New distributed algorithm lets robot swarms maintain multi-coverage without central control

A team led by Mariem Guitouni and Aaron T. Becker from the University of Houston has developed a distributed multi-coverage algorithm for robot swarms, accepted at the International Conference on Swarm Intelligence (ANTS 2026) and published by Springer Nature. The core challenge addressed is maintaining reliable coverage of critical assets—such as surveillance targets, environmental sensors, or infrastructure nodes—when redundant observation by multiple robots is required (multi-coverage). Prior work solved this optimally with centralized integer programming, but that approach breaks down under real-world constraints: limited communication range, onboard computation restrictions, and partial system failures that must not abort the mission.

The new algorithm operates entirely with local sensing and local communication, requiring no global coordination or central planner. Each robot only needs to know the positions and coverage requirements of nearby assets and the presence of neighboring robots. The swarm self-organizes to ensure each asset is observed by the required number of robots, dynamically reassigning roles when a robot fails or moves. This is particularly relevant for autonomous drone swarms in surveillance, environmental monitoring, and infrastructure inspection, where resilience against individual failures is critical. The authors emphasize that the distributed approach scales well and is robust to communication disruptions and robot attrition, marking a significant step toward practical deployment of swarm systems in unstructured environments.