An on-orbit link identification method for ionospheric anomaly event optimization based on sweeping mode

Abstract

The application discloses a kind of on-orbit link identification ionospheric anomaly event optimization method based on sweeping mode, comprising the following steps: step 1, receiver and GNSS satellite parameters are obtained, link sweeping speed is solved based on link geometry configuration;Step 2, link data is obtained by using receiver measurement, and link data is corrected based on link sweeping speed;Step 3, the identification method of ionospheric anomaly event is established, and the ionospheric anomaly event is identified based on the corrected link data.The optimization method disclosed by the application corrects effective link data based on link sweeping mode, gives more reasonable physical principle for identification method in specific application scenario, and then optimizes the identification method of ionospheric anomaly event, with the beneficial effect of improving identification ability.

Description

Technical Field

[0001] This invention belongs to the field of spaceborne ionospheric sensing and identification, and specifically relates to an optimized method for identifying ionospheric anomalies based on on-board link data. Background Technology

[0002] Human society is entering a new historical era where the space age and the digital age converge and develop. Communication satellites are also entering a golden age of high-quality digitalization and intelligent development. As an organic intelligent system, the "brain" of a communication satellite should include intelligent perception, recognition, and behavioral systems. Among these, the satellite communication environment effect model is a crucial foundational component of the satellite's intelligent recognition system. Extensive observation and research have shown that various ionospheric anomalies (events) can affect the communication quality across multiple frequency bands. Summary of the Invention

[0003] The technical problem to be solved by this invention is to provide an optimized method for identifying ionospheric anomalies in satellite links based on a sweeping approach.

[0004] The present invention adopts the following technical solution:

[0005] An optimized method for identifying ionospheric anomalies in satellite links based on a sweeping approach is improved by including the following steps:

[0006] Step 1: Obtain receiver and GNSS satellite parameters, and solve for the link sweep velocity based on the link geometry.

[0007] Obtain the receiver position, denoted as R, which is the location of the satellite where the receiver is located; obtain the GNSS satellite position, denoted as G.

[0008] Based on the GNSS satellite position G and the receiver position R, establish link L, which points from the GNSS satellite position G to the receiver position R;

[0009] Obtain the receiver velocity, oriented tangentially to the satellite orbit, denoted as V. R ;

[0010] The aforementioned GNSS satellite positions and receiver velocities were obtained through two-planet ephemeris, broadcast ephemeris, and precise ephemeris methods;

[0011] Based on the geometry of link L, the receiver speed V R Solve for the sweep speed of the link:

[0012] Receiver speed V R The component along the L direction of the link is denoted as V. O Receiver speed V R The component perpendicular to the link L direction is denoted as V.N , according to the geometric relationship, we have:

[0013]

[0014] Let the position of the ionospheric anomaly event be P, and the link scanning speed at the ionospheric anomaly event P be V P , according to the geometric relationship, the link scanning speed V P solution formula:

[0015]

[0016] Since the link scanning speed V P changes with time, we have V P =V P (i), where i is a sequence number;

[0017] For the case where the receiver is mounted on a low-earth orbit satellite, the solution formula (2) for link scanning speed is further simplified as follows:

[0018] Based on the actual altitude of the GNSS satellite, GP ≈ 20000km; based on the altitude of the low-earth orbit satellite where the receiver is located and the altitude of the ionospheric anomaly event, PR = [0, 2000km];

[0019] Since PR << GP, the solution formula (2) for link scanning speed is simplified as:

[0020]

[0021] Step 2, using the receiver to measure link data, and correcting the link data based on the link scanning speed:

[0022] Since the link data measured by the receiver changes with time, the obtained link data sequence measured by the receiver is denoted as S(i), where i is a sequence number, and the time of the link data sequence is recorded as t(i), where i is a sequence number;

[0023] Based on the link scanning speed V P (i), the spatial gradient of the link data in the link scanning direction is solved from the link data sequence S(i), and the corrected link data ΔS(i) is obtained, which is expressed as:

[0024]

[0025] It is required that the link scanning speed V P (i) is not less than the link scanning speed threshold V TH , the threshold V TH is set based on experience, and the typical value is 2000m / s;

[0026] Step 3: Establish a method for identifying ionospheric anomalies, based on the corrected link data:

[0027] Based on the corrected link data, a neural network input sample for identifying ionospheric anomalies is established. Based on the neural network input sample, a target sample for the neural network is established accordingly. For input samples with ionospheric anomalies, the target sample value is Y, and for input samples without ionospheric anomalies, the target sample value is N. The number of input layer nodes and the number of hidden layer nodes are set.

