CN115575978B - Grid ionosphere delay correction method and device for user side and receiver - Google Patents
Grid ionosphere delay correction method and device for user side and receiver Download PDFInfo
- Publication number
- CN115575978B CN115575978B CN202211471091.8A CN202211471091A CN115575978B CN 115575978 B CN115575978 B CN 115575978B CN 202211471091 A CN202211471091 A CN 202211471091A CN 115575978 B CN115575978 B CN 115575978B
- Authority
- CN
- China
- Prior art keywords
- grid
- neural network
- ionospheric
- delay
- training
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
- G01S19/01—Satellite radio beacon positioning systems transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
- G01S19/03—Cooperating elements; Interaction or communication between different cooperating elements or between cooperating elements and receivers
- G01S19/07—Cooperating elements; Interaction or communication between different cooperating elements or between cooperating elements and receivers providing data for correcting measured positioning data, e.g. DGPS [differential GPS] or ionosphere corrections
- G01S19/072—Ionosphere corrections
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- General Physics & Mathematics (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computer Networks & Wireless Communication (AREA)
- Position Fixing By Use Of Radio Waves (AREA)
Abstract
The invention relates to a grid ionosphere delay correction method, a device and a receiver of a user side; the delay correction method comprises the following steps: building a neural network for grid ionosphere delay correction; constructing a training set for training a neural network; the training set comprises training sample data and label data corresponding to the training sample; training a neural network by using a training set; and arranging the trained neural network in a navigation receiver of a user side, and when the receiver demodulates the ionosphere vertical delay grid graph at the moment to be corrected, utilizing the neural network to perform grid ionosphere delay correction to obtain the corrected grid ionosphere delay amount. The invention improves the correction precision of the ionospheric delay of the grid points in the central area of the service area, obviously improves the ionospheric delay estimation precision of the edges of the service area and the grid points in the low latitude areas, and can improve the positioning precision of the user by adopting the corrected grid to correct the ionospheric delay.
Description
Technical Field
The invention belongs to the technical field of satellite navigation, and particularly relates to a grid ionosphere delay correction method and device for a user side and a receiver.
Background
The ionospheric delay is an important error source of the satellite navigation positioning system and is also a main correction quantity broadcast by the satellite-based augmentation system. Both the American Wide Area Augmentation System (WAAS) and the Beidou satellite based augmentation System (BDSBAS) under construction use grid correction methods. The existing satellite-based augmentation system mostly adopts plane fitting (Planar) and Kriging (Kriging) to interpolate through point ionosphere delay obtained by observation of a ground reference station to calculate grid point ionosphere delay, quantifies the grid point ionosphere delay, encodes augmentation messages and broadcasts the augmentation messages to users by GEO satellites, and the ionosphere correction precision is higher in a service center area under most situations of broadcasting grids, but the following defects still exist: (1) The existing method only selects penetration points near grid points to participate in calculation, and the fitting precision is sensitive to the measurement error of an individual station or an individual path; (2) When the distribution condition of the penetration points around the grid point is poor, such as insufficient number of the penetration points or over-small elevation angle, the model error of the existing method is obvious, and the estimation precision is poor; (3) Due to the limitation of the message information rate, the calculated grid ionosphere delay is quantized and then is broadcast to the user, and quantization errors exist.
Disclosure of Invention
In view of the above analysis, the present invention aims to disclose a method, an apparatus and a receiver for correcting grid ionospheric delay at a user end, which are used to solve the problem of insufficient accuracy of broadcast grid ionospheric delay of a satellite-based augmentation system.
The invention discloses a grid ionosphere delay correction method of a user side, which comprises the following steps:
s1, building a neural network for grid ionosphere delay correction;
the input of the neural network is that a receiver demodulates an ionospheric vertical delay grid diagram at the moment to be corrected; outputting the corrected grid ionosphere delay amount;
s2, constructing a training set for training a neural network;
the training set comprises training sample data and label data corresponding to the training sample; each training sample data comprises ionospheric vertical delay historical data of a plurality of grid points to be corrected; the label data is standard data corresponding to data to be corrected in the training sample data;
s3, training a neural network by using a training set;
s4, utilizing the trained neural network to perform grid ionosphere delay correction;
the trained neural network is arranged in a navigation receiver of a user side, when the receiver demodulates an ionospheric vertical delay grid chart at a time to be corrected, the neural network is used for correcting grid ionospheric delay to obtain corrected grid ionospheric delay.
