基于网格化的单星导航融合定位技术研究

发布时间：2018-07-07 09:54

本文选题：单星定位 + 网格化　；参考：《电子科技大学》2017年硕士论文

【摘要】：目前,全球范围内的导航定位产业快速发展,人们对导航定位服务的需求也愈来愈强烈。卫星导航定位不仅局限于军事领域,还在汽车导航、精密授时、气象观测、海洋救援、地理勘测、载人航天等各个领域有广泛的应用,成为了国民经济建设发展的重要战略支撑。单星定位是卫星导航定位系统中的重要组成部分,因其特有的优点在很多领域都有其独特的作用。目前,单星定位精度主要集中在1~10km,既不能作为一种辅助的民用导航定位手段,也不能在重大灾难面前作为应急导航定位的补充。为此,本文首先介绍了常规导航定位系统的特点,分析了常规多星定位的定位算法和模型;其次,研究了单星导航定位的应用背景和发展趋势,分析了单星导航定位与多星体制相比存在的优点,建立了单星导航定位的系统模型;最后,仿真并分析了不同定位算法下的误差来源和误差传播模型。并从以下两个方面对单星快速导航定位算法的性能进行了优化:1.提高解算速度。通过对地理信息系统(Geographic Information System,GIS)的研究,构建了空天地一体化的网格化模型。设计了基于网格化的多普勒定位算法和基于网格化的伪距定位算法,仿真证明了算法的解算速度要优于单纯的多普勒/伪距定位定位算法。本文还设计了波束扫描定位算法,通过仿真证明其不仅提高了解算速度,还提高了解算的成功率。2.提高算法精度。波束扫描算法虽然提高了解算速度,但是其对卫星位置和半功率角的依赖比较强烈。为此,本文还设计了基于波束扫描的边缘概率加权算法,使波束信息在解算过程中发挥更加“平等”的作用。仿真证明新的算法既提高了算法的精度也提高了速度。针对全天候下的定位问题,本文设计了神经网络与波束扫描的融合定位算法,通过神经网络学习已有的波束信息,构造出一个预测模型,对定位结果进行预测输出。本文提出的基于网格化的单星定位方法有效地改善了单星情况下的快速响应定位解算速度,能够在复杂的电磁环境或恶劣气候条件下使单星定位的误差稳定在1~5km范围内,通过神经网络的预测模型后,算法的精度能够达到百米量级。
[Abstract]:At present, with the rapid development of navigation and positioning industry worldwide, the demand for navigation and positioning services is becoming more and more intense. Satellite navigation and positioning are not only limited to the military field, but also widely used in automotive navigation, precision timing, meteorological observation, ocean rescue, geographical survey, manned spaceflight, and so on. Has become an important strategic support for the development of national economic construction. Single satellite positioning is an important part of satellite navigation and positioning system. At present, the accuracy of single satellite positioning is mainly concentrated at 110km, which can not be used as an auxiliary means of civil navigation and positioning, nor can it be used as a supplement to emergency navigation and positioning in the face of major disasters. Therefore, this paper first introduces the characteristics of conventional navigation and positioning system, analyzes the localization algorithm and model of conventional multi-satellite positioning, and secondly, studies the application background and development trend of single-satellite navigation and positioning. This paper analyzes the advantages of single satellite navigation and positioning compared with multi-star system, and establishes the system model of single satellite navigation and positioning. Finally, the error sources and error propagation models under different positioning algorithms are simulated and analyzed. The performance of the single star fast navigation and location algorithm is optimized from the following two aspects: 1. Improve the speed of solution. Based on the research of Geographic Information system (GIS), a gridded model of the integration of space and earth is constructed. A grid-based Doppler location algorithm and a grid-based pseudo-range location algorithm are designed. The simulation results show that the algorithm is faster than the pure Doppler / pseudo-range location algorithm. This paper also designs a beam-scanning localization algorithm, which not only improves the speed of computing, but also improves the success rate of computing. Improve the accuracy of the algorithm. Although the beam-scanning algorithm improves the computational speed, it is strongly dependent on satellite position and half-power angle. Therefore, the edge probability weighting algorithm based on beam scanning is designed to make the beam information play a more equal role in the calculation process. Simulation results show that the new algorithm not only improves the accuracy of the algorithm, but also improves the speed. In order to solve the problem of all-weather localization, a fusion localization algorithm of neural network and beam scanning is designed in this paper. By learning the existing beam information from neural network, a prediction model is constructed and the location results are predicted and outputted. In this paper, the grid-based single satellite positioning method can effectively improve the solution speed of the fast response location in the single satellite case, and can stabilize the error of the single satellite positioning in the 1~5km range under the complex electromagnetic environment or the bad weather conditions. Through the prediction model of neural network, the accuracy of the algorithm can reach the order of 100 meters.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN966

【相似文献】