Research On Trajectory Planning Algorithm For Quadrotor UAVs

With the progress of modern science and technology, the technology of mini multi-rotor UAVs is becoming more and more mature. And the application fields is becoming widely, and with this demands of trajectory planning is constantly increasing. On the condition that the UAVs satisfied with the performance and some specific constrains, the trajectory planning of unmanned aerial vehicle(UAV) refers to planning an optimal or sub-optimal flight from the starting point to the target point. It is an important part for the UAV to realize autonomous cruise.The traditional trajectory planning algorithm such as Ant colony optimization and Random search algorithm and so on does not considering the kinematic constraints of the trajectory. While the UAVs need to have a flight trajectory between each point when flying. In order to make full use of the performance of UAVs with hovering and all direction can mobile, also aiming at the defects of the traditional trajectory planning algorithm, and taking the ability of the operation speed is relatively weak into account. This paper proposes an brand new approach for the Sigmoid function-based trajectory planning between the multiple paths points, and at the same time we have designed two strategies for the angle adjustment for the UAVs. This approach applies the smooth and derivable Sigmoid function as the position function, and therefore determines the state trajectories between those midway points, meanwhile it guarantees the smoothness and derivability of the velocity functions.At the same time, we also used a approached algorithm in our UAVs. The algorithm is called Dubins trajectory planning algorithm, which is widely used in robot trajectory planning What is more we also give a detailed description on the three-dimensional cruising tasks.Finally this paper also gives the simulation results obtained by means of applying the dynamic model and tracking control scheme of a quadrotor UAVs are presented, showing the effectiveness, feasibility and real-time feature of the two algorithm.