杂波环境下AUV纯方位目标跟踪方法研究
文献类型：学位论文
| 作者 | 梅登峰
|
| 学位类别 | 硕士 |
| 答辩日期 | 2014-05-28 |
| 授予单位 | 中国科学院沈阳自动化研究所 |
| 导师 | 刘开周 |
| 关键词 | 自主水下机器人 容积卡尔曼滤波 概率数据关联滤波器 联合概率数据关联滤波器 广义概率数据关联滤波器 无人水面机器人 在线路径规划 速度避障法 混合整数线性规划 |
| 其他题名 | Research of AUV bearings-only target tracking in the cluttered environment |
| 学位专业 | 控制工程 |
| 中文摘要 | 随着科技的进步，尤其是电子技术、计算机技术、传感器技术、导航控制技术的飞速发展，自主水下机器人（autonomous underwater vehicle, AUV）的成本逐渐降低，同时，其智能水平和作业能力不断提高。AUV作为一种新型的水下航行器，人们不断地赋予它新的使命，其中一个非常重要的使命就是让AUV对水下运动目标进行纯方位跟踪。AUV单观测站纯方位目标跟踪可以应用于海洋反恐，海洋环境监控等领域。 本文以国家自然科学基金（No.61273334）弱观测复杂海洋环境下AUV动态目标跟踪算法研究、辽宁省自然科学基金（No.2011010025-401）弱观测条件下AUV动态目标跟踪及机动策略研究、中国科学院知识创新工程重要方向项目（No.YYYJ-0917）等科研课题为依托，对杂波环境下的AUV单观测站纯方位目标跟踪进行了深入系统的研究。本论文的主要工作内容如下： 1) 论文简要介绍了AUV单观测站纯方位目标跟踪的研究背景和意义，概述了AUV国内外研制现状和纯方位目标跟踪以及数据关联的研究现状，最后，描述了本文的研究内容与安排。 2) 研究了AUV纯方位目标跟踪的原理和常速度（CV）模型，将CV模型用于AUV纯方位目标跟踪问题的建模，该数学模型为后面的研究奠定了基础。 3) 无杂波理想环境下的AUV单观测站单目标纯方位跟踪问题本质上是一个不完全可观测的非线性估计问题。本文针对这种非线性估计问题，深入研究了扩展卡尔曼滤波（EKF）、无迹卡尔曼滤波（UKF）和容积卡尔曼滤波（CKF）三种非线性滤波算法。先是在理论上对三种非线性滤波算法进行了研究，然后通过仿真实验对它们进行了验证。 4) 杂波环境下的AUV单观测站单目标纯方位跟踪问题比理想环境下的跟踪问题复杂很多，因为它不仅涉及到非线性滤波还涉及到多个测量数据与单个目标的关联。针对该问题，本文深入研究了概率数据关联滤波器（PDAF）和最大熵模糊概率数据关联滤波器（MEF-PDAF）算法。由于单观测站纯方位目标跟踪系统是一个强非线性系统，针对这一点，本论文对MEF-PDAF算法进行了改进，提出了基于无味变换的MEF-PDAF（UT-MEF-PDAF）算法。在进行深入的理论研究之后，通过数值仿真实验和外场数据实验对上述三种方法进行了验证和分析。 5) 杂波环境下的AUV单观测站多目标纯方位跟踪问题比单目标跟踪问题更加复杂，因为它涉及到多个测量数据与多个目标的关联问题。针对该问题，论文深入研究了联合概率数据关联滤波器（JPDAF）算法和广义概率数据关联滤波器（GPDAF）算法。在理论研究之后，通过两个仿真实例对这两种数据关联滤波器进行了实验分析研究，实验结果说明了GPDAF算法优于传统的JPDAF算法。 |
| 索取号 | TP242.3/M44/2014 |
| 英文摘要 | With the progress of science and technology, especially the rapid development of electronic technology, computer technology, sensor technology, navigation and control technology, the cost of AUV (autonomous underwater vehicle) is reduced, at the same time, its intelligence and working capability is enhanced. AUV, as a new underwater vehicle, more and more missions are entrusted to it by human, and one of the most important missions is to track the underwater moving target using the bearings-only tracking method. AUV single observer bearings-only target tracking can be used in sea anti-terrorist, marine environment surveillance and so on. This paper is supported by National Natural Science Foundation of China (No.61273334), Provincial Natural Science Foundation of Liaoning (No.2011010025-401) and Key Project of Innovation Knowledge of Chinese Academy of Sciences (No.YYYJ-0917). A deep and scientific research in AUV single observer bearings-only target tracking is done in this paper. The main contents of this paper are as follows: 1) The research background and significance of AUV single observer bearings-only target tracking is described. The development of AUV, the current research status of bearings-only target tracking and data association is present. Last, the main research contents and the arrangement of the paper are described. 2) The principle of AUV single observer bearings-only target tracking and the constant velocity (CV) model is studied, and then, the CV model is applied to modeling the problem of AUV single observer bearings-only target tracking. The new model paves the way for the later research. 3) AUV single observer single target bearings-only tracking under uncluttered ideal conditions is essentially an incomplete observable nonlinear estimation problem. For this problem, three kinds of nonlinear filters, which includes extended Kalman filter (EKF), unscented Kalman filter (UKF) and cubature Kalman filter (CKF), are deeply studied. Firstly, theory research on the three algorithms is done, and then the simulation is conducted to verify them. 4) The problem of AUV single observer single target bearings-only tracking in the cluttered environment is more difficult than tracking under ideal conditions, because it involves not only nonlinear filtering but also data association of multiple measurements with single target. For the above problems, probability data association filter (PDAF) algorithm and maximum entropy fuzzy probabilistic data association filter (MEF-PDAF) algorithm are deeply studied. For the nonlinearity of bearings-only target tracking, the MEF-PDAF algorithm is modified, and MEF-PDAF based on unscented transformation (UT-MEF-PDAF) algorithm is proposed. After the deep theory research, simulation and field experiments is conducted to verify and analyze the three algorithms. 5) The problem of AUV single observer multiple targets bearings-only tracking in cluttered environment is more difficult than single target tracking problem in cluttered environment, because it involves the problem of associating multiple measurements with multiple targets. For this problem, the generalized probability data association filter (GPDAF) algorithm and joint probabilistic data association filter (JPDAF) algorithm are deeply studied. After the deep theory research, two simulations are conducted to verify and analyze the two algorithm |
| 语种 | 中文 |
| 公开日期 | 2015-08-31 |
| 产权排序 | 1 |
| 页码 | 57页 |
| 分类号 | TP242.3 |
| 源URL | [http://ir.sia.ac.cn/handle/173321/14814]  |
| 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 | 梅登峰. 杂波环境下AUV纯方位目标跟踪方法研究[D]. 中国科学院沈阳自动化研究所. 2014. |
入库方式： OAI收割
来源：沈阳自动化研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。