Further, the acquiring process of the training sample data includes:
1) Receiving and storing broadcast messages of the satellite-based augmentation system at m historical moments by a plurality of reference stations of the satellite-based augmentation system; the coverage range of the plurality of reference stations comprises a service range for grid ionospheric delay correction for a user;
2) From broadcasting telegramMT26Demodulate out andmcorresponding to individual historical timemA tensile ionospheric vertical delay grid graph;
3) FrommTensile ionospheric vertical delaySelecting from the grid imagemAll the time is effectivenThe ionospheric vertical delay data of the grid points is dimensioned asm×nTraining set matrix ofX m n(×) As training sample data;
further, in the training sample datanAnd the grid points cover the grid points required by the user for grid ionospheric delay correction.
Furthermore, the label data adopts the vertical delay of the ionized layer calculated according to the historical time corresponding to the training sample data provided by the European orbit determination center and the posterior ionized layer total electron content grid graph of the corresponding grid points, and the formed dimension ism×nTraining set label matrix ofY m n(×) 。
Furthermore, the neural network is a single hidden layer fully-connected neural network, and the number of the neurons of the input layer and the output layer is equalnThe number of hidden layer neurons iskThe hidden layer activation function is selected as a non-linear functionf(x) The output layer does not use the activation function.
Further, the loss function in the neural network training process is an average absolute error function; the optimizer is Adam, and training is carried out according to the set learning rate and the batch size;
the trained network parameter matrix of the neural network comprises a connection input layer and a hidden layern×kDimensional weight matrixW 1 ,1×kDimension hidden layer bias vectorB 1 Connecting the hidden layer to the output layerk×nDimensional weight matrixW 2 And 1 is as effective askDimension output layer bias vectorB 2 。
Further onThe number of the neurons of the input layer of the single hidden layer full-connection neural network and the output layer of the hidden layer are alln(ii) a Learning rate for neural network trainingr=0.001, batch size is the full training set.
The invention also discloses a grid ionosphere delay correction device of the user terminal; the grid ionosphere delay correction device comprises a neural network trained in the grid ionosphere delay correction method; and demodulating an ionospheric vertical delay grid graph at the time to be corrected by the input receiver through the neural network to perform grid ionospheric delay correction, and outputting the corrected grid ionospheric delay quantity.
The invention also discloses a receiver; the receiver comprises the grid ionosphere delay correction device of the user end to correct the grid ionosphere delay.
The invention can realize one of the following beneficial effects:
the method solves the problems that the estimation precision of the ionospheric delay of some grid points is reduced due to measurement errors, the estimation precision of the ionospheric delay of grid points at the edge of a service area is poor due to insufficient number of penetration points or over-low elevation angle, and the ionospheric delay correction precision of a grid point at a user end is limited due to quantization errors;
and also,
the user can correct the ionospheric delay grid graph of the receiving demodulation without changing the hardware of the receiver;
the precision of the ionized layer delay grid graph after correction is obviously improved on most grid points; especially the ionospheric delay estimation accuracy of the service area edge and grid points in low latitude areas;
the correction method provided by the invention has small calculation amount and can meet the real-time requirement of the system.
Drawings
The drawings are only for purposes of illustrating particular embodiments and are not to be construed as limiting the invention, wherein like reference numerals are used to designate like parts throughout the drawings;
fig. 1 is a flowchart illustrating a grid ionospheric delay correction method at a user equipment according to an embodiment of the present invention;
fig. 2 is a graph of the loss function value and the number of training rounds in the training process according to the second embodiment of the present invention.
Detailed Description
The preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which form a part hereof, and which together with the embodiments of the invention serve to explain the principles of the invention.