[0028] Based on the input and target samples, as well as the number of nodes in the input layer and the number of nodes in the hidden layer, the neural network is trained to obtain the trained neural network f. NN ,

[0029] Using neural network f NN The specific method for identifying ionospheric anomalies is as follows:

[0030] Create a sample to be identified, the number of data points of which should be consistent with the number of nodes in the input layer. Input the sample to be identified into the neural network f. NN In the process, if the output result is Y, it means that the identification result is an ionospheric anomaly event; if the output result is N, it means that the identification result is an ionospheric anomaly event.

[0031] Furthermore, ionospheric anomalies include sudden E-layer events, plasma bubbles, plasma blocks, F-layer irregularities, sudden ionospheric disturbances, proton bursts and polar cap absorption events, ionospheric storms, ionospheric traveling disturbances, equatorial anomalies, the terminator zone, and space activities.

[0032] Furthermore, in step 3, the number of input samples is above 1000, the number of input layer nodes is 120, and the hidden layer is set to two layers, with the number of hidden layer nodes H1 = 30 and the number of hidden layer nodes H2 = 10; training is carried out using a neural network, with the training samples accounting for 70%.

[0033] The beneficial effects of this invention are:

[0034] The optimization method disclosed in this invention corrects effective link data based on link scanning and provides a more reasonable physical principle for the identification method in specific application scenarios, thereby optimizing the identification method for ionospheric anomalies and improving the identification capability.

[0035] The optimization method disclosed in this invention, through the modification and optimization of the identification method, more accurately identifies ionospheric anomalies, which has the beneficial effect of improving the ability of communication satellites to identify environmental effects and supporting the decision-making of the satellite's brain behavior system, and is of great significance to the protection of satellite communication. Attached Figure Description

[0036] Figure 1 This is a flowchart illustrating the optimization method of the present invention;

[0037] Figure 2 This is a schematic diagram of the link configuration used to solve for the link sweep speed;

[0038] Figure 3 This is a graph showing the change in link sweep speed over time;

[0039] Figure 4 This is a diagram of the corrected link data results;

[0040] Figure 5 This is a diagram of a neural network structure for identifying ionospheric anomalies;

[0041] Figure 6 This is a confusion matrix diagram of the recognition results. Detailed Implementation

[0042] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0043] To ensure that the satellite's "brain" system can make timely and accurate decisions, it is necessary to achieve panoramic perception and dynamic (real-time) identification of ionospheric anomalies. The ionosphere, as a vast and complex atmospheric system, frequently experiences various types of anomalies, including sudden E-layer disturbances, plasma bubbles, plasma masses, F-layer irregularities, sudden ionospheric disturbances, proton bursts and polar cap absorption events, ionospheric storms, traveling ionospheric disturbances, equatorial anomalies, the terminator zone, and space activities. When using an onboard radio receiver to receive GNSS (Global Navigation Satellite System) signals from high-orbit satellites, the link data will be immediately affected when the receiver-GNSS link passes through the space region of ionospheric anomalies. By analyzing the affected link data, i.e., the valid data, ionospheric anomalies can be identified. Existing identification methods include those based on analytical criteria and those based on pattern recognition.

[0044] In terms of specific technical issues, due to the extremely high speed of satellite movement, the data detected by the link primarily reflects the migration and changes of ionospheric anomalies rather than local variations. Therefore, the effective data is mainly determined by two factors: the spatial distribution of the ionospheric anomalies themselves and the link's sweeping method. Here, we note that the link sweeping method can be determined through the link configuration. Therefore, if the effective data can be corrected based on the link sweeping method, it is hoped that the ionospheric anomaly identification method can be optimized, thereby improving the identification effect.

[0045] Example 1 discloses an optimized method for identifying ionospheric anomalies in on-board links based on a sweeping approach, such as... Figure 1 As shown, it includes the following steps:

[0046] Step 1: Obtain receiver and GNSS satellite parameters, and solve for the link sweep velocity based on the link geometry.

[0047] To ensure good detection coverage and scanning speed, receivers based on link-based ionospheric detection are usually mounted on low-Earth orbit satellites. The receiver position is obtained and denoted as R. The receiver position is the position of the satellite where the receiver is located.

[0048] Existing GNSS satellites include GPS, BD, GLONASS, and Galileo constellations, typically located in medium to high Earth orbits. The position of a GNSS satellite is denoted as G.

