A Hybird Trajectory Planning Algorithm for UAVs in Cluttered Environments

This paper proposes a hybrid trajectory planning algorithm for UAV, which can handle cluttered environments effectively. The proposed algorithm combines the Sampling-based method called RRT∗ with Dubins trajectory generation method to produce a trajectory satisfy the dynamic constrains of the UAV. Simultaneously, we modify the collision checking method combined the characteristics of Dubins trajectory, which guarantees the trajectory planning by the algorithm have a safe distance between with the obstacle in the environment. Further, the algorithm inherits the asymptotic optimality of the RRT∗ method, as the number of nodes increases in the tree graph, the cost of trajectories gradually decreases. Totally, the trajectory planned by the hybrid trajectory planning algorithm can satisfy the constrains of dynamic, smoothly, sufficient spacing from obstacles, and asymptotic optimality. The algorithm is validated through simulation and shown that our algorithm is able to reliably produce a safety and flyable trajectory in all the situations.