基于自适应卡尔曼滤波的GPS精密单点定位研究

发布时间：2018-08-21 14:24

【摘要】：精密单点定位是指利用IGS等机构提供的高精度轨道和钟差数据,利用单台GNSS接收机所采集的双频伪距和载波相位观测量进行定位解算的高精度绝对定位方法。卡尔曼滤波是精密单点定位普遍采用的解算模式。精密单点定位的解算过程中存在着难点和局限性:第一、周跳探测方法受到观测量类型的限制,数据预处理难度较大。第二、需要考虑众多测量误差然后消除或减弱。第三、整周模糊度的固定难以完成导致收敛速度较慢,定位精度也受到影响。第四、实时精密轨道和钟差文件的实时性难以保证。同时卡尔曼滤波也存在着局限性:在动态精密单点定位中运动模型的噪声可能无法精确模拟或者测量噪声不是正态分布,因此使用卡尔曼滤波进行动态解算结果的精度可能会受到影响。本文采用精密单点定位经典模式,对精密单点定位中的误差项进行减弱或消除;在数据预处理中采用TurboEdit进行周跳探测,并且考虑了钟跳和GPS硬件延迟的影响;采用拉格朗日插值对精密轨道和钟差数据进行加密;在精密单点定位中添加了改进新息序列自适应卡尔曼滤波和自适应抗差卡尔曼滤波两种算法。本文将两种自适应卡尔曼滤波分别与标准卡尔曼滤波进行对比,分析它们能否抑制状态异常和观测异常对定位解算的影响。
[Abstract]:Precise single point positioning is a high-precision absolute positioning method which uses the high-precision orbit and clock difference data provided by IGS and so on, and uses the dual-frequency pseudo-range and carrier phase observations collected by a single GNSS receiver. Kalman filter is widely used in precise single point positioning. There are some difficulties and limitations in the calculation of precise single point positioning. Firstly, the method of cycle slip detection is limited by the type of observation, and the data preprocessing is difficult. Second, a large number of measurement errors need to be considered and then eliminated or weakened. Third, the difficulty of fixing the integer ambiguity leads to the slow convergence speed and the influence of positioning accuracy. Fourth, the real-time precision track and clock error file is difficult to guarantee. At the same time, Kalman filter also has some limitations: the noise of moving model may not be accurately simulated or the measurement noise is not normal distribution in dynamic precise single point positioning. Therefore, the accuracy of dynamic solution using Kalman filter may be affected. In this paper, the classical mode of precision single point positioning is used to reduce or eliminate the error in precision single point positioning, and TurboEdit is used to detect cycle slips in data preprocessing, and the effects of clock slips and GPS hardware delays are taken into account. Lagrangian interpolation is used to encrypt the precision orbit and clock error data, and two improved adaptive Kalman filtering and robust Kalman filtering algorithms are added to the precision single point positioning. In this paper, two kinds of adaptive Kalman filter are compared with standard Kalman filter, and the influence of state anomaly and observation anomaly on location calculation is analyzed.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P228.4

【参考文献】