Example one
An embodiment of the present invention discloses a grid ionosphere delay correction method for a user end, as shown in fig. 1, including the following steps:
s1, building a neural network for grid ionosphere delay correction; the input of the neural network is that a receiver demodulates an ionospheric vertical delay grid diagram at the moment to be corrected; outputting the corrected grid ionosphere delay amount;
s2, constructing a training set for training a neural network;
the training set comprises training sample data and label data corresponding to the training sample; each training sample data comprises ionospheric vertical delay historical data of a plurality of grid points to be corrected; the label data is standard data corresponding to data to be corrected in the training sample data;
s3, training a neural network by using a training set;
s4, utilizing the trained neural network to perform grid ionosphere delay correction;
and arranging the trained neural network in a navigation receiver of a user side, and when the receiver demodulates the ionosphere vertical delay grid graph at the moment to be corrected, utilizing the neural network to perform grid ionosphere delay correction to obtain the corrected grid ionosphere delay amount.
Specifically, in step S1, the neural network used for grid ionospheric delay correction is a three-layer fully-connected neural network including a single hidden layer; wherein, the number of neurons in the input layer and the output layer of the neural network are bothnThe number of hidden layer neurons iskHiding the layer miningWith a non-linear activation function, the output layer does not use the activation function.
Specifically, in step S2, in the training set used for training the neural network in step S1, the obtaining of the training sample data includes:
1) Receiving and storing broadcast messages of the satellite-based augmentation system at m historical moments by a plurality of reference stations of the satellite-based augmentation system; the coverage range of the plurality of reference stations comprises a service range for grid ionospheric delay correction for a user;
2) From broadcasting telegramMT26Demodulate out andmcorresponding to individual historical timemA tensile ionospheric vertical delay grid graph;
3) FrommSelecting from the vertical delay grid of the tensile ionospheremAll the time is effectivenThe ionospheric vertical delay data of the grid points is dimensioned asm×nTraining set matrix ofX m n(×) As training sample data;
and, said selectedmAll the time is effectivenIonospheric vertical delay data for individual grid pointsnAnd the grid points cover the grid points required by the user for grid ionosphere delay correction.
Specifically, in step S2, in the training set for training the neural network in step S1, the label data is the time corresponding to the data to be corrected in the training sample data and the vertical delay standard value of the corresponding grid point, and the dimension is established asm×nTraining set label matrix ofY m n(×) ;
Preferably, the tag data is calculated by adopting a grid graph of the total electron content of the ionosphere after events of the corresponding grid points and the historical time corresponding to the training sample data provided by the European orbit determination centerIs formed of dimensions ofm×nTraining set label matrix ofY m n(×) 。
In step S3, the training set and the training set label matrix established in step S2 are used (X,Y) Training the neural network according to the set learning rate and the batch size; the loss function in the training process is an average absolute error function; adam is selected as the optimizer. Training until the loss function converges and tends to be stable; and recording the network parameter matrix at the moment. The network parameter matrix comprises a connection input layer and a hidden layern×kDimensional weight matrixW 1 ,1×kDimension hidden layer bias vectorB 1 Connecting the hidden layer to the output layerk×nDimensional weight matrixW 2 And 1-kDimension output layer offset vectorB 2 。
In a more preferred embodiment, the neural network for grid ionospheric delay correction has the input layer, hidden layer and output layer of each neuron numbern(ii) a Learning rate in training processr=0.001, batch size is the full training set.
Specifically, in step S4, the trained neural network is built in the navigation receiver, and when the receiver demodulates the ionospheric vertical delay grid graph at the time to be corrected, the neural network is used to perform grid ionospheric delay correction to obtain the corrected grid ionospheric delay amount.
