CN117592381B - Atmospheric waveguide parameter inversion model training method, device, equipment and medium - Google Patents
Atmospheric waveguide parameter inversion model training method, device, equipment and medium Download PDFInfo
- Publication number
- CN117592381B CN117592381B CN202410070672.3A CN202410070672A CN117592381B CN 117592381 B CN117592381 B CN 117592381B CN 202410070672 A CN202410070672 A CN 202410070672A CN 117592381 B CN117592381 B CN 117592381B
- Authority
- CN
- China
- Prior art keywords
- waveguide
- model
- determining
- parameters
- evaporation
- 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
- 238000012549 training Methods 0.000 title claims abstract description 196
- 238000000034 method Methods 0.000 title claims abstract description 88
- 238000001704 evaporation Methods 0.000 claims abstract description 191
- 230000008020 evaporation Effects 0.000 claims abstract description 191
- 230000007246 mechanism Effects 0.000 claims abstract description 83
- 239000000725 suspension Substances 0.000 claims abstract description 38
- 238000013507 mapping Methods 0.000 claims description 47
- 238000012937 correction Methods 0.000 claims description 38
- 238000001514 detection method Methods 0.000 claims description 31
- 239000013598 vector Substances 0.000 claims description 23
- 238000012545 processing Methods 0.000 claims description 22
- 238000000605 extraction Methods 0.000 claims description 18
- 238000004590 computer program Methods 0.000 claims description 16
- 238000003860 storage Methods 0.000 claims description 14
- 230000004927 fusion Effects 0.000 claims description 10
- 238000010606 normalization Methods 0.000 claims description 10
- 239000005433 ionosphere Substances 0.000 claims description 5
- 230000006870 function Effects 0.000 description 21
- 238000010586 diagram Methods 0.000 description 16
- 238000004891 communication Methods 0.000 description 8
- 230000008569 process Effects 0.000 description 8
- 238000004422 calculation algorithm Methods 0.000 description 6
- 239000005436 troposphere Substances 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 238000005259 measurement Methods 0.000 description 4
- 238000012544 monitoring process Methods 0.000 description 4
- 239000000758 substrate Substances 0.000 description 4
- 230000009466 transformation Effects 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 238000010276 construction Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000007667 floating Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000003746 surface roughness Effects 0.000 description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Chemical compound O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 3
- 238000009825 accumulation Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000001427 coherent effect Effects 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000007935 neutral effect Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000001902 propagating effect Effects 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 238000009834 vaporization Methods 0.000 description 2
- 230000008016 vaporization Effects 0.000 description 2
- 101000822695 Clostridium perfringens (strain 13 / Type A) Small, acid-soluble spore protein C1 Proteins 0.000 description 1
- 101000655262 Clostridium perfringens (strain 13 / Type A) Small, acid-soluble spore protein C2 Proteins 0.000 description 1
- 101000655256 Paraclostridium bifermentans Small, acid-soluble spore protein alpha Proteins 0.000 description 1
- 101000655264 Paraclostridium bifermentans Small, acid-soluble spore protein beta Proteins 0.000 description 1
- 238000010521 absorption reaction Methods 0.000 description 1
- 230000001174 ascending effect Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 238000013517 stratification Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- 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
- G06N3/045—Combinations of networks
-
- 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)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Geometry (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The embodiment of the invention discloses an atmospheric waveguide parameter inversion model training method, device, equipment and medium. The method comprises the following steps: determining a first time delay related quantity of sea surface scattering signals and a second time delay related quantity of occultation direct signals; generating a first sample training set of first delay related quantities for a first sample tag with real evaporation waveguide parameters and a second sample training set of second delay related quantities for a second sample tag with real surface suspended waveguide parameters; the first sample training set and the second sample training set are respectively input into a preset multi-head attention mechanism model for model training, and an evaporation waveguide parameter prediction model, an evaporation waveguide parameter, a surface suspension waveguide parameter prediction model and a surface suspension waveguide parameter can be obtained. According to the technical scheme provided by the embodiment of the invention, the atmospheric waveguide parameters are inverted through the preset multi-head attention mechanism model, so that the performance and accuracy of atmospheric waveguide inversion are further improved.
