Research On Identification Method Of Spaceborne GNSS TEC Disturbance Based On Artificial Neural Network

Ionospheric TEC disturbance can have a serious impact on electronic information systems such as early warning detection,satellite reconnaissance,satellite communication,satellite navigation,short wave communication and so on.The research on the disturbance law of spaceborne TEC and the development of TEC disturbance identification algorithm based on artificial neural network are of great significance to the development and application technology development of spacebased space environment remote sensing technology in China.Based on the simulation data of small satellite GNSS receiver,this paper carries out the time domain,frequency domain and time-frequency domain feature analysis of TEC caused by background ionosphere and equatorial anomaly,and establishes the identification model of ionospheric equatorial anomaly based on artificial neural network.The main research contents are divided into the following three parts:The time domain characteristics of TEC,TEC’s first-order time gradient and TEC’s second-order time gradient caused by background ionosphere and cross equatorial anomalies are analyzed.The typical disturbance characteristics of the cross equatorial anomaly are given,including three peak structure,deepening valley structure,double peaks with close spacing and double peaks with far spacing,as well as the sharp peak structure,gradient inversion structure and wave packet structure generated by TEC time step and second-order time gradient.These disturbance characteristics are significantly different from the TEC variation characteristics in the background ionosphere,It can be used as the input sample set of artificial neural network for training.Using the methods of discrete cosine transform and wavelet transform,the frequency domain characteristics and time-frequency domain characteristics of TEC,TEC’s first-order time gradient and TEC’s second-order time gradient of background ionospheric and cross equatorial anomalies are analyzed.Compared with the equatorial anomaly,the resonance structure of the equatorial anomaly has a stronger amplitude and a narrower bandwidth in the frequency domain,and there are a significant difference between the two for most events.However,some significant time-domain perturbation features will be "weakened" in the frequency and time-frequency domains,so the frequency and time-frequency domain features of equatorial anomalies are not enough to be used as independent means for the identification of equatorial anomalies.A recognition model of ionospheric equatorial anomalies based on artificial neural network is established.When TEC,TEC’s first-order time gradient and TEC’s secondorder time gradient are used as the input samples of neural network,the recognition effect of equatorial anomaly is better,not the recognition effect of equatorial anomaly is poor.The combined judgment of three networks can effectively improve the recognition effect of equatorial anomaly.