During correction, the telegraph text of the satellite-based augmentation system at the moment to be calculated is received and demodulatedMT26To obtain the time of daynIonospheric delay vectors of individual grid points to be correctedD n(1×) (ii) a Will be provided withD n(1×) Inputting the trained neural network to obtain the corrected grid ionosphere delay vectorI n(1×) The calculation process of the neural network can be expressed asI=f(DW 1 +B 1 )W 2 +B 2 。
In summary, the grid ionospheric delay correction method of this embodiment is applied to a satellite navigation satellite-based augmentation system to obtain an ionospheric delay grid map obtained after correction, and compared with a grid map broadcast and demodulated by a GEO satellite of the satellite-based augmentation system, the method can further improve the correction accuracy of ionospheric delay of grid points in a central area of a service area, and significantly improve the ionospheric delay estimation accuracy of grid points in an edge of the service area and a low-latitude area, and the positioning accuracy of a user can be improved by correcting the ionospheric delay by using the corrected grid.
Example two
In this embodiment, the Beidou satellite-based enhancement system ionosphere delay grid is used to specifically implement grid ionosphere delay correction.
The Beidou satellite-based augmentation system (BDSBAS) starts to broadcast a complete ionospheric delay grid map of the Chinese region from 6 months in 2020, and the total number of grid points is 117.
The grid ionospheric delay correction method of this embodiment includes:
s1, building a neural network for grid ionosphere delay correction; the input of the neural network is an ionospheric vertical delay grid graph at the time to be corrected which is demodulated by a receiver; outputting the corrected grid ionosphere delay amount;
specifically, a three-layer (single hidden layer) fully-connected neural network is built by utilizing a TensorFlow machine learning algorithm library in Python 3.7, and the number of neurons of an input layer and an output layer is set to ben=50, number of hidden layer neuronk=50, hidden layer activation function is chosen as nonlinear functionI.e. Sigmoid function, the output layer does not use the activation function.
S2, constructing a training set and a verification set for training a neural network;
receiving and storing BDSBAS enhanced telegraph text from 2020 to 2022 in month 7MTDemodulating 18170 ionospheric delay grid graphs for 23 time points (every integer point except 0 point) every day by 26; selecting 50 grid points with availability greater than 99.9% from 117 grid points in each grid graph, and extracting the unavailable moment of the grid points from the dataRemoved, leaving 18137 Zhang Gewang fig; selecting the period from 6 months in 2020 to 4 months in 2022mThe =16000 Zhang Gewang graph is used in the algorithm training process, and the remaining 2137 Zhang Gewang graph is used for simulating real-time data inspection correction performance; will be selected previouslyn=50 points are determined as grid points to be corrected, and the 50 grid points are guaranteed to be available at the selected 16000 moments; acquiring 16000 post-event ionospheric total electron content grid graphs of 50 corresponding grid points at corresponding time provided by European orbit determination Center (CODE) and calculating ionospheric vertical delay as a standard value; and obtaining corresponding ionospheric delay value standard values of the remaining 2137 test grids for later use.
Establishing a training set matrix with the dimension of 16000 multiplied by 50 by utilizing each obtained 16000 Zhang Gewang graph containing 50 grid point ionospheric delaysX (16000×50) (ii) a Establishing a training set label matrix with dimension of 16000 multiplied by 50 by using 16000 time 50 grid point vertical delay standard values obtained and calculated from European orbit determination center CODEY (16000×50) 。
S3, training a neural network by using a training set;
selecting average absolute error (MAE) as a loss function of the network, selecting Adam as an optimizer, and setting learning rate asrSet as batch size of =0.001m=16000, i.e. the total training set size; using the training set and training set label matrix (S) established in step S2X,Y) Training the neural network built in the step S1 until the loss function value is converged and tends to be stable, wherein the number of training rounds (epochs) at the moment is 10000, and a curve of the loss function value and the number of training rounds in the training process is shown in FIG. 2; recording the network parameter matrix at this time, including a weight matrix having dimension of 50 × 50 and connecting the input layer and the hidden layerW 1 Hidden layer bias vector of dimension 1 × 50B 1 A weight matrix of dimension 50 x 50 connecting the hidden layer and the output layerW 2 And a sum dimension of 1 × 50 output layer bias vectorB 2 The number of network parameters totals 5100.
