【作者】 罗超；

【导师】 熊红云；

【作者基本信息】 中南大学 ， 电子信息（专业学位）， 2023， 硕士

【摘要】 无人机作为机器人领域的研究热点之一,近年来吸引了越来越多的研究人员的关注。在GPS无法使用的复杂环境下,无人机需要依靠自身携带的摄像头等传感器获取局部环境信息,然后依靠机载电脑进行在线路径规划。但是,由于机载电脑的算力有限,自主无人机的路径规划问题仍然面临着诸多挑战。针对复杂环境中自主无人机路径规划问题的相关难点进行研究,具体研究内容如下:提出一种设置了可视化检查和拐角代价函数的改进A*算法,解决了传统A*算法存在效率不高、路径拐角多的问题。该算法引入了可视检查的思想,通过设置节点之间的可视检查和增加拐角代价函数的方法,减少了路径上的转角次数和转角幅度。此外,算法将拓展节点的方式修改为拓展固定半径内的邻居节点,增加了拓展节点的灵活性。对比实验表明,改进的A*算法在算法效率、路径转角次数和转角幅度等性能指标上表现出了优越性。提出一种结合改进A*和B样条的无人机路径规划算法,提高了无人机长距离路径规划的实时性和安全性。该算法将轨迹表示成一个优化B样条曲线控制点的优化问题,通过对B样条曲线的控制点施加动力学可行性、安全性和轨迹的连续性等软约束,获得一条局部最优的轨迹。仿真实验表明,该方法在保证无人机安全的同时,也能够缩短飞行时间和飞行距离。搭建了自主无人机的仿真平台和四旋翼无人机的硬件平台,并通过实验验证了算法的有效性。首先基于可视化仿真平台搭建了自主无人机的仿真平台,在其中进行了无人机自主导航的硬件在环仿真实验。其次搭建了四旋翼无人机的硬件平台,在室内场景和室外树林场景中完成了自主无人机的自主导航,验证了路径规划算法的可行性。图47幅,表4个,参考文献76篇更多还原

【Abstract】 As one of the hot research topics in the field of robotics,unmanned aerial vehicles(UAVs)have attracted increasing attention from researchers in recent years.In complex environments where GPS is unavailable,UAVs need to rely on sensors such as cameras carried onboard to obtain local environmental information and then perform online path planning on the onboard computer.However,the computing power of the onboard computer is limited,so the path planning problem of autonomous UAVs still faces many challenges.An improved A* algorithm with the visual check and the corner cost function is proposed to solve the problems of low efficiency and many corners in the traditional A* algorithm.An improved A* algorithm with the visual check and the corner cost function is proposed to solve the problems of low efficiency and many corners in the traditional A* algorithm.This algorithm introduces the idea of smooth optimization and reduces the number of turns and amplitude of turns on the path by setting the visual inspection between nodes and increasing the corner cost function.In addition,the method of expanding nodes is modified to expand neighbor nodes within a fixed radius,which increases the flexibility of expanding nodes.Comparative experiments show that the improved A* algorithm has advantages in performance indexes such as algorithm efficiency,number of path turns,and Angle amplitude.A UAV path planning algorithm combining improved A* algorithm and B-spline curve is proposed to improve the real-time and safety of UAV long-distance path planning.In this algorithm,the trajectory is expressed as an optimization problem of optimizing B-spline curve control points,and a local optimal trajectory is obtained by applying soft constraints such as dynamic feasibility,safety,and trajectory continuity to B-spline curve control points.The simulation results show that this method can not only ensure the safety of UAVs,but also shorten the flight time and flight distance.To verify the feasibility of the proposed path planning algorithms,we built a simulation platform for autonomous UAVs and a hardware platform for the quadrotor and validated the effectiveness of the algorithms through experiments.First,based on the visualization simulation platform,we built a simulation platform for autonomous UAVs and conducted a hardware-inthe-loop simulation experiment for autonomous UAV navigation.Second,we built a hardware platform for the quadrotor and completed autonomous UAV navigation in indoor and outdoor forest scenes,verifying the feasibility of the path planning algorithm.更多还原