A Multi-Objective Approach for Unmanned Aerial Vehicle Mapping

Many commercial applications require aerial mapping with multiple UAVs. Mapping is a mission planning problem which requires meeting a set of constraints while optimizing key factors that may conflict with each other, such as fuel/battery consumption, make-span, and the associated risks. Solving this Multi-Objective Optimization (MOO) will therefore result in a set of trade-offs (Pareto optimal solutions) that will be supplied to a decision-maker. Given that the Pareto set can be of a very large size, we propose a Multi-criteria Decision Making (MCDM) system that relies on user's preferences to bring down this set to a manageable size. More precisely, the proposed system captures user's qualitative preferences and uses them through the Fuzzy Vikor to filter and rank Pareto optimal solutions. The designed system is able to work with both or either fixed-wing and multi-rotor UAVs. To evaluate the performance of our system, we conducted a set of experimental simulations considering several scenarios. The findings show that fixed-wing UAVs have higher energy consumption and mission time than multi-rotors due to Dubin's turns, assuming both types have the same charging/fueling endurance and the same velocity. Lastly, it is found that heterogeneity will not always lead to a better mission duration than homogeneous UAV fleets.