A Dynamic Path Generation Method for a UAV Swarm in the Urban Environment

In this paper, a dynamic path planning method for swarming of unmanned aerial vehicles (UAVs) in the urban environment is presented. For the missions of a team of UAVs in a complex environment, conflict-free paths for all vehicles should be calculated dynamically in real-time using the latest information on the changes in the surroundings such as pop-up threats. Therefore, we propose a hierarchical dynamic path planner that consists of a offline path planning and a real-time model predictive trajectory generator. In this framework, many existing and proven off-line algorithms can be deployed for globally optimal path planning. The pre-computed trajectory is sent to the model predictive layer, which generates locally feasible trajectory free from conflicts. In this manner, each vehicle is able to fly along its designated path to reach the destination while avoiding obstacles or vehicles in collision path with minimal deviation from the designated path. For validation, the proposed algorithm is applied to a deployment scenario of sixteen rotary-wing UAVs flying in a cluttered urban area and showed a satisfactory performance.