A Hybrid Method for Online Safe Trajectory Generation in Quadrotor Flight

In this paper, we propose a safe and efficient online trajectory generation method for quadrotors. This algorithm combines the sample-based motion planning method with the time-constrained trajectory optimization method. Based on the linearized state-space model of the quadrotor, we use the ellipsoid heuristic to directly sample the map and consider reasonable energy loss and completion time under the motion constraints, and construct a smooth and feasible initial trajectory for obstacle avoidance through multiple iteration optimization strategies. By using B-spline trajectory optimization and convex hull property, we minimize the loss of smoothing term, collision term, and dynamic feasible term to ensure fast and safe trajectory generation online. The proposed method is verified by numerical experiments, and compared with the classical RRT* algorithm and Kinodynamic RRT* algorithm, it is proved that the proposed method can generate feasible quadrotor trajectories at a faster speed.