Description
Technical Field
The embodiment of the invention relates to the technical field of atmospheric waveguides and computers, in particular to an atmospheric waveguide parameter inversion model training method, an atmospheric waveguide parameter inversion model training device, atmospheric waveguide parameter inversion model training equipment and an atmospheric waveguide parameter inversion model training medium.
Background
The refractive index of the atmosphere affects the propagation characteristics of electromagnetic waves in the atmosphere, and when there is a reverse temperature in the troposphere or when water vapor becomes extremely small with the height, the phenomenon of the super-refractive propagation of electromagnetic waves occurs in the troposphere, and the electromagnetic waves repeatedly refract and propagate forward between the upper and lower walls of the atmosphere of the troposphere, as if propagating in a waveguide, and the troposphere is called an atmospheric waveguide. The atmospheric waveguide can be divided into an evaporation waveguide, a surface waveguide and a suspension waveguide from the production mechanism and the distribution height, wherein the evaporation waveguide is generally below 40 meters on the sea surface, the surface waveguide is generally 40-300 meters, and the suspension waveguide is generally 300-3000 meters.
In the prior art, the atmospheric waveguide can be directly detected by means of radar, sounding balloon or rocket, etc., namely, a weather element sensor is utilized to calculate an atmospheric correction refractive index profile, and the position of the atmospheric correction refractive index and the highly negative gradient is the point for generating the atmospheric waveguide. However, the direct detection mode of the atmospheric waveguide adopted in the prior art has the problems of high cost, inconvenient detection, to-be-improved performance and lower accuracy.
Disclosure of Invention
The invention provides an atmospheric waveguide parameter inversion model training method, device, equipment and medium, which are used for improving the performance and accuracy of atmospheric waveguide inversion.
In a first aspect, an embodiment of the present invention provides a method for training an atmospheric waveguide parameter inversion model, where the method includes:
Determining a first time delay related quantity of sea surface scattering signals and determining a second time delay related quantity of occultation direct light signals;
Generating a first sample training set with a first delay related quantity of a first sample tag; the first sample tag includes real evaporation waveguide parameters; and
Generating a second sample training set with a second delay related quantity of a second sample label; the second sample tag comprises real surface suspended waveguide parameters;
inputting the first sample training set into a preset multi-head attention mechanism model for model training to obtain predicted evaporation waveguide parameters output by the model, and performing model training according to the predicted evaporation waveguide parameters and the real evaporation waveguide parameters until a preset first model training ending condition is met to obtain an evaporation waveguide parameter prediction model for inverting the evaporation waveguide parameters in the atmospheric waveguide parameters; and
Inputting the second sample training set into a preset multi-head attention mechanism model for model training to obtain predicted surface suspended waveguide parameters output by the model, and performing model training according to the predicted surface suspended waveguide parameters and the real surface suspended waveguide parameters until the preset second model training ending condition is met to obtain a surface suspended waveguide parameter prediction model for inverting the surface suspended waveguide parameters in the atmospheric waveguide parameters.
In a second aspect, an embodiment of the present invention provides an apparatus for training an atmospheric waveguide parameter inversion model, including:
the correlation quantity determining module is used for determining a first time delay correlation quantity of the sea surface scattering signal and determining a second time delay correlation quantity of the occultation direct signal;
A first training set determining module for generating a first sample training set with a first delay related quantity of a first sample tag; the first sample tag includes real evaporation waveguide parameters; and
A second training set determining module, configured to generate a second sample training set with a second delay related quantity of a second sample label; the second sample tag comprises real surface suspended waveguide parameters;
The evaporation parameter obtaining module is used for inputting the first sample training set into a preset multi-head attention mechanism model to carry out model training to obtain predicted evaporation waveguide parameters output by the model, carrying out model training according to the predicted evaporation waveguide parameters and the real evaporation waveguide parameters until the preset first model training ending condition is met to obtain an evaporation waveguide parameter prediction model, and inverting the evaporation waveguide parameters in the atmospheric waveguide parameters; and
The suspension parameter obtaining module is used for inputting the second sample training set into a preset multi-head attention mechanism model to carry out model training, obtaining predicted surface suspension waveguide parameters output by the model, carrying out model training according to the predicted surface suspension waveguide parameters and the real surface suspension waveguide parameters until the preset second model training ending condition is met, obtaining a surface suspension waveguide parameter prediction model, and inverting the surface suspension waveguide parameters in the atmospheric waveguide parameters.
