中国科学院沈阳自动化研究所机构知识库
Advanced  
SIA OpenIR  > 水下机器人研究室  > 学位论文
题名: 自主水下机器人动态目标跟踪关键技术研究
其他题名: The key technology Research on Dynamic Target Tracking of UUV
作者: 徐进宝
导师: 封锡盛
分类号: TP242
关键词: 自主水下机器人 ; 鲁棒滤波 ; 数据关联 ; RB粒子滤波
索取号: TP242/X74/2011
学位专业: 模式识别与智能系统
学位类别: 博士
答辩日期: 2011-11-25
授予单位: 中国科学院沈阳自动化研究所
学位授予地点: 中国科学院沈阳自动化研究所
作者部门: 水下机器人技术研究室
中文摘要: 自主水下机器人(UUV, Unmanned Underwater Vehicle)作为一个费用低、可以替代人或可以遥控的系统已经得到了广泛的应用,如在海洋开发、军事活动等领域发挥的重要作用。但随着人们对UUV 能力的要求不断提高,使得UUV 的使命变得越来越复杂,然而其自主性不高却限制了它在很多方面的潜在应用。由于UUV在军事和民用方面的重要性,很多国家已经制定了水下机器人发展规划,其中重点研究使其具有三大自主性,自主能量、自主导航、自主决策。鉴于对UUV自主性发展的迫切要求,所以进行UUV的自主性研究是必要的。进行UUV的自主性研究必须要UUV自身能够感知周围的环境,然后通过对环境的感知和自身的态势进行决策。UUV能够感知环境,必须需要UUV能够具有对传感器信息的处理,融合能力。 本文的研究工作依托中国科学院重大基础科研项目等课题,在总结国内外研究成果的基础上,借鉴其它领域的研究成果,重点研究了单(多)UUV对纯方位信息的处理(融合)能力。具体的研究内容为: (1)水下机器人纯方位跟踪的基本理论 本部分对应用于水下机器人纯方位跟踪基本理论的各部分做了介绍。首先介绍了纯方位跟踪中的笛卡儿坐标系和修正极坐标系,并给出了两种坐标系下的数学模型;然后介绍了用于跟踪的各种经典滤波算法,重点介绍了基于UT变换的无迹卡尔曼滤波理论和粒子滤波理论;最后还给出了后面用于滤波的初始化方法。这些理论为单目标水下机器人纯方位跟踪奠定理论基础。 (2)鲁棒滤波在纯方位跟踪中的应用 由于实际跟踪系统的模型具有很多不确定性,如初始化的不确定性、观测噪声的不确定性,状态噪声不确定性等等。这些不确定性使得在使用基于卡尔曼滤波框架的最小均方差估计滤波算法实施跟踪过程中,鲁棒性会很差。本部分根据线性滤波算法,利用SUT变换去近似估计随机变量函数的均值与方差,提出了滤波算法,并分别推导出了基于笛卡尔坐标系的纯方位滤波算法和基于修正极坐标系纯方位滤波算法。并在修正极坐标系和笛卡儿坐标系下进行了仿真验证,得到了修正坐标系下的滤波跟踪最优的结果。 (3)杂波环境下,水下机器人纯方位跟踪方法 实际的跟踪环境往往是多目标环境或多杂波环境,也就是说水下机器人所携带的传感器得到的信息可能是含有杂波的单目标信息或者是含有杂波的多目标信息。所以在水下机器人进行纯方位跟踪时要首先要利用数据关联技术将传感器的观测与目标相关联,然后再进行滤波的更新。本部分首先总结了用于多杂波多目标环境的数据关联方法,根据RB粒子滤波算法,并结合前人的研究成果,根据水下纯方位跟踪的特殊环境,提出了基于RB粒子滤波数据关联水下纯方位跟踪算法,并将该种数据关联算法与上一部分的鲁棒滤波算法和无迹卡尔曼滤波相结合,实现水下单观测载体和双观测载体在多杂波环境下的纯方位目标跟踪算法。通过仿真验证,能对目标实施有效的跟踪。 (4)水下机器人纯方位跟踪实验 根据前面的理论知识,通过对现场实验数据进行在两种坐标系下的扩展卡尔曼滤波处理,得到修正极坐标系下的滤波方法比笛卡尔坐标系优的特点。最后通过对两组相似实验数据通过给定不同的初始距离,进行无迹卡尔曼滤波和鲁棒滤波进行算法比较,通过这两组数据的分析,鲁棒滤波的性能要好于无迹卡尔曼滤波。
英文摘要: Unmanned Underwater Vehicle (UUV) is widely used as a low-cost human replacement or a remote-controlled system in ocean development, military activities, and etc.. With the emerging demand of UUV, their functions are getting increasingly complex. However, the relatively low autonomy limited the applications of the existing UUV systems. Many plans of development of UUVs have been recently seen globally, of which the main interests are focused on automatic energy, automatic navigation and unmanned decision making. Hence, the research of autonomy for UUVs is becoming progressively significant. It is essential that UUV are able to sense the status of surroundings, whereby decisions are made. Here, the ability of information processing and fusion is also key to the research. This paper relies on the fundamental research project of Chinese Academy of Science. It employs global research outcome from diverse fields. Moreover, this research concentrates on processing (or fusion) of bearings-only information for single (or multiple UUVs) UUV. The rest of the thesis is organized as the follows. (1)Principles of bearings-only tracking in UUV. This section introduces the principles of Bearings-Only Tracking (BOT) in UUV. Firstly, Cartesian Coordinates (CC) and Modified Polar Coordinates (MPC) in BOT is introduced. Subsequently, mathematical models of BOT in both coordinates are described. Secondly, a range of conventional filtering algorithm, which can be used in BOT, are introduced. Here, mainly Unscented Kalman Filter (UKF) based on Unscented Transform (UT) and Particle Filter (PF) are looked into. Lastly, the following methods about filter initialization that can be used in filtering are investigated. To conclude, this section gives the principles of the proposed single-target BOT for UUV in the following sections. (2) The application of robust filter in BOT. Due to the uncertainties in actual tracking system, for instance the uncertainties of initialization, observation noise, and state noise, poor robustness are seen in tracking using Kalman filter based on the minimum Mean Square Error estimation. In this section, a novel filtering algorithm, ,  is proposed, based on linear robust filtering algorithm, with the mean and variance for the function of random variable estimated by Scale Unscented Transform (SUT). The bearings-only filtering algorithm in CC and MPC of the proposed algorithm are derived, and, respectively. Simulations of both of the variations are carried out. The results show that the robust filter in MPC returns a superior performance. (3) The UUV bearings-only tracking in clutter. Generally, in real world, tracking conditions are multi-target or cluttered. In other words, the data that the on-board sensors acquired are usually cluttered single-target or multi-target. Hence, measurements of sensors need to be associated with targets using data association techniques before filter update. This section firstly reviews the data association techniques that can be used in multi-target or cluttering conditions. Then, an RB particle filter based data association method is proposed specifically targeting under water conditions. Later, this method is integrated with the robust filtering algorithm and UKF presented in Section 2. Finally, simulations are carried out, and the results show that the proposed data association method can effective deal with under water conditions as intended.  (4) Experiment of bearing-only tracking for UUV. Founded on the theoretical results, field experiment data and analyzed using extend Kalman filter in CC and MPC coordinate systems. It is also concluded that filtering algorithm provides an advanced performance in MPC compared to CC. Finally, two sets of similar experimental data, which has different initialization distances, are analyzed with both robust filtering and UKF. The result shows that robust filtering is superior compared the latter. 
语种: 中文
产权排序: 1
内容类型: 学位论文
URI标识: http://ir.sia.cn/handle/173321/9290
Appears in Collections:水下机器人研究室_学位论文

