A swarm-based algorithm for optimal spatial coverage of an unknown region

This paper presents an algorithm for optimal spatial coverage of an unknown region by a swarm of agents. The algorithm is based on the Ant Colony Optimization heuristic which is mapped and adapted to solve the current optimization problem. Each agent will leave a virtual pheromone trail during its movement through the unknown region, either attractor or repellent, represented by a positive or negative value, that decays in time. A novel stimergy technique is used to coordinate the agents' behavior when deciding to follow or to move away from the pheromone trail, depending on the pheromone value. For exploring the unknown areas a combination of a greedy technique based on the concept of rejection vector and a probabilistic technique for selecting the agents' rotating angles are employed. The obtained results are promising showing that our solution manages to obtain a coverage with up to 40% higher than classic rejection algorithm and with up to 25% that the distributed rejection algorithm.