Model of Surveillance in Complex Environment Using a Swarm of Unmanned Aerial Vehicles

This paper examines the model of autonomous surveillance using a swarm of unmanned aerial vehicles with the simultaneous detection principle. This model enables to specify a number of sensors needed to detect an object of interest located in the area of interest; objects are detected only if scanned by the specified number of sensors simultaneously. The model plans deployment of individual vehicles in the swarm during the surveillance operation in a such a way that the surveillance is performed in the maximum quality; the quality is measured as a percentage of the area of interest that is covered during the operation. Furthermore, the surveillance is assumed to be conducted in the complex area of operations (including urban environments, build-up areas, or mountain environments with very uneven terrain) where occlusions caused by obstacles or terrain may occur often. For solution, the metaheuristic algorithm based on the simulated annealing is proposed. This algorithm deploys the number of waypoints, from which the monitoring is performed, maximizing the surveillance quality and taking the simultaneous detection principle into consideration. The algorithm is verified by a set of experiments based on typical surveillance scenarios.