Downwash-Aware Trajectory Planning for Quadrotor Swarms

We describe a trajectory planning pipeline for large quadrotor teams in obstacle-rich environments. We construct a sparse roadmap in the environment and use a boundedsuboptimal conflict-based graph search to generate a discrete plan. We then refine this plan into into smooth trajectories using a spatial partition and Bézier curve basis. We model downwash directly, allowing safe flight in dense formations. We show simulation results with up to 200 robots and a real-robot experiment with 32 quadrotor. To our knowledge, our approach is the first solution which can compute safe and smooth trajectories for hundreds of quadrotor in dense environments with obstacles in a few minutes.