In a third aspect, an embodiment of the present invention further provides an atmospheric waveguide parameter inversion method, where the method includes:
Determining a first delay correlation quantity to be measured of a sea surface scattering signal to be measured; determining a second time delay related quantity to be detected of the occultation direct signal to be detected;
Inputting the first delay related quantity to be measured into an evaporation waveguide parameter prediction model to obtain evaporation waveguide parameters output by the model; and
Inputting the second delay related quantity to be measured into a surface suspended waveguide parameter prediction model to obtain surface suspended waveguide parameters output by the model;
generating an atmospheric waveguide parameter comprising an evaporation waveguide parameter and a surface-suspended waveguide parameter;
The evaporation waveguide parameter prediction model and the surface suspended waveguide parameter prediction model are respectively generated by adopting an atmospheric waveguide parameter inversion model training method as described above.
In a fourth aspect, an embodiment of the present invention further provides an atmospheric waveguide parameter inversion apparatus, where the apparatus includes:
The to-be-measured correlation quantity determining module is used for determining a first to-be-measured delay correlation quantity of the to-be-measured sea surface scattering signal; determining a second time delay related quantity to be detected of the occultation direct signal to be detected;
The evaporation parameter output module is used for inputting the first delay related quantity to be measured into the evaporation waveguide parameter prediction model to obtain evaporation waveguide parameters output by the model; and
The suspension parameter output module is used for inputting a second delay related quantity to be detected into the surface suspension waveguide parameter prediction model to obtain surface suspension waveguide parameters output by the model;
the waveguide parameter generation module is used for generating atmospheric waveguide parameters including evaporation waveguide parameters and surface suspension waveguide parameters;
The prediction model generation module is used for generating an evaporation waveguide parameter prediction model and a surface suspension waveguide parameter prediction model by adopting the atmospheric waveguide parameter inversion model training device respectively.
In a fifth aspect, an embodiment of the present invention further provides an electronic device, including:
at least one processor; and
A memory communicatively coupled to the at least one processor; wherein,
The memory stores a computer program executable by the at least one processor to enable the at least one processor to perform an atmospheric waveguide parameter inversion model training method or an atmospheric waveguide parameter inversion method of any of the embodiments of the present invention.
In a sixth aspect, embodiments of the present invention further provide a storage medium containing computer-executable instructions that, when executed by a computer processor, enable the computer processor to perform any one of the atmospheric waveguide parameter inversion model training methods or an atmospheric waveguide parameter inversion method provided by the embodiments of the present invention.
According to the embodiment of the invention, the first time delay related quantity of the sea surface scattering signal is determined, and the second time delay related quantity of the occultation direct signal is determined; generating a first sample training set with a first delay related quantity of a first sample tag; the first sample tag includes real evaporation waveguide parameters; and generating a second sample training set with a second delay related quantity of a second sample label; the second sample tag comprises real surface suspended waveguide parameters; inputting the first sample training set into a preset multi-head attention mechanism model for model training to obtain predicted evaporation waveguide parameters output by the model, and performing model training according to the predicted evaporation waveguide parameters and the real evaporation waveguide parameters until a preset first model training ending condition is met to obtain an evaporation waveguide parameter prediction model for inverting the evaporation waveguide parameters in the atmospheric waveguide parameters; and inputting the second sample training set into a preset multi-head attention mechanism model for model training to obtain predicted surface suspended waveguide parameters output by the model, and performing model training according to the predicted surface suspended waveguide parameters and the real surface suspended waveguide parameters until a preset second model training ending condition is met to obtain a surface suspended waveguide parameter prediction model for inverting the surface suspended waveguide parameters in the atmospheric waveguide parameters. According to the technical scheme provided by the embodiment of the invention, the atmospheric waveguide parameters are inverted through the preset multi-head attention mechanism model, so that the performance and accuracy of atmospheric waveguide inversion are further improved.
