Combining Lévy Walks and Flocking for Cooperative Surveillance using Aerial Swarms

Continuous area coverage missions are a fundamental part of many swarm robotics applications. One of such missions is cooperative surveillance, where the main aim is to deploy a swarm for covering predefined areas of interest simultaneously by k robots, leading to better overall sensing accuracy. However, without prior knowledge of the location of these areas, robots need to continuously explore the domain, so that up-to-date data is gathered while maintaining the benefits of simultaneous observations. In this paper, we propose a model for a swarm of unmanned aerial vehicles to successfully achieve cooperative surveillance. Our model combines the concept of Lévy Walk for exploration and Reynolds’ flocking rules for coordination. Simulation results clearly show that our model outperforms a simple collision avoidance mechanism, commonly found in Lévy-based multi-robot systems. Further preliminary experiments with real robots corroborate the idea.