And step S4: 2137 grid graphs in total are taken as a real-time grid to be corrected at 23 moments (every whole point except 0 point) per day received and demodulated from 5 months to 7 months in 2022;
each piece still contains 50 grid points, and as the moments are moments after the training data, the conditions of simulating real-time data are met; will be firsti(i50 grid points to be corrected ionospheric delays constituting vectors of =1,2, … and 2137) real-time momentsD i (1×50) (ii) a Will be provided withD i Inputting the trained neural network to obtain the corrected grid ionosphere delay vector at the corresponding momentI i (1×50) The calculation process of the neural network can be expressed as:
I i =Sigmoid(D i W 1 +B 1 )+B 2 ;
whereinW 1 ,W 2 ,B 1 ,B 2 Is the parameter matrix recorded in step S3.
According to step S4, 2137 real-time grids are all corrected by the neural network and a new grid graph is recorded, and the broadcast grid and the corrected grid are compared with the ionospheric delay standard grid obtained in step one, which will be described with reference to data.
Firstly, the grid ionospheric delay values directly demodulated by an enhanced telegraph broadcast by a GEO satellite are discrete, the minimum value is 0, and the quantization interval is 0.125 meters, so that errors caused by quantization inevitably exist; the ionospheric delay values corrected by the neural network are continuous and have no quantization error.
Next, this example evaluated 2137 reports of the accuracy of the estimates of the ionospheric delays of the broadcast grid and the corrected grid by comparison with a grid of standard ionospheric delay values, the mean absolute error of the ionospheric delay for each grid point being shown in tables 1 and 2, and the standard deviation of the ionospheric delay error for each grid point being shown in tables 3 and 4, respectively. It can be seen from tables 1 and 2 that the average absolute error of the ionospheric delay grid after correction is significantly reduced except for individual grid points, so that the method has a significant correction effect on the ionospheric delay absolute error of the broadcast grid existing for a long time.
TABLE 1 Mean Absolute Error (MAE) of the tested 2137 Zhang Bofa grid plot versus standard value
TABLE 2 average absolute error (MAE) m of 2137 corrected grid plots versus standard values tested
Finally, it can be seen from tables 3 and 4 that the standard deviation of the ionospheric delay error of the grid after correction also decreases significantly except for individual grid points, especially grid points at the edge of low latitude areas and service areas; the standard deviation of the error represents the stability of the error, and the smaller the standard deviation is, the closer the calculated value is to the distribution of the standard value, so that the correction precision of most of the grid point ionospheric delay after correction by the method is also closer to the reference value.
TABLE 3 error Standard deviation (STD) meters of the tested 2137 Zhang Bofa grid plot versus standard value
TABLE 4 error Standard deviation (STD) meters of 2137 calibrated grid plots versus the standard
EXAMPLE III
The embodiment discloses a grid ionospheric delay correction device at a user end, which includes a neural network trained in the grid ionospheric delay correction method in the first embodiment, demodulates an ionospheric vertical delay grid map at a time to be corrected for an input receiver, corrects the received ionospheric vertical delay grid map, and outputs a corrected grid ionospheric delay amount.
The specific technical details of the grid ionospheric delay correction method in this embodiment are the same as those in the previous embodiment. Please refer to the previous embodiment, which is not repeated herein.
Example four
The embodiment discloses a receiver; in the receiver, a grid ionospheric delay correction device as in the third embodiment is built in to perform grid ionospheric delay correction.
The specific technical details of the grid ionospheric delay correction apparatus in this embodiment are the same as those in the previous embodiment. Please refer to the previous embodiment, which is not repeated herein.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention.