It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the invention or to delineate the scope of the invention. Other features of the present invention will become apparent from the description that follows.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of an atmospheric waveguide parameter inversion model training method according to a first embodiment of the present invention;
FIG. 2 is a schematic diagram illustrating measurement of GNSS occultation signals according to a first embodiment of the present invention;
FIG. 3 is a schematic diagram of GNSS sea surface scattering signals propagating in an evaporation waveguide according to a first embodiment of the present invention;
FIG. 4 is a schematic diagram of a multi-parameter atmospheric refractive index profile model provided in accordance with a first embodiment of the present invention;
FIG. 5 is a schematic diagram of a multi-head self-attention mechanism according to a first embodiment of the present invention;
FIG. 6 is a schematic diagram of an inversion model based on a self-attention mechanism according to a first embodiment of the present invention;
FIG. 7 is a flow chart of another method for atmospheric waveguide parameter inversion according to a second embodiment of the present invention;
FIG. 8 is a schematic structural diagram of an apparatus for training an inversion model of atmospheric waveguide parameters according to a third embodiment of the present invention;
FIG. 9 is a schematic structural diagram of an atmospheric waveguide parameter inversion apparatus according to a fourth embodiment of the present invention;
Fig. 10 is a schematic structural diagram of an electronic device for an atmospheric waveguide parameter inversion model training method or an atmospheric waveguide parameter inversion method according to a fifth embodiment of the present invention.
Detailed Description
In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and the claims of the present invention and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the invention described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
In addition, it should be noted that, in the technical scheme of the invention, the related processes of collection, storage, use, processing, transmission, provision, disclosure and the like of the related data and the like all conform to the regulations of related laws and regulations and do not violate the popular regulations.
Example 1
Fig. 1 is a flowchart of an atmospheric waveguide parameter inversion model training method according to an embodiment of the present invention, where the embodiment is applicable to inversion of atmospheric waveguide parameters, and the method may be performed by an atmospheric waveguide parameter inversion model training device, and the atmospheric waveguide parameter inversion model training device may be implemented in hardware and/or software, and the atmospheric waveguide parameter inversion model training device may be configured in an electronic device, and the electronic device may be a terminal device or a server, and the embodiment of the present invention is not limited thereto.
As shown in fig. 1, the method for training the atmospheric waveguide parameter inversion model provided by the embodiment of the invention specifically includes the following steps:
s110, determining a first time delay related quantity of the sea surface scattering signal and determining a second time delay related quantity of the occultation direct signal.
Specifically, the global navigation satellite system (Global Navigation SATELLITE SYSTEM, GNSS), also known as a global satellite navigation system, is an air-based radio navigation positioning system that can provide all-weather three-dimensional coordinates and velocity and time information to a user at any location on the earth's surface or near-earth space. The GNSS occultation direct signal refers to a GNSS signal with a zero altitude angle or a negative altitude angle, the sea surface GNSS scattering signal can be modulated by a sea surface evaporation waveguide in a propagation path, and the occultation signal with the zero altitude angle or the negative altitude angle can be modulated by a low-altitude surface waveguide or a suspension waveguide in the propagation process. If a zero or negative altitude GNSS signal can be received, it is indicated that there is a occultation and an atmospheric waveguide is present in the propagation path. In response to a need to invert the atmospheric waveguide parameters, a first time-delay related quantity of the sea surface scattering signal and a second time-delay related quantity of the occultation direct signal can be determined.
Optionally, determining the first time-delay related quantity of the sea surface scattering signal includes: acquiring navigation position information of a Global Navigation Satellite System (GNSS) satellite, receiving position information of a GNSS scattered signal receiving antenna and reflecting position information of a sea surface specular reflection point under a geocentric rectangular coordinate system; and determining a first time delay related quantity of the sea surface scattering signal according to the navigation position information, the receiving position information and the reflection position information.
Specifically, when the first time delay related quantity of the sea surface scattering signal is determined, navigation position information of a Global Navigation Satellite System (GNSS) satellite in a geocentric rectangular coordinate system, receiving position information of a GNSS scattering signal receiving antenna and reflecting position information of a sea surface specular reflection point can be obtained. And then, according to the acquired navigation position information, the received position information and the reflected position information, determining a first time delay related quantity of the sea surface scattering signal. In particular, the embodiment of the invention does not limit the acquisition mode of navigation position information of a Global Navigation Satellite System (GNSS) satellite, the acquisition mode of receiving position information of a GNSS scattered signal receiving antenna and the acquisition mode of reflection position information of a sea surface specular reflection point under a geocentric rectangular coordinate system. In addition, the method for determining the first delay related quantity of the sea surface scattering signal according to the navigation position information, the receiving position information and the reflection position information is not limited.
