Combining stigmergic and flocking behaviors to coordinate swarms of drones performing target search

Due to growing endurance, safety and non-invasivity, small drones can be increasingly experimented in unstructured environments. Their moderate computing power can be assimilated into swarm coordination algorithms, performing tasks in a scalable manner. For this purpose, it is challenging to investigate the use of biologically-inspired mechanisms. In this paper the focus is on the coordination aspects between small drones required to perform target search. We show how this objective can be better achieved by combining stigmergic and flocking behaviors. Stigmergy occurs when a drone senses a potential target, by releasing digital pheromone on its location. Multiple pheromone deposits are aggregated, increasing in intensity, but also diffused, to be propagated to neighborhood, and lastly evaporated, decreasing intensity in time. As a consequence, pheromone intensity creates a spatiotemporal attractive potential field coordinating a swarm of drones to visit a potential target. Flocking occurs when drones are spatially organized into groups, whose members have approximately the same heading, and attempt to remain in range between them, for each group. It is an emergent effect of individual rules based on alignment, separation and cohesion. In this paper, we present a novel and fully decentralized model for target search, and experiment it empirically using a multi-agent simulation platform. The different combination strategies are reviewed, describing their performance on a number of synthetic and real-world scenarios.