【作者】 寇义民；

【导师】 王常虹；

【作者基本信息】 哈尔滨工业大学 ， 控制科学与工程， 2010， 博士

【摘要】 地磁导航作为一种新兴的导航技术,具有不受地形、位置、气候等外部环境限制,可实现全地域、全天候导航的优点,能够有效弥补现有导航方法的不足,因而具有广阔的应用前景。地磁场的测量与测绘、对干扰磁场的分离与补偿、磁图匹配定位技术、以地磁场为参照的组合导航技术是地磁导航这一研究领域中的重要技术难点与热点。本文针对上述问题展开了研究,研究工作主要集中在以下几个方面:对传统航磁校正算法进行了研究和改进。针对经典航磁补偿方法中数学模型系数计算复杂、未考虑载体运动耦合、参数求解易陷入奇异等问题,对航磁补偿数学模型进行了重新推导,以姿态阵和角速度矩阵取代了原始模型中复杂的高阶三角函数来作为模型系数,获得了全姿态航磁校正模型,有效克服了载体运动耦合所带来的问题;在参数求解上,采用滤波的方式替代原始的线性方程求解方式,提高了抗干扰能力并更充分的利用了数据。针对水下磁场测绘载体难以经常上浮修正导航参数而造成的测绘误差过大问题,提出了一种边导航边测绘的地磁测绘新算法。该算法在对第一次测绘航行获得的初始图进行预处理,提取必要信息的基础上,进行二次测绘;在二次测绘中载体利用初始图中的信息,结合惯导输出与当前磁场测量值进行组合滤波,从而达到在校正自身运动参数的同时对初始图进行修正的目的。理论分析和仿真实验证明了该算法可以通过两次低精度测绘获得较高精度的测绘结果。低频电磁场干扰信号因难以通过磁场屏蔽以及带通滤波器等传统技术手段消除,是阻碍地磁导航最终实现的重要障碍。本文针对这个问题提出了一种信号分离与匹配算法。该算法对载体姿态进行了限制,并把复杂干扰源拆分成若干简单等效干扰源,从而将各种磁场信号的幅值叠加由非线性转换成了线性叠加,进而采用独立特征量分析法(ICA)对各种场源信号进行分离,最后用相关系数匹配来解决ICA的幅值与顺序不确定问题。针对传统MAGCOM以及ICCP算法的不足,提出了两个新的匹配算法。一个是基于等高线相交的地磁场、重力场组合定位方法,该方法有效克服了现有匹配方法依赖航迹累积的问题,在实时性方面有所提高。另一个是基于多级子图的磁场曲面线性化定位方法,该方法充分利用了在地磁异常微弱区域磁场空间分布平缓的特点,提出将磁场矢量分解后各自建立位置估测区域,通过各估测区域的交集来获取载体概率上最可能的位置;进而采用多级子方式图由粗到精进行定位;进一步又给出了该方法的适用条件。对基于滤波的地磁/惯性组合导航方法进行了研究和改进。针对标准方法在地磁场曲面缺乏起伏的平缓区域容易发散的问题,提出采用磁场矢量分解对观测方程进行改进以提高系统的局部能观性;针对标准方法在地磁场变化剧烈区线性化误差过大,导致系统性能下降的问题,提出采用UT变换取代随机线性化方法来改进系统性能;针对磁场测量误差统计量不确定问题,提出采用自适应滤波的方式对测量误差方差进行实时跟踪;将相关匹配算法的结果作为系统的观测量,构建组合导航系统,提升系统性能并避免偶然失配现象对系统造成的不良影响。更多还原

【Abstract】 As a new technology the geomagnetic navigation system has several obvious advantages. The most important one is that its effectiveness is not affected by external factors such as weather, location and topographical factors, thus it is an ideal complement to traditional navigation systems and has a broad research and application prospects. The branches of this technology include the following research fields: geomagnetic survey and mapping, disturbance separation, geomagnetic matching algorithms and filtering technologies. This thesis covers several important techniques in the branches above, and its research results are summarized as follows:The traditional aeromagnetic compensation method is improved and a new method for underwater geomagnetic mapping and correction is developed. Focusing on the movement coupling problems in the traditional aeromagnetic compensation method, the compensation model is rebuilt based on rotation matrices and angular velocities, which is much simpler than the old model and can work in all attitudes. Filtering method is applied to substitute equation solving method, which guarantees the new algorithm to have stronger anti-interference ability. Underwater vehicles can’t float frequently to receive GPS signals, so mapping errors are often caused by their inertial navigation systems. Focusing on this problem, a new algorithm for GPS independent geomagnetic mapping method is developed. The algorithm is divided into three steps: Firstly, an inaccurate map is created with magnetic measurement data and position output from inertial navigation systems. Secondly, the map is analyzed and processed to extract shape information. Thirdly, another survey is conducted along the same route, and an extended Kalman filter is constructed with the shape information of the original map, the status of the inertial navigation system as well as the new measurement data, which helps to form a much more accurate map. The effectiveness of the new method is verified by both theoretical analysis and simulation results.The low-frequency electromagnetic interference is one of the major obstacles for geomagnetic navigation systems because it is hard to isolate, model or filter. A new method based on independent components analysis is developed to solve the problem. The features of magnetic field produced by different types of circuits are analyzed, and then the branches of a circuit are divided into several independent interference sources, the attitude of the carrier is also regulated. A linear superposition model of all magnetic compositions is built based on all efforts above, which guarantees the application prerequisites of ICA are fulfilled. A map matching method based on correlation coefficient is proposed to conquer the ambiguity drawbacks of ICA.Focusing on the deficiencies of traditional geomagnetic matching methods, two new algorithms are proposed. One is based on intersected contour lines of geomagnetic field and gravitational field, which doesn’t need the trajectory shape constraint, thus has better real-time performance. The other one is based on linear fitting of multi-level sub maps. This algorithm builds up location estimation areas and utilizes sub maps to perform coarse to fine locating strategy that works efficiently in areas with smooth geomagnetic surfaces. The application conditions of this algorithm are also given.A series of improvements for the traditional integrated geomagnetic navigation system based on EKF are made. To solve the divergence problem caused by flat geomagnetic terrain shapes, vector decomposition is implemented to improve the system’s local observability; To solve the performance degradation caused by big linearization errors, unscented transformation is utilized to replace linearization methods in areas with severe undulating geomagnetic terrains; An adaptive filtering method is carried out to detect and track the switches of disturbance variance, which helps to overcome the uncertainty of sensor noise in complex environment; The new map-matching algorithm 3proposed in this thesis is also used to improve the integrated navigation system’s performance.更多还原