移动机器人导航关键技术的研究与实现
发布时间：2018-11-24 11:16
【摘要】：移动机器人在帮助社会产生巨大效益的前提之一就是它能够自主操作,不受到人为参与。其核心技术之一就是自主导航。自主导航是移动机器人在预先已知或者未知地图环境的情况下,从一个地点自主导航到另外一个地点来完成人们给机器人预先安排的任务。机器人从一个地点行走到另外一个地点需要考虑的问题有三个,包括机器人所在地图环境中的位置、机器人所到达的目的地、机器人如何到达指定地点。以上三个问题实际上就是导航系统三大内容:地图创建,机器人定位,路径导航。本论文是以水利水电行业电力智能巡检机器人项目的研发为基础,其重点研究的是机器人移动导航技术的实现。其中包括对地图环境的建模,主要采用拓扑地图抽象表示机器人导航所在的环境信息;采用ORB-SLAM和超宽带结合定位方法对机器人定位;提出一种基于贪心策略处理包含中间节点集的最短路径算法;提出了一种简单但高效的局部路径避障算法;以及处理机器人行走过程中所遇到的偏离航向、速度、方向等问题。本文所完成的工作以及贡献详细说明如下:(一)利用PHP+MYSQL+IIS开发一套web上位机系统。主要用于工作人员查看机器人状态、机器人位置;编写代码实现了Dijkstra最短路径算法,即给定目的地后,系统能够生成一条全局最短导航路径;编写代码实现了基于贪心策略来处理中间节点集的最短路径算法;可以通过可视化界面控制机器人前后左右运动;利用了多线程和非阻塞通信机制实现机器人与上位机系统通信、传输全局导航路径数据、机器人所在位置数据、控制机器人前后左右运动数据。(二)利用C/C++语言设计了一套移动机器人导航主体软件系统。机器人主体采用了两台mini电脑,一台控制机器人运动,进行局部路径规划的避障算法,控制机器人运动速度和方向,以及运行超宽带定位算法。一台处理ORB-SLAM算法和Kinect摄像头数据,并将所得到的数据通过以太网的形式传递给另一台电脑。(三)本文结合了ORB-SLAM定位算法和超宽带定位算法各自的优点,设计出在地图环境下不同位置采用不同的算法定位方法。机器人在直线行走时采用ORB-SLAM定位算法,而机器人在转弯时则采用超宽带定位算法。这样极大地提高了机器人的定位精度,并且所受到的外界干扰也有明显改善;同时本文还针对机器人在全局路径导航过程中,所出现的中间节点集情况下,提出一种基于贪心策略的最短路径算法。并且通过反复实验,证实了本算法有很强的实用性;最后本文在局部路径导航中针对水电巡检机器人特定的地图环境提出了一种简单但高效快速的避障算法,能够让机器人迅速地避开障碍物。
[Abstract]:One of the prerequisites for mobile robots to produce great benefits in helping society is that they can operate independently without human participation. One of its core technologies is autonomous navigation. Autonomous navigation is a mobile robot that navigates from one location to another to complete the tasks that people prearranged for the robot under the condition that the mobile robot knows or does not know the map environment in advance. There are three problems that the robot has to consider from one place to another, including the location of the robot in the map environment, the destination reached by the robot, and how the robot can reach the designated place. The above three problems are actually three main contents of navigation system: map creation, robot positioning and path navigation. This paper is based on the research and development of power intelligent inspection robot project in water conservancy and hydropower industry, which focuses on the realization of robot mobile navigation technology. It includes modeling the map environment, mainly using topological map to express the environment information of robot navigation, using ORB-SLAM and ultra-wideband localization method to locate the robot. Based on greedy strategy, a shortest path algorithm including intermediate node set is proposed, a simple but efficient local path obstacle avoidance algorithm is proposed, and some problems such as deviation from course, velocity and direction are dealt with in the course of walking. The work and contribution of this paper are described in detail as follows: (1) A web host computer system is developed by using PHP MYSQL IIS. It is mainly used to check the status and position of the robot, write code to realize the Dijkstra shortest path algorithm, that is, after a given destination, the system can generate a global shortest navigation path. The shortest path algorithm based on greedy strategy is developed to deal with the intermediate node set, and the robot movement can be controlled by visual interface. The multithreading and non-blocking communication mechanisms are used to realize the communication between the robot and the host computer, to transmit the global navigation path data, the position data of the robot, and to control the moving data before and after the robot. (2) A mobile robot navigation software system is designed by C / C language. Two mini computers are used in the main body of the robot, one is used to control the motion of the robot, the obstacle avoidance algorithm is used for local path planning, the speed and direction of the robot are controlled, and the ultra-wideband positioning algorithm is run. One handles ORB-SLAM algorithm and Kinect camera data and transfers the resulting data to another computer in the form of Ethernet. (3) combining the advantages of ORB-SLAM localization algorithm and UWB location algorithm, this paper designs different location algorithms in the map environment. The robot uses ORB-SLAM algorithm when walking in a straight line, while the robot uses ultra-wideband localization algorithm when turning. In this way, the positioning accuracy of the robot is greatly improved, and the external interference is obviously improved. At the same time, this paper proposes a shortest path algorithm based on greedy strategy for the intermediate node set in the global path navigation. And through repeated experiments, it is proved that this algorithm has strong practicability. Finally, this paper presents a simple but efficient obstacle avoidance algorithm for the specific map environment of hydropower patrol robot in local path navigation, which enables the robot to avoid obstacles quickly.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242
