A Multi-Population Genetic Algorithm for UAV Path Re-Planning under Critical Situation

This paper studies the path planning for Unmanned Aerial Vehicles (UAVs) under critical situations, where the aircraft has to execute a hard landing. Such critical situations can be provoked by equipment failures or extreme environmental situations that demand the UAV to abort the mission running and to land the aircraft without risk for people, properties and itself. First, a mathematical formulation is introduced to describe this problem. A planner system is proposed based on a multi-population genetic algorithm and a greedy heuristic. Computational results are conducted over a large set of scenarios with different levels of difficulty. Also, some simulations are executed using FlightGear simulator to illustrate the UAV's behaviour when landing under different wind velocities. The results achieved indicate the greedy heuristic is able to define faster feasible landing paths, whose quality can be improved by the evolutionary approach always within a short computation time.