[0049] Based on the GNSS satellite position G and the receiver position R, establish link L, which points from the GNSS satellite position G to the receiver position R;

[0050] Obtain the receiver velocity, oriented tangentially to the satellite orbit, denoted as V. R ;

[0051] The aforementioned GNSS satellite positions and receiver velocities were obtained through two-planet ephemeris, broadcast ephemeris, and precise ephemeris methods;

[0052] Based on the geometry of link L, the receiver speed V R Solve for the sweep speed of the link:

[0053] The geometry of the link is as follows Figure 2 As shown, the receiver speed is decomposed along the link direction and perpendicular to the link direction. Receiver speed V R The component along the L direction of the link is denoted as V. O Receiver speed V R The component perpendicular to the link L direction is denoted as V. N According to geometric relations, we have:

[0054]

[0055] Denote the position of the ionospheric anomaly event as P, and the link scanning velocity at the ionospheric anomaly event P as V P , and the solving expression for the link scanning velocity V is obtained according to the geometric relationship P :

[0056]

[0057] Since the link scanning velocity V P changes with time, we have V P = V P (i), where i is a serial number;

[0058] For the case where the GNSS receiver is mounted on a low-earth orbit satellite, the solving expression (2) of the link scanning velocity is further simplified:

[0059] According to the actual altitude of a medium-high earth orbit GNSS satellite, GP≈20000km is obtained; according to the altitude of the low-earth orbit satellite where the GNSS receiver is located (the typical value is 600km) and the altitude of the ionospheric anomaly event (the typical value is 300km), PR=[0, 2000km] is obtained;

[0060] Since PR<<GP, the solving expression (2) of the link scanning velocity is simplified as:

[0061]

[0062] Step 2: Obtain link data through measurement by the receiver, and correct the link data based on the link scanning velocity:

[0063] Obtain link data through measurement by the receiver. Since the link data measured by the receiver changes with time, a sequence of link data measured by the receiver is obtained, which is denoted as S(i), where i is a serial number. The time instants corresponding to the link data sequence are recorded at the same time, which are denoted as t(i), where i is a serial number;

[0064] Based on the link scanning velocity V P (i), the spatial gradient of the link data in the link scanning direction is solved from the link data sequence S(i), and the corrected link data ΔS(i) is obtained, wherein:

[0065]

[0066] It is noted that when the link scanning velocity is low, the link scanning is not obvious, which is unfavorable for the detection of ionospheric anomaly events. Therefore, requirements can be put forward for the magnitude of the link scanning velocity: the link scanning velocity V P (i) is not less than the link scanning velocity threshold V TH , the threshold V TH is set based on experience, and the typical value is 2000m / s;

[0067] The selected link data is podTec_C002.2010.001.18.15.0034.G09.01_2013.3520_nc. The receiver is on the C002 satellite of the COSMIC constellation, and the GNSS satellite is the G09 satellite of the GPS constellation. The measurement data type is Total Electron Content (TEC). A measurement link is established based on the C002 and G09 satellites. The link sweep speed is obtained using the method described in step 1. The results are as follows... Figure 3 As shown, the upper part is the link data, and the lower part is the change of the link sweep speed over time. The dashed line is the sweep speed threshold.

[0068] Further utilizing the method described in step 2, the corrected link data is obtained. The results are as follows... Figure 4 As shown. This result is the spatial derivative of the link data in the link sweep direction, i.e., the corrected link data.

[0069] Step 3: Establish a method for identifying ionospheric anomalies, based on the corrected link data:

[0070] Ionospheric anomalies include sudden E-layer events, plasma bubbles, plasma masses, F-layer irregularities, sudden ionospheric disturbances, proton bursts and polar cap absorption events, ionospheric storms, ionospheric traveling disturbances, equatorial anomalies, the terminator region, and space activities.

[0071] Based on the corrected link data, establish neural network input samples for identifying ionospheric anomalies. The number of input samples should be as large as possible while ensuring computational performance, and it is recommended to set the value above 1000.

[0072] The number of input layer nodes should be moderate. To improve real-time recognition, the number of input layer nodes should be smaller; to ensure a certain level of recognition accuracy, the number of input layer nodes can be appropriately increased. A value of around 120 is recommended.

[0073] Based on the input samples of the neural network, the target samples of the neural network are established accordingly. For input samples with ionospheric anomalies, the target sample value is Y (Yes), and for input samples without ionospheric anomalies, the target sample value is N (No).

[0074] The hidden layer can be set to a single layer or multiple layers; two layers are recommended. The number of nodes in the hidden layer is H. i Where i is the hidden layer index. It is recommended that the number of hidden layer nodes H be [value missing]. i The values ​​are H1 = 30 and H2 = 10.

[0075] When using neural networks for training, it is recommended that the training samples account for 70% of the total sample size. Then, identify ionospheric anomalies and obtain the identification results.

[0076] Based on the input and target samples, as well as the number of nodes in the input layer and the number of nodes in the hidden layer, the neural network is trained to obtain the trained neural network f. NN ,

[0077] Using neural network f NN The specific method for identifying ionospheric anomalies is as follows:

[0078] Create a sample to be identified, the number of data points of which should be consistent with the number of nodes in the input layer. Input the sample to be identified into the neural network f. NN In the process, if the output result is Y, it means that the identification result is an ionospheric anomaly event; if the output result is N, it means that the identification result is an ionospheric anomaly event.