本文编号：2353484
[Abstract]:One of the prerequisites for mobile robots to produce great benefits in helping society is that they can operate independently without human participation. One of its core technologies is autonomous navigation. Autonomous navigation is a mobile robot that navigates from one location to another to complete the tasks that people prearranged for the robot under the condition that the mobile robot knows or does not know the map environment in advance. There are three problems that the robot has to consider from one place to another, including the location of the robot in the map environment, the destination reached by the robot, and how the robot can reach the designated place. The above three problems are actually three main contents of navigation system: map creation, robot positioning and path navigation. This paper is based on the research and development of power intelligent inspection robot project in water conservancy and hydropower industry, which focuses on the realization of robot mobile navigation technology. It includes modeling the map environment, mainly using topological map to express the environment information of robot navigation, using ORB-SLAM and ultra-wideband localization method to locate the robot. Based on greedy strategy, a shortest path algorithm including intermediate node set is proposed, a simple but efficient local path obstacle avoidance algorithm is proposed, and some problems such as deviation from course, velocity and direction are dealt with in the course of walking. The work and contribution of this paper are described in detail as follows: (1) A web host computer system is developed by using PHP MYSQL IIS. It is mainly used to check the status and position of the robot, write code to realize the Dijkstra shortest path algorithm, that is, after a given destination, the system can generate a global shortest navigation path. The shortest path algorithm based on greedy strategy is developed to deal with the intermediate node set, and the robot movement can be controlled by visual interface. The multithreading and non-blocking communication mechanisms are used to realize the communication between the robot and the host computer, to transmit the global navigation path data, the position data of the robot, and to control the moving data before and after the robot. (2) A mobile robot navigation software system is designed by C / C language. Two mini computers are used in the main body of the robot, one is used to control the motion of the robot, the obstacle avoidance algorithm is used for local path planning, the speed and direction of the robot are controlled, and the ultra-wideband positioning algorithm is run. One handles ORB-SLAM algorithm and Kinect camera data and transfers the resulting data to another computer in the form of Ethernet. (3) combining the advantages of ORB-SLAM localization algorithm and UWB location algorithm, this paper designs different location algorithms in the map environment. The robot uses ORB-SLAM algorithm when walking in a straight line, while the robot uses ultra-wideband localization algorithm when turning. In this way, the positioning accuracy of the robot is greatly improved, and the external interference is obviously improved. At the same time, this paper proposes a shortest path algorithm based on greedy strategy for the intermediate node set in the global path navigation. And through repeated experiments, it is proved that this algorithm has strong practicability. Finally, this paper presents a simple but efficient obstacle avoidance algorithm for the specific map environment of hydropower patrol robot in local path navigation, which enables the robot to avoid obstacles quickly.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242
【参考文献】
相关期刊论文 前8条
1 王东署;段谊海;王佳;;未知环境中移动机器人的环境探索与地图构建[J];郑州大学学报(理学版);2014年03期
2 赵世伟;马少池;吴书明;林洋;;红外线与超声波组合定位技术研究[J];价值工程;2014年14期
3 ;中国工业机器人领跑智能制造装备产业升级[J];可编程控制器与工厂自动化;2014年03期
4 肖竹;王勇超;田斌;于全;易克初;;超宽带定位研究与应用:回顾和展望[J];电子学报;2011年01期
5 朱大奇;颜明重;;移动机器人路径规划技术综述[J];控制与决策;2010年07期
6 李季;孙秀霞;;基于改进A-Star算法的无人机航迹规划算法研究[J];兵工学报;2008年07期
7 张殿东;无线射频识别(RFID)技术[J];电信技术;2005年02期
8 孙维君;欧式空间中向量的叉积及其应用[J];山东理工大学学报(自然科学版);2004年01期
相关硕士学位论文 前2条
1 杨飞;基于三维激光雷达的运动目标实时检测与跟踪[D];浙江大学;2012年
2 于克;室内自主移动机器人的定位研究[D];燕山大学;2009年
，本文编号：2353484
本文链接：https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/2353484.html