The navigation position information of the GNSS satellites of the global navigation satellite system under the geocentric rectangular coordinate system can be acquired through a positioning antenna for receiving the satellite navigation positioning information; for the receiving position information of the GNSS scattering signal receiving antenna, the GNSS scattering signal of the left-hand circularly polarized antenna can be used as the receiving antenna to receive the common GNSS signal scattered by the sea surface, and the antenna can be understood as the four-system full-frequency external measuring antenna covering Beidou, a global positioning system (Global Positioning System, GPS), a Gellana satellite navigation system (Global Navigation SATELLITE SYSTEM, GLONASS) and a Galileo satellite navigation system (Galileo Satellite Navigation System), so that the requirements of high precision and multi-system compatibility of measuring equipment can be met.
In an alternative embodiment, when determining the first time delay related quantity of the sea surface scattering signal, the code delay of the scattering signal to the direct signal may be caused by the difference of the wave path of the scattering signal and the direct signal, and the code delay calculation method for the sea surface scattering signal may be that, under the geocentric rectangular coordinate system, the integrated path code delay is calculated by using the navigation position of the GNSS satellite, the receiving position of the GNSS scattering signal receiving antenna and the reflecting position of the sea surface specular reflection point. Exemplary, if the navigation position of the GNSS satellite is/>The receiving position of the GNSS scattered signal receiving antenna is/>The reflection position of the sea surface specular reflection point is/>The code delay is formulated as follows:
；
the code delay of the sea surface scattering signal obtained at this time Can be used as a subsequent input.
Optionally, determining a second delay related quantity of the occultation direct signal includes: acquiring signal power of GNSS occultation; determining the signal additional phase of the GNSS occultation according to the clock error and the position distance between the GNSS occultation signal receiver and the signal transmitter; and determining a second delay related quantity of the occultation direct signal according to the signal power and the signal additional phase.
Specifically, when determining the second delay related quantity of the occultation direct signal, acquiring signal power of GNSS occultation; according to the clock error and the position distance between the GNSS occultation signal receiver and the signal transmitter, determining the signal additional phase of the GNSS occultation; from the signal power and the signal additional phase, a second delay related amount of the occultation direct signal may be determined. In particular, the method for acquiring the signal power of the GNSS occultation, the method for determining the additional phase of the signal of the GNSS occultation and the method for determining the second delay related quantity of the direct signal of the occultation are not limited in the embodiment of the invention. In addition, the embodiment of the invention does not limit the determination mode of the second time delay related quantity for determining the occultation direct signal according to the signal power and the signal additional phase.
When receiving the GNSS occultation direct signal, the right-hand circularly polarized antenna can be used for receiving, in particular, two right-hand circularly polarized antennas which are arranged in a 180-degree back direction can be deployed, the occultation direct signal in two opposite directions is received to increase the frequency of occultation signal monitoring, and when deployed in the open sea area and other regions, the frequency of occultation signal monitoring can be doubled. In addition, a multipoint feeding scheme is adopted for the GNSS occultation signal receiving antenna, so that the coincidence of the antenna phase center and the geometric center is ensured, the positioning accuracy is improved, and the low-zero-altitude angle signal and the negative-altitude angle signal are received; the antenna unit has high gain, wide directional diagram wave beam and built-in low noise amplifying module, and adopts multistage filtering to filter interference signals, thereby ensuring normal operation in severe electromagnetic environment.
In an alternative embodiment, when processing the GNSS occultation signal, the selected GNSS occultation signal processing algorithm may be to use a single differential technique (a difference between a Low Earth Orbit (LEO) satellite and a GNSS occultation observation and a reference GNSS satellite) to eliminate a clock error of the LEO satellite-borne receiver, and obtain an additional phase of the occultation observation by using LEO, a GNSS satellite precision Orbit and GNSS clock error data. In order to track the weak occultation signal, an open loop tracking technology and a coherent accumulation gain technology can be adopted to improve the capturing and tracking sensitivity of the occultation signal. The reason is that in the low troposphere of the earth atmosphere, the atmosphere condition is complex, the water vapor content is rich, and the atmosphere refractive index gradient is large. Propagation of signals in low troposphere is susceptible to effects such as atmospheric refraction, scattering, absorption, multipath effects, super-refraction effects and noise. In addition, the occultation signal receiver can generate various errors in the process of receiving signals. The open loop tracking technology can detect the occultation with lower height, and can track the ascending occultation and descending occultation at the same time, so that the observed quantity of the occultation is improved, and meanwhile, the occultation data obtained by the open loop technology can effectively study the characteristics of the atmospheric boundary layer, and partial key data of the atmospheric boundary layer can be obtained. The coherent accumulation means that in-phase signals and quadrature signals of the correlation result are respectively accumulated, so that the observation sensitivity of weak signals can be improved.
