Multiobjective evolutionary flight planning of autonomous unmanned aerial vehicles for exploration and surveillance

This article presents a multiobjective evolutionary approach for computing flight plans for a fleet of unmanned aerial vehicles to perform exploration and surveillance missions. The static off-line planning subproblem is addressed, which is useful to determine initial flight routes to maximize the explored area and the surveillance of points of interest in the zone. A specific flight planning solution is developed, to be applied in low-cost commercial Bebop 2. The experimental analysis is performed in realistic instances of the surveillance problem. Results indicate that the proposed multiobjective evolutionary algorithm is able to compute accurate flight plans, significantly outperforming a previous evolutionary method applying the linear aggregation approach.