[0079] Establish a neural network to identify ionospheric anomalies. Figure 5 The neural network architecture diagram is given. The input samples are from data such as the COSMIC constellation, with a total of 9983 samples. The input layer has 120 nodes. There are two hidden layers with 30 and 10 nodes respectively. The neural network is trained, with the training samples accounting for 70% of the total sample size. The trained neural network is then obtained.

[0080] The trained neural network was used to identify ionospheric anomalies and obtain the identification results. Figure 6 The confusion matrix for a single recognition result shows an overall recognition accuracy of 98.3%. After multiple recognitions and averaging, the average overall recognition accuracy is 97.4%. For comparison, the recognition accuracy without using the correction method of this invention was examined, and the value was 96.9%, which is lower than that of this invention.

Claims

1. An optimized method for identifying ionospheric anomalies in satellite on-links based on a sweeping approach, characterized in that, Comprising the following steps: Step 1: acquiring parameters of a receiver and a GNSS satellite, and solving for link sweeping velocity based on a link geometry configuration: acquiring a receiver position, denoted as R, wherein the receiver position is the position of the satellite where the receiver is located; acquiring a GNSS satellite position, denoted as G; establishing a link L based on the GNSS satellite position G and the receiver position R, wherein the link L points from the GNSS satellite position G to the receiver position R; Obtain the receiver velocity, oriented tangentially to the satellite orbit, denoted as V. R ; the above GNSS satellite position and receiver velocity are acquired by means of two-line ephemeris, broadcast ephemeris and precise ephemeris; Based on the geometry of link L, the receiver speed V R Solve for the sweep speed of the link: Receiver speed V R The component along the L direction of the link is denoted as V. O Receiver speed V R The component perpendicular to the link L direction is denoted as V. N According to geometric relations, we have: Let P be the location of the ionospheric anomaly event, and V be the link sweep velocity at point P. P Based on geometric relationships, the link sweeping speed V is obtained. P The solution formula is: Due to the link sweeping speed V P It changes over time, hence V. P =V P (i), where i is the sequence number; for the case where the receiver is mounted on a low-earth orbit satellite, further simplify the solving equation (2) for the link sweeping velocity: according to the actual altitude of the GNSS satellite, GP≈20000km is obtained; according to the altitude of the low-earth orbit satellite where the receiver is located and the altitude of the ionospheric anomaly event, PR=[0, 2000km] is obtained; since PR<<GP, the solving equation (2) for the link sweeping velocity is simplified as: Step 2: measuring link data by using the receiver, and correcting the link data based on the link sweeping velocity: since the link data measured by the receiver changes with time, a sequence of link data measured by the receiver is obtained, denoted as S(i), where i is a sequence number, and the time of the link data sequence is recorded simultaneously, denoted as t(i), where i is a sequence number; Based on link sweep speed V P (i) The spatial gradient of the link data in the link sweep direction is calculated from the link data sequence S(i) to obtain the corrected link data ΔS(i), which is: Required link sweep speed V P (i) Not less than the link sweep speed threshold V TH Threshold V TH Based on experience, the typical value is set at 2000 m / s; Step 3: establishing an identification method for ionospheric anomaly events, and identifying ionospheric anomaly events based on the corrected link data: based on the corrected link data, establishing neural network input samples for identifying ionospheric anomaly events, correspondingly establishing target samples of the neural network according to the input samples of the neural network, wherein for an input sample with an ionospheric anomaly event, the target sample takes a value Y, and for an input sample without an ionospheric anomaly event, the target sample takes a value N, and setting the number of input layer nodes and the number of hidden layer nodes; Based on the input and target samples, as well as the number of nodes in the input layer and the number of nodes in the hidden layer, the neural network is trained to obtain the trained neural network f. NN , Using neural network f NN The specific method for identifying ionospheric anomalies is as follows: Create a sample to be identified, the number of data points of which should be consistent with the number of nodes in the input layer. Input the sample to be identified into the neural network f. NN In the process, if the output result is Y, it means that the identification result is an ionospheric anomaly event; if the output result is N, it means that the identification result is an ionospheric anomaly event.

2. The optimized method for identifying ionospheric anomalies in on-board links based on a sweeping approach according to claim 1, characterized in that, ionospheric anomaly events include sporadic E, plasma bubbles, plasma blobs, F-layer irregularities, sudden ionospheric disturbance, proton storms and polar cap absorption events, ionospheric storms, traveling ionospheric disturbances, equatorial anomaly, the dusk-dawn transition region and space activities.

3. The optimization method for identifying ionospheric anomaly events through on-satellite link based on a sweeping mode according to claim 1, characterized in that: in step 3, the number of input samples is set to be more than 1000, the number of input layer nodes is 120, two hidden layers are arranged, the number of nodes of the first hidden layer H1=30, the number of nodes of the second hidden layer H2=10; training is carried out by using the neural network, and training accounts for 70% of all samples.

Images