论文部分内容阅读
针对微小型四旋翼无人机的路径规划问题,引入运动学习框架,提出了一种基于动态运动基元的路径规划方法。该方法通过对给定运动样本的学习提取出运动基元,并将学习结果推广到新的飞行目标,从而泛化出相应的运动轨迹。有障碍物的情况下,在已有学习基础上通过设计耦合因子规划出避障路径,然后将规划的轨迹点集提供给微小型四旋翼无人机完成路径跟踪飞行任务。该路径规划方法的可行性通过微小型四旋翼无人机不同目标点的飞行任务仿真得到了验证。仿真实验还验证了该方法在三维空间进行有效避障的性能。
Aiming at the path planning problem of a miniature four-rotor UAV, a motion learning framework is introduced and a path planning method based on dynamic motion primitives is proposed. In this method, the motion primitives are extracted from the learning of a given motion sample, and the learning result is extended to a new flight target to generalize the corresponding motion trajectory. In the case of obstacles, the obstacle avoidance path is planned by designing the coupling factor based on existing learning, and then the planned trajectory point set is provided to the micro-quadrotor UAV to complete the path tracking mission. The feasibility of this path planning method is validated by the simulation of mission missions on different target points of a miniature quadrotor UAV. Simulation experiments also verify the performance of this method in three-dimensional space for effective obstacle avoidance.