水下机器人管线跟踪方法研究

CORC > 沈阳自动化研究所 > 中国科学院沈阳自动化研究所 > 水下机器人研究室

题名	水下机器人管线跟踪方法研究
作者	葛利亚
学位类别	硕士
答辩日期	2015-05-26
授予单位	中国科学院沈阳自动化研究所
授予地点	中国科学院沈阳自动化研究所
导师	李硕
关键词	水下机器人水下管线跟踪声呐图像处理行为规划导航与控制
其他题名	Research on Pipeline Tracking Method of Underwater Vehicle
学位专业	控制工程
中文摘要	近年来，人类对海洋不断进行开发和勘探，具有自主能力的水下机器人作为海洋探测的重要工具逐步得到了广泛的应用。多年来，在海底铺设了大量的水下管线（包括海底管道和海底电缆等），水下管线的正常运行是海上油气田生产及各国之间通信的重要保障。水下机器人以定期巡航的方式作业，使其具备自动寻找和跟踪水下管线的能力，是本文的主要研究目的。前视声纳作为水下机器人上搭载的重要传感器，可以为水下机器人提供有效的环境信息。因此，本文主要研究了基于前视声呐的水下机器人管线跟踪问题。具体研究内容如下：1) 简要介绍了课题的研究背景和研究意义，分析了水下管线检测和跟踪技术的国内外研究现状，前视声纳图像处理技术的研究现状及管线跟踪控制方法的相关研究现状，指出了本文的研究重点。2) 建立了基于前视声纳的水下管线跟踪系统结构，定义了大地和载体坐标系，建立了水下机器人运动学和动力学模型，分析了前视声纳的工作原理及前视声纳图像特点。3) 根据声呐图像的特点，重点研究声纳图像处理技术。从声呐图像中提取管线特征大体分为三个步骤：声纳图像预处理、图像分割及管线特征提取。研究包括图像滤波、灰度拉伸、二值化、形态学变换、连通域分割及面积、形状和亮度特征的提取等的相关方法和过程，并对各种处理算法进行了对比研究；最后对坐标转换过程进行分析，即将声纳图像中的管线信息如何转换到载体坐标系及大地坐标系的过程。4) 对比了三种基本搜索方式，并针对不同的情况选用不同的搜索方式。提出了两种跟踪策略：先搜索后跟踪和边搜索边跟踪。5) 将水下管线跟踪控制使命分为决策层和执行层两层，决策层又分为使命规划层、任务规划层，执行层又分为行为执行层和数字驱动层。设计了使命规划器和五个任务规划器，分析了几种比较典型的行为：下潜、上浮、管线搜索、管线跟随器、接近管线和跟随管线等。在行为执行层主要分析了PD控制和模糊P+ID控制两种运动控制算法。最后，在水下机器人运动特性的基础上，通过MATLAB仿真、三维视景仿真和水池试验，验证本论文提出的声纳图像处理和管线跟踪算法。仿真和试验结果都证明了水下管线跟踪算法的有效性和可行性
索取号	TP242/G38/2015
英文摘要	With human’s continuous development and exploration of the oceans, as an important tool for marine exploration, underwater vehicles which own autonomy have been widely used. Over the years, a large number of underwater pipelines (including submarine pipelines and submarine cables, etc.) have been lay undersea. The normal operation of underwater pipelines is an important guarantee for offshore oil production, gas production and communication between countries. Underwater vehicle works in the regularly cruising way. To make it have the ability to find and track underwater pipeline automatically is the main purpose of this research. As an important sensor underwater vehicle equipped with, forward looking sonar can provide effective environmental information for underwater vehicle. Therefore, this paper mainly studies the pipeline tracking problem for underwater vehicle based on forward looking sonar. Specific studies are as follows:1) This paper introduces the research background and significance of the subject briefly. The research status of pipeline detection and tracking technology in and outside of the country, research status of forward looking sonar image processing technology, and the related research status of pipeline tracking control technology are analyzed. Besides, research focus of this paper is pointed out. 2) The architecture of underwater pipeline tracking system is described. In addition, the earth-fixed and body-fixed coordinate system are defined, and motion and dynamics model of underwater vehicle are established. Furthermore, the working principle of forward-looking sonar and characteristics of forward looking sonar image are analyzed. 3) According to the characteristics of sonar image, sonar image processing technology is focused on. Extracting features of pipeline from sonar image can be divided into three steps: sonar image preprocessing, image segmentation and feature extraction of pipeline. Studies include associated methods and processes of image filtering, gray stretch, binarization, morphological transformation, segmentation of connected domain and extraction of size, shape and brightness characteristics etc. And various processing algorithms are studied comparatively. Finally, the process of coordinate conversion is introduced, meaning that how to convert pipeline information of sonar image to earth-fixed and body-fixed coordinates. 4) Three basic search methods are compared, and different search methods for different situations are chosen. Two tracking strategies are proposed: tracking after searching algorithm and tracking while searching algorithm, and they are introduced separately. 5) Underwater pipeline tracking control mission is divided into two layers: decision-making layer and implementation layer. The decision-making layer is divided into mission planning layer and task planning layer, besides, execution layer is divided into behavior execution layer and digital driver layer. Mission planner and five task planners are designed, in addition, several typical behaviors are analyzed: drive, float, pipeline searching, pipeline follower, approaching pipeline and following pipeline, etc. In the behavior execution layer, two motion control algorithms of PD control and fuzzy P+ID control are analyzed. Finally, algorithm proposed in this paper is verified through MATLAB simulation, 3D visual simulation and pool test based on the characteristics of the underwater vehicle. The simulation and experimental results show that the proposed algorithm can meet the real-time performance of tracking system. And the effectiveness and feasibility of the algorithm is demonstrated for underwater pipeline tracking tasks.
语种	中文
产权排序	1
页码	84页
内容类型	学位论文
源URL	[http://ir.sia.ac.cn/handle/173321/16787]
专题	沈阳自动化研究所_水下机器人研究室
推荐引用方式 GB/T 7714	葛利亚. 水下机器人管线跟踪方法研究[D]. 中国科学院沈阳自动化研究所. 中国科学院沈阳自动化研究所. 2015.