A RRT*-based kinodynamic trajectory planning algorithm for Multirotor Micro Air Vehicle

This paper presents an online local trajectory planning algorithm for multirotor micro air vehicle(MAV) which applies the motion primitive generator proposed by Meuller et al. in RRT* algorithm to extend new nodes. The obtained results satisfy the dynamic constraints of multirotor MAV and avoid obstacles. Given initial state and final state, the motion primitive generator proposed by Meuller et al. generates a segment of polynomial trajectory to minimize the jerk of the trajectory. If initial state and final state are not feasible, the calculated motion primitive will not satisfy with the multirotor multirotor micro air vehicle dynamics. To resolve this problem, we use the numerical fitting method to obtain a mapping relation which gives the feasible range of the final state of the corresponding initial state. In the operation, such mapping can be called directly. Meanwhile, the trajectory is modified to decrease the velocity and acceleration of the end state of the whole trajectory as much as possible, so as to reduce the risk of colliding with obstacles not considered in the planning. Finally, we verified the effectiveness of the algorithm in the simulation environment.