Files in This Item:
File Name/ File Size Content Type Version Access License
自主水下机器人动态目标跟踪关键技术研究.pdf(1832KB)----限制开放 联系获取全文

Recommended Citation:
徐进宝.自主水下机器人动态目标跟踪关键技术研究.[博士 学位论文].中国科学院沈阳自动化研究所.2011
Service
Recommend this item
Sava as my favorate item
Show this item's statistics
Export Endnote File
Google Scholar
Similar articles in Google Scholar
[徐进宝]'s Articles
CSDL cross search
Similar articles in CSDL Cross Search
[徐进宝]‘s Articles
Related Copyright Policies
Null
Social Bookmarking
Add to CiteULike Add to Connotea Add to Del.icio.us Add to Digg Add to Reddit
所有评论 (0)
暂无评论
 
评注功能仅针对注册用户开放,请您登录
您对该条目有什么异议,请填写以下表单,管理员会尽快联系您。
内 容:
Email:  *
单位:
验证码:   刷新
您在IR的使用过程中有什么好的想法或者建议可以反馈给我们。
标 题:
 *
内 容:
Email:  *
验证码:   刷新

Items in IR are protected by copyright, with all rights reserved, unless otherwise indicated.

 

 

Valid XHTML 1.0!
Copyright © 2007-2016  中国科学院沈阳自动化研究所 - Feedback
Powered by CSpace