Claims (6)
1. A grid ionospheric delay correction method for a user terminal, comprising:
s1, building a neural network for grid ionosphere delay correction;
the input of the neural network is that a receiver demodulates an ionospheric vertical delay grid diagram at the moment to be corrected; outputting the corrected grid ionosphere delay amount;
the neural network is a single hidden layer full-connection neural network, and the number of the neurons of the input layer and the output layer is equalnThe number of hidden layer neurons iskThe hidden layer activation function is selected as a nonlinear function, and the output layer does not use the activation function;
s2, constructing a training set for training a neural network;
the training set comprises training sample data and label data corresponding to the training sample; each training sample data comprises ionospheric vertical delay historical data of a plurality of grid points to be corrected; the label data is standard data corresponding to data to be corrected in the training sample data;
s3, training a neural network by using a training set;
s4, utilizing the trained neural network to perform grid ionosphere delay correction;
the trained neural network is arranged in a navigation receiver of a user side, and when the receiver demodulates an ionosphere vertical delay grid graph at the time to be corrected, grid ionosphere delay correction is carried out by utilizing the neural network to obtain the corrected grid ionosphere delay amount;
the process of acquiring the training sample data comprises the following steps:
1) Multiple reference station reception and storage by satellite based augmentation systemmBroadcasting a power generation message by the satellite-based augmentation system at each historical moment; the coverage range of the plurality of reference stations comprises a service range for grid ionospheric delay correction for a user;
2) From broadcasting telegramMT26Demodulate andmcorresponding to individual historical timemA tensile ionospheric vertical delay grid graph;
3) FrommSelecting from the vertical delay grid of the tensile ionospheremAll the time is effectivenIonospheric vertical delay data for a grid point is dimensioned asm×nTraining set matrix ofX m n(×) As training sample data;
in the training sample datanThe grid points cover grid points required by a user for grid ionosphere delay correction;
the label data adopts the ionospheric vertical delay calculated according to the historical time corresponding to the training sample data provided by the European orbit determination center and the posterior ionospheric total electron content grid graph of the corresponding grid points, and the formed dimension ism×nTraining set label matrix ofY m n(×) 。
2. The grid ionospheric delay correction method at the user end according to claim 1, wherein,
the loss function in the neural network training process is an average absolute error function; the optimizer is Adam, and training is carried out according to the set learning rate and the batch size;
the trained network parameter matrix of the neural network comprises a connection input layer and a hidden layern×kDimensional weight matrixW 1 ,1×kDimension hidden layer bias vectorB 1 Connecting the hidden layer to the output layerk×nDimension weight matrixW 2 And 1 is as effective askDimension output layer bias vectorB 2 。
4. The grid ionospheric delay correction method at the user end according to claim 3, wherein,
the number of the neurons of the input layer, the hidden layer and the output layer of the single hidden layer fully-connected neural network is alln(ii) a Learning rate for neural network trainingr=0.001, batch size is the full training set.
5. A grid ionospheric delay correction apparatus at a user end, wherein the grid ionospheric delay correction apparatus comprises a neural network trained by the grid ionospheric delay correction method according to any one of claims 1 to 4; and demodulating an ionospheric vertical delay grid graph at the time to be corrected by the input receiver through the neural network to perform grid ionospheric delay correction, and outputting the corrected grid ionospheric delay quantity.
