• Publisher-DOI

Planning flight paths for unmanned aerial vehicles in urban areas requires consideration of safety, legal, and economic aspects as well as attention to social factors for gaining public acceptance. To solve this many-objective path planning problem in the three-dimensional space, we propose a hybrid framework combining an exact Dijkstra search and a metaheuristic evolutionary optimization. Given a start and an endpoint, we optimize a path regarding the risk in case of a system failure, the radio signal disturbance between the aerial vehicle and a ground station, the energy consumption, and the noise immission on city residents. The optimization includes constraints for static obstacle collision avoidance and compliance with the minimum flight altitude. The result is a set of smooth and three-dimensional paths that realize different trade-offs between the defined objectives. As an example, we consider an urban transportation application for aerial vehicles in San Francisco. For all tests, we use real-world data from OpenStreetMap. In a statistical evaluation, we test the efficiency of our framework against different state-of-the-art optimizers. Moreover, we extend the framework with two features that allow the user to integrate arbitrary objectives and unknown scenarios into the path planning framework.