Wide-area Ionospheric Disturbance Monitoring And Its Application Based On GNSS

With the rapid development of satellite navigation,scholars have paid more attention to the study of ionosphere,the temporal and spatial variation of ionosphere and anomaly monitoring have gradually become a research hotspot in the field of satellite navigation.Mastering the temporal and spatial characteristics of the ionosphere is very necessary,on the one hand,it can improve the positioning accuracy,and on the other hand it can predict extreme events such as magnetic storms by monitoring the abnormal changes in the ionosphere.Compared with traditional ionospheric monitoring techniques,ionospheric monitoring using the global navigation satellite system(GNSS)has unparalleled advantages,its appearance has greatly promoted the development of ionospheric monitoring.In the wide-area ionosphere monitoring,due to the insufficient number of ground-based observatories and satellites,limited distribution,a large number of grids in the ionosphere with no GNSS rays pass through,making the electron density after inversion less accurate.Aiming at the shortcomings in ionospheric tomography algorithms,this paper proposes a wide-area ionospheric tomography algorithm(WITA)to effectively improve the accuracy of electron density inversion,and proposes a ionospheric forecast model based on deep learning recurrent neural network(RNN),followed by analysis of ionospheric abnormal disturbances during magnetic storms.This paper focuses on three aspects:1?Proposing the WITA,the algorithm uses the smooth distribution of electron density in the grid space,the objective law that the closer the grid is,the more likely it is to meet the same regression model,to recalculate the electron density values for grids that do not have GNSS signal rays crossed,thus overcomes the dependence of the electron density on the initial values.Experiments show that in the wide-area ionosphere monitoring,the new algorithm has significantly better electron density than the traditional multiplicative algebraic reconstruction techniques(MART).2?Traditional ionospheric prediction models have high prediction accuracy when ionosphere is calm,but once the ionosphere is disturbed,the accuracy does not meet the requirements.This paper proposes a ionospheric prediction model based on deep learning RNN,fully exploiting the temporal correlation characteristics of ionospheric TEC data,and combining with solar activity and geomagnetic activity index as input factors to predict ionospheric TEC.3?Analyzing the ionospheric disturbances during the magnetic storm,and comprehensively considering the solar activity before and after the magnetic storm.From the VTEC times series analysis and the global ionospheric TEC anomaly analysis,it is concluded that the ionospheric TEC anomaly intensity is closely related to the geomagnetic index.