6. A receiver, characterized in that the receiver comprises the grid ionospheric delay correction apparatus at the user end according to claim 5, and performs grid ionospheric delay correction.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211471091.8A CN115575978B (en) | 2022-11-23 | 2022-11-23 | Grid ionosphere delay correction method and device for user side and receiver |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211471091.8A CN115575978B (en) | 2022-11-23 | 2022-11-23 | Grid ionosphere delay correction method and device for user side and receiver |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115575978A CN115575978A (en) | 2023-01-06 |
CN115575978B true CN115575978B (en) | 2023-04-07 |
Family
ID=84590547
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211471091.8A Active CN115575978B (en) | 2022-11-23 | 2022-11-23 | Grid ionosphere delay correction method and device for user side and receiver |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115575978B (en) |
Family Cites Families (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103455702A (en) * | 2012-11-28 | 2013-12-18 | 东南大学 | Method for determining regional ionospheric layer delay |
US20180024248A1 (en) * | 2016-07-20 | 2018-01-25 | Honeywell International Inc. | Systems and methods for nequick modeling using neural networks |
CN108828626B (en) * | 2018-07-02 | 2020-11-06 | 中国人民解放军战略支援部队信息工程大学 | Network RTK ionosphere delay interpolation method and system based on real-time grid |
CN110377969B (en) * | 2019-06-27 | 2023-02-21 | 东南大学 | ELM-based regional ionospheric delay real-time modeling method |
CN111045046B (en) * | 2019-12-03 | 2021-10-19 | 东南大学 | Short-term ionosphere forecasting method and device based on NARX |
CN111880207B (en) * | 2020-07-09 | 2022-12-23 | 南京航空航天大学 | Visual inertial satellite tight coupling positioning method based on wavelet neural network |
US11914047B2 (en) * | 2020-12-14 | 2024-02-27 | CACI, Inc.—Federal | Systems and methods for predicting ionospheric electron content |
CN114152962B (en) * | 2021-11-14 | 2024-05-14 | 中国电子科技集团公司第二十研究所 | Method for determining service range of star-based enhanced system |
CN115293198A (en) * | 2022-06-29 | 2022-11-04 | 中国空间技术研究院 | Method for improving GNSS-R height finding inversion accuracy based on multi-hidden-layer neural network |
-
2022
- 2022-11-23 CN CN202211471091.8A patent/CN115575978B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN115575978A (en) | 2023-01-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Yuan et al. | The BeiDou global broadcast ionospheric delay correction model (BDGIM) and its preliminary performance evaluation results | |
Behrens et al. | A new multi-sensor approach to simulation assisted tsunami early warning | |
CN107390233A (en) | A kind of low orbit satellite navigation enhancing Ionospheric delay correcting parametric technique | |
CN106468774A (en) | A kind of ephemeris star clock being applied to satellite-based augmentation system corrects parameter and spacing wave integrity parameter method | |
Xue et al. | Estimation of differential code biases for Beidou navigation system using multi-GNSS observations: How stable are the differential satellite and receiver code biases? | |
EP3115807A2 (en) | Gnss receiver with an on-board capability to implement an optimal error correction mode | |
Zhang et al. | Evaluation of global modeling of M (3000) F2 and hmF2 based on alternative empirical orthogonal function expansions | |
Maksyutov et al. | A high-resolution inverse modelling technique for estimating surface CO 2 fluxes based on the NIES-TM–FLEXPART coupled transport model and its adjoint | |
CN116204756B (en) | Comprehensive method and system for multi-analysis-center precise station coordinate products | |
JP2010112842A (en) | Method and apparatus for creating tec map and measuring receiver bias | |
JP2008122100A (en) | Ionosphere electron density computing apparatus | |
Uwamahoro et al. | Reconstruction of storm-time total electron content using ionospheric tomography and artificial neural networks: A comparative study over the African region | |
Wang et al. | Tropospheric wet refractivity tomography based on the BeiDou satellite system | |
CN115575978B (en) | Grid ionosphere delay correction method and device for user side and receiver | |
Klobuchar et al. | Comparative range delay and variability of the earth's troposphere and the ionosphere | |
Homam | Initial prediction of Total Electron Content (TEC) at a low latitude station using neural network | |
Pietrella et al. | A comparative study of ionospheric IRIEup and ISP assimilative models during some intense and severe geomagnetic storms | |
CN113093225B (en) | Wide-area and local-area fused high-precision ionospheric scintillation model establishment method | |
CN116165728A (en) | Method and device for estimating atmospheric precipitation, computer equipment and storage medium | |
KR102562273B1 (en) | Apparatus for forecasting 24-hour global total electron content using deep learning based on conditional generative adversarial neural network and method thereof | |
Grishin et al. | Use of ITU-R recommendations in calculating tropospheric signal attenuation in the simulation modeling problems of satellite systems | |
Yang et al. | Overbounding residual zenith tropospheric delays to enhance GNSS integrity monitoring | |
CN111814855A (en) | Global ionospheric total electron content prediction method based on residual seq2seq neural network | |
Erol et al. | Statistical characterization of the ionosphere using GPS signals | |
Kim et al. | Modified kriging based double-difference tropospheric correction interpolation method for Network RTK user |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |