverage Method for UAV Swarms Under Communication-Limited Conditions

Multi-area coverage by unmanned aerial vehicle (UAV) swarms is vulnerable to limited inter-UAV communication and UAV failures, which can cause unbalanced resource allocation and slow recovery of coverage capability. This paper proposes a resilient self-organizing coverage method for UAV swarms under communication-limited conditions. The method integrates memory-based stochastic target selection, distributed color-state transitions, and virtual-force-based motion control, enabling UAVs to autonomously discover coverage areas and adjust their coverage participation using only locally perceived relative-position and visual-state information, without explicit multi-hop inter-UAV message exchange. UAVs not currently required for coverage are organized as a patrol reserve that continuously monitors coverage-state changes and can be redirected when local coverage degrades. To accelerate post-failure recovery, a localized threshold-adjustment mechanism temporarily weakens repulsive interactions for patrol UAVs moving toward affected areas, enabling rapid replenishment without swarm-wide parameter adjustment. A task-oriented multi-area resilience metric evaluates cumulative requirement satisfaction and effective coverage. Simulations assess coverage across different perception ranges and area numbers, while repeated runs under single- and multi-area failures show higher mean task-time, task-performance, and comprehensive resilience than the comparison methods. These results demonstrate that color-state-mediated local perception, patrol redundancy, and localized reconfiguration can effectively enhance resilient multi-area coverage of UAV swarms.