视感知下的空地协同路径规划算法研究

发布时间：2018-12-25 18:00

【摘要】：近年来,伴随着智能机器人领域的快速发展,种类繁多的智能机器人按照不同的分工融入了我们的日常生活和生产。智能机器人在各行各业中扮演着越来越重要的角色。同时也为人类带来了极大的便利与解放[1]。尤其是在电力系统行业,各类巡检以及救援机器人代替工作人员在高电压高辐射的环境中作业,在提高工作效率的同时保证了电力行业工作人员的人身安全[2]。作为智能机器人中至关重要的一环,移动机器人导航与定位的研究价值毋庸置疑[3]。其中,对于移动机器人的路径规划算法的研究确保了机器人在各类约束条件下自主判断是否存在由起点至终点的最优路径。本文针对视感知下的空地协同路径规划问题,分为了规划区域尺度可变的局部路径规划,三维空间中和躲避动态障碍物的路径规划问题,以及视感知下的空地协同路径规划三个子问题逐步展开研究,主要工作包含以下三个方面(9):(1)对于移动机器人来说,按照规划的范围来分,可以分为局部路径规划和全局路径规划。其中局部实时路径规划对机器人在未知环境下的导航至关重要。然而,目前绝大多数局部路径规划都是在一个固定大小的局部范围内进行的。在一些情况下,这种规划区域尺度固定的路径规划会造成规划时间和资源的浪费。因此,本文了提出一种规划区域尺度可变的局部路径规划算法,使机器人可根据将要行进区域内障碍物的数量来决定规划区域的大小。仿真实验结果表明,移动机器人能够以更有效的方式来规划路径。另外,所提算法在交通道路地图中的也有较好应用。(2)目前,伴随着移动机器人行业的飞速发展,仅针对二维平面内的路径规划算法研究,已经无法满足一些特定类型的机器人(空中无人机、水下机器人等)的规划需求。因此,本文提出了移动机器人在三维空间中的路径规划算法,新的维度提高了找到更短路径的可能。另外,之前的研究都是在静态环境中进行的,而实际环境中,移动机器人往往会遇到动态障碍物。因此,本文提出了一种移动机器人在包含动态障碍物环境中的实时局部路径规划算法。仿真实验结果表明,移动机器人在规划路径的同时,具有实时躲避动态障碍物的能力。(3)在未知环境中,地面机器人受制于自身感知周围环境的能力的限制。在一些情况下,地面机器人会陷入障碍物的包围而不自知,这种情况下,需要浪费更多的时间和路程才能挣脱困境到达目的地。因此,为了避免这类情况的发生,本文提出了一种视感知下的空地协同路径规划算法,使用空中无人机为地面机器人提供高空视野。在地面机器人启动之前,无人机通过所搭载的视觉传感器扫描地面机器人将要行驶的未知环境,并将障碍物信息告知地面机器人。仿真实验结果表明,无人机扫描范围越大,提供的环境信息越多,地面机器人的路径规划效果越好。
[Abstract]:In recent years, with the rapid development of intelligent robot field, a variety of intelligent robots are integrated into our daily life and production according to different division of labor. Intelligent robots play a more and more important role in various industries. At the same time, it also brought great convenience and liberation to mankind. Especially in the power system industry, all kinds of inspection and rescue robots work in the environment of high voltage and high radiation instead of workers, which improves the working efficiency and ensures the personal safety of the workers in the power industry [2]. As an important part of intelligent robot, the research value of mobile robot navigation and localization is beyond doubt. Among them, the research of path planning algorithm for mobile robot ensures that the robot can independently judge whether there is an optimal path from the beginning to the end under all kinds of constraints. In this paper, the cooperative path planning problem under visual perception is divided into local path planning with variable regional scale, path planning with three-dimensional space and avoiding dynamic obstacles. And three sub-problems of cooperative path planning based on visual perception are studied step by step. The main work includes the following three aspects (9): (1) for mobile robots, according to the range of planning, It can be divided into local path planning and global path planning. Local real-time path planning is very important for robot navigation in unknown environment. However, at present, most of the local path planning is carried out in a fixed local range. In some cases, the planning of a fixed regional scale path planning will lead to a waste of planning time and resources. Therefore, in this paper, a local path planning algorithm with variable regional scale is proposed, in which the robot can decide the size of the planning area according to the number of obstacles in the moving area. The simulation results show that the mobile robot can plan the path in a more efficient way. In addition, the proposed algorithm also has a good application in traffic road map. (2) at present, with the rapid development of mobile robot industry, only for the two-dimensional in-plane path planning algorithm research, The planning needs of certain types of robots (aerial drones, underwater vehicles, etc.) have not been met. Therefore, this paper proposes a path planning algorithm for mobile robots in three dimensional space. The new dimension improves the possibility of finding shorter paths. In addition, previous studies are carried out in static environment, but in real environment, mobile robots often encounter dynamic obstacles. Therefore, this paper presents a real-time local path planning algorithm for mobile robots in environments containing dynamic obstacles. The simulation results show that the mobile robot has the ability to avoid the dynamic obstacles in real time while planning the path. (3) in the unknown environment, the ground robot is limited by its ability to perceive the surrounding environment. In some cases, the ground robot will be surrounded by obstacles without knowing it. In this case, it will take more time and distance to get out of this dilemma to get to its destination. Therefore, in order to avoid this kind of situation, a cooperative path planning algorithm based on visual perception is proposed in this paper, in which the aerial unmanned aerial vehicle (UAV) is used to provide high altitude vision for the ground robot. Before the ground robot is started, the UAV scans the unknown environment of the ground robot through the vision sensor and informs the robot of obstacles. The simulation results show that the larger the scan range of UAV is, the more environmental information is provided, and the better the path planning effect of ground robot is.
【学位授予单位】：兰州理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【参考文献】