CASIA OpenIR  > 毕业生  > 硕士学位论文
基于高斯和粒子滤波的地形辅助导航方法研究
其他题名Terrain Aided Navigation Algorithm Using Gaussian Sum Particle Filter
廖威
学位类型工学硕士
导师台宪青
2011-05-28
学位授予单位中国科学院研究生院
学位授予地点中国科学院自动化研究所
学位专业计算机应用技术
关键词地形辅助导航 数字高程模型 非线性非高斯模型 高斯和粒子滤波 混合高斯模型 Terrain Aided Navigation Digital Elevation Model Nonlinear Non-gaussian System Model Gaussian Sum Particle Filter Gaussian Mixture Model
摘要利用地形辅助导航(Terrain Aided Navigation: TAN)技术来进行远程、精确的自主定位与导航是飞行器导航领域的一个重要研究方向。随着飞行器空间活动范围的不断扩展,对其自主导航系统性能的要求越来越高。地形辅助惯性导航系统是一种应用非常广泛的自主导航系统,开展高性能的地形辅助导航方法研究具有非常重要的意义。 论文主要以无人机的地形辅助导航系统为应用背景,开展基于高斯和粒子滤波(Gaussian Sum Particle Filter: GSPF)的地形辅助导航方法研究,并搭建了仿真试验环境,对方法的有效性和可靠性进行了验证。论文主要工作包括: 1. 分析地形高度匹配方法的原理和应用情况,并对现有的地形辅助导航方法进行综述,指出了目前的方法所存在的不足之处。由于地形辅助导航问题本质上是非线性非高斯的滤波估计问题,采用契合该类滤波问题模型的高斯和粒子滤波方法更具优势,而且计算开销相对较低。 2. 对地形辅助导航系统所需的数据进行了建模与仿真。包括:讨论了地形的数学模型,运用随机数字地形生成技术,仿真得到数字地图;分析了DEM的误差模型,用空间自相关随机场模拟DEM与实际地形的误差;运用最近点雷达高度表模型,生成了TAN系统的序列测高数据。 3. 给出了基于高斯和粒子滤波的地形辅助导航方法。基于高斯和粒子滤波的理论,设计了适合TAN系统的地形辅助导航系统新模型和GSPF方法,并在不同特性的地形区域进行了仿真实验,与基于其它典型非线性滤波技术的方法相比,新方法在匹配精度与适应性上具有明显的优势。
其他摘要Using TAN(Terrain Aided Navigation) technology for remote, precise autonomous positioning and navigation has become an important research topic in aircraft navigation area. With the expansion of the aircraft space range and the improving task requirement, there are increasing demands for the performance of their autonomous navigation system. Therefore, it has been very useful and vital to study the high performance TAN algorithms. Based on the application background of the UAVs' TAN system, a concrete research on the TAN algorithm using GSPF (Gaussian sum particle filter) has been carried out and the simulation experiment environment has been built up in order to verify the reliability and effectiveness of the method. The main work is summarized as follows: 1. The principle of TAN and its application area are introduced. By reviewing the recent TAN methods, the main drawbacks are pointed out. Since the TAN problem is essentially a matter of nonlinear non-gaussian filter estimation, there are more advantages to adopt the gaussian sum particle filter method which is coincidence with such filtering problem model. What's more, it brings out higher computational efficiency. 2. The modeling and simulation of the data demanded by the TAN system are carried out. First of all, the terrain map is derived through the discussion of the mathematics model of terrain and the use of random digital terrain generation technology. Then, through the simulation of the difference between the real terrain and DEM data, the error model of the DEM (digital elevation model) is analyzed. Finally, the sequences of the altimeter data are obtained using the nearest point radar altimeter model. 3. A TAN approach based on Gaussian sum particle filter is proposed. Based on the GSPF theory, we build a new TAN dynamic system model and GSPF method for TAN system. Meanwhile, the simulation experiments are carried out in areas with different features. Meanwhile, the simulation experiments are carried out in areas with different features. What is more, compared with the typical methods which based on nonlinear filtering technology, the new method shows obvious advantages in match precision and adaptability.
馆藏号XWLW1649
其他标识符200828014629075
语种中文
文献类型学位论文
条目标识符http://ir.ia.ac.cn/handle/173211/7570
专题毕业生_硕士学位论文
推荐引用方式
GB/T 7714
廖威. 基于高斯和粒子滤波的地形辅助导航方法研究[D]. 中国科学院自动化研究所. 中国科学院研究生院,2011.
条目包含的文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
CASIA_20082801462907(2150KB) 暂不开放CC BY-NC-SA请求全文
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[廖威]的文章
百度学术
百度学术中相似的文章
[廖威]的文章
必应学术
必应学术中相似的文章
[廖威]的文章
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

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。