A multi-UAV minimum time search planner based on ACO R

This paper presents a new planner based on Ant Colony Optimization for Real-coded domains (ACO R ) for optimizing the trajectories of multiple Unmanned Aircraft Vehicles (UAVs) in Minimum Time Search (MTS) missions, where the UAVs have to shorten the detection time of a given target while avoiding collisions and Non-Flying Zones (NFZ). Therefore, the planner has to identify the UAV trajectories that minimize the Expected Time of Target Detection (ETTD) and nullify the total of NFZ overflights and of UAV collisions. To achieve it, the planner is backed by an ACO R that 1) ensures the feasibility of the trajectories by encoding them as a sequence of input UAV control commands and by decoding them through complex UAVs kinematic/dynamic models, 2) handles the uncertainty of the sensor and of the target location in the computation of the ETTD using Bayesian theory, and 3) improves the planning process with a heuristic that has been especially designed to exploit the probability and spatial properties of the problem. All these properties let our ACO R based planner handle successfully minimum time target detection missions in real world scenarios, as the results analyzed in this paper, obtained over different setups, show.