基于GNSS的区域电离层模型研究

发布时间：2018-05-03 19:32

本文选题：GNSS + CORS　；参考：《东南大学》2015年博士论文

【摘要】：电离层是日地空间的重要组成部分,电离层的存在使得地球上的生物免受太阳紫外线和x射线的直接辐射作用,但是对在电离层中传播的无线电波产生反射、折射、散射和吸收等效应,从而对无线电通讯、航天活动和全球定位系统的导航定位等都将会产生严重的影响。因此,积极地研究与监测电离层中的各种现象,并揭示现象背后的各种物理机制,从而探索电离层对人类生存环境的影响,是空间天气研究的重要内容,对社会经济的发展十分重要。本文基于电离层探测仪数据和地面GNSS接收站的CORS (Continuously Operating Reference Stations)数据,针对电离层模型研究中存在的关键问题,重点研究了电离层的预报方法和电离层层析方法。本文的主要研究工作和成果有：(1)提出了一种基于误差补偿和神经网络技术的单站临界频率foF2预测的新方法(BP-IRI模型)。利用瑞典利克赛(Lycksele, Sweden)电离层观测站27年(1958-1984年)的电离层观测数据、相应的地磁数据和太阳指数、IRI2012模型的foF2预测数据等,对该方法进行了仔细分析和验证。其结果表明,该方法的平均预测精度较IRI2012模型提高了19.1%。(2)提出了一种基于误差补偿和神经网络技术的电离层TEC/VTEC预测的新方法(BP-DPM模型)。结合传统的区域电离层二维多项式模型(two-dimensional polynomial ionosphere delay correction model,2-DPM)和中国江苏省的CORS数据,通过对新方法的验证和分析,其结果表明,该方法的平均预测精度较2-DPM模型提高了14.0%。(3)提出了一种基于多尺度剖分的电离层层析新方法(MST模型)。该方法将待反演的电离层三维空间按不同的像素尺度进行剖分,由此得到了多个不同的单尺度电离层层析模型(single-scale ionospheric tomography, SST),将这些单尺度模型的未知变量进行统一解算,并根据不同的权重因子,最终加权得到多尺度层析模型的解,即重构待反演区域的电离层电子密度分布。结合GNSS观测数据和IRI2012模型,该方法的可行性、准确性、合理性得到了有效验证,相比于其他传统电离层层析模型,更高精度的电子密度分布通过该方法得以重构,且其重构精度较CSST模型(基于约束代数重构算法(constrained algebraic reconstruction technique, CART)的SST模型)提高了20.8%、较传统的SST模型提高了22.2%、较IRI2012模型提高了77.7%。(4)提出了一种基于可变像素高度的电离层层析新方法(VHCIT模型)。该方法在传统电离层层析模型(SST)基础上,对每个像素尺度的高度增加了一个扰动变量,和待反演的电子密度一并解算;当该方法反演结束后,相当于对待反演区域的电离层结构进行了第二次或多次剖分,并且通过重构得到了不同于传统电离层层析模型的电离层结构。结合GNSS观测数据和IRI2012模型,对该方法的可行性、准确性、合理性进行了验证,相比于其他传统电离层层析模型,更高精度的、更切合实际的电子密度分布通过该方法得以重构,且其重构精度较CSST模型提高了8.6%、较SST模型提高了14.8%、较IRI2012模型提高了75.3%。
[Abstract]:The ionosphere is an important part of the earth's space. The existence of the ionosphere prevents the living of the earth from the direct radiation of the sun's ultraviolet rays and X rays, but produces reflection, refraction, scattering and absorption effects on the radio waves propagating in the ionosphere, so as to navigate to non wire communications, space activities and global positioning systems. It is important to study and monitor all kinds of phenomena in the ionosphere and reveal the various physical mechanisms behind the phenomenon, so as to explore the influence of the ionosphere on the human living environment. It is an important part of the study of the space weather and is very important for the development of the society. This paper is based on the ionosphere detector. The data and the CORS (Continuously Operating Reference Stations) data of the ground GNSS receiving station, in view of the key problems in the ionosphere model research, focus on the method of ionosphere prediction and the ionosphere tomography. The main research work and results of this paper are as follows: (1) an error compensation and neural network technique is proposed. A new method (BP-IRI model) for the single station critical frequency foF2 prediction (BP-IRI model). The ionospheric observation data of 27 years (1958-1984 years) of the Lycksele (Sweden) ionospheric observation station, the corresponding geomagnetic data and the sun index, and the foF2 prediction data of the IRI2012 model are carefully analyzed and verified. The average prediction accuracy of the method is higher than the IRI2012 model 19.1%. (2). A new method of ionospheric TEC/VTEC prediction based on error compensation and neural network technology (BP-DPM model) is proposed. Combined with the traditional regional ionospheric two-dimensional polynomial model (two-dimensional polynomial ionosphere delay correction model, 2-DPM) and the China River The CORS data of Jiangsu Province, through the verification and analysis of the new method, shows that the average prediction accuracy of the method is improved by 14.0%. (3) than the 2-DPM model. A new method of ionospheric tomography based on multiscale division (MST model) is proposed. This method divides the three-dimensional space of the ionized layer into different pixel scales. Several different single scale ionospheric tomography (single-scale ionospheric tomography, SST) are obtained, and the unknown variables of these single scale models are unify, and the solution of the multi-scale tomography model is obtained by the final weighting according to the different weighting factors. That is, the distribution of the ionosphere electron density in the area to be retrieved is reconstructed. Combined with GNSS The feasibility, accuracy and rationality of the method are effectively verified by the observation data and the IRI2012 model. Compared with other traditional ionospheric tomography models, the higher precision of the electronic density distribution is reconstructed by this method, and its reconstruction precision is compared with the CSST model (constrained algebraic reconstruction Tec based on the constraint algebraic reconstruction algorithm. The SST model of hnique, CART) increased by 20.8%, improved by 22.2% compared with the traditional SST model, and improved 77.7%. (4) by IRI2012 model (4). A new method of ionospheric tomography based on variable pixel height (VHCIT model) was proposed. The method added a disturbance to the height of each pixel scale on the basis of the traditional ionospheric tomography (SST). After the inversion is completed, the ionospheric structure of the inversion region is divided into second or more times, and the ionospheric structure different from the traditional ionosphere tomography model is obtained by reconstructing the method. The feasibility of the method is combined with the GNSS observation data and the IRI2012 model. The accuracy and rationality are verified. Compared with other traditional ionospheric tomography models, the more accurate and more practical electronic density distribution can be reconstructed by this method, and its reconstruction accuracy is 8.6% higher than that of the CSST model, 14.8% higher than the SST model, and a higher 75.3%. than the IRI2012 model.

【学位授予单位】：东南大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：P228.4;P352

【相似文献】