In particular, the GNSS observation phase can be expressed as follows, ignoring the integer ambiguity and measurement noise:
；
Wherein subscript k=1, 2; observing the phase for the signal; c is the speed of light in vacuum; /(I) Is the geometric distance between the transmitter and the receiver at the sea surface; /(I)Generating a phase delay for the atmospheric neutral layer; /(I)Generating a phase delay for the ionosphere; t and T are clock errors of the occultation transmitter and the occultation receiver, respectively.
As can be seen from the GNSS mask signal measurement schematic diagram shown in fig. 2, R is the mask receiver, and a is the earth radius of curvature. When the occultation occurs, the occultation signal receiver observes the occultation GNSS satellite and another non-occultation reference GNSS satellite signal at the same time. On the premise of neglecting ambiguity and noise, the receiver and the occultation GNSS can be used for observing additional phasesAnd observations of reference GNSS satellites/>The difference between them to eliminate the clock error of the receiver, as shown in the following equation.
；
Wherein determining the signal additional phase of the GNSS occultation comprises: determining a first phase delay of the GNSS occultation signal in the atmosphere central layer according to the clock error and the position distance between the GNSS occultation signal receiver and the signal transmitter; determining a second phase delay generated by the GNSS occultation signal in the ionosphere according to the clock error and the position distance between the signal receiver and the signal transmitter of the GNSS reference star; and determining the signal additional phase of the GNSS occultation according to the first phase delay and the second phase delay.
Specifically, when determining the signal additional phase of the GNSS occultation, determining the first phase delay of the GNSS occultation signal in the atmosphere central layer according to the clock error and the position distance between the signal receiver and the signal transmitter of the GNSS occultation; according to the clock error and the position distance between the signal receiver and the signal transmitter of the GNSS reference star, the second phase delay of the GNSS occultation signal generated in the ionosphere can be determined; then, based on the acquired first phase delay and second phase delay, an additional phase of the GNSS occultation signal can be determined.
In particular, the signal-added phase term of the final GNSS occultationThe following formula can be used:
；
Wherein, single difference The method can be directly obtained from observed quantity, wherein in the formula, the superscript o represents a occultation star, and the r represents a reference; /(I)And/>Respectively representing geometric distances among the occultation transmitter, the reference GNSS satellite transmitter and the receiver; Representing a clock error between the occultation satellite and the reference GNSS satellite; /(I) Ionospheric phase delay terms for reference GNSS satellites; /(I)For reference GNSS satellite neutral layer. It should be noted that noise contained in the additional phase signal may affect the inversion result, and thus noise in the original phase needs to be filtered out.
Further, when the power delay related quantity of the occultation direct signal is extracted, the power delay related quantity of the sea surface scattering signal can be extracted as the input of the subsequent inversion, that is to say, the code delay quantity of the sea surface scattering signal can be extracted, and it is required to be noted that the occultation comprises Beidou, GLONASS and Galileo satellite navigation systems. The power delay related quantity of the occultation direct signal can be shown as the following formula:
；
wherein p is the signal power of the received occultation signal; An additional phase calculated for the foregoing; /(I) And/>Is an adjustable coefficient; /(I)For the correlation quantity calculated last. It should be noted that the power and the delay related quantity of the finally obtained occultation signal can be used as the input of the subsequent surface suspended waveguide inversion module.
Acquiring navigation position information of a Global Navigation Satellite System (GNSS) satellite, receiving position information of a GNSS scattered signal receiving antenna and reflection position information of a sea surface specular reflection point under a geocentric rectangular coordinate system; the first delay related quantity of the sea surface scattering signal can be determined according to the acquired navigation position information, the received position information and the reflected position information; acquiring signal power of GNSS occultation; the signal additional phase of the GNSS occultation can be determined according to the clock error and the position distance between the GNSS occultation signal receiver and the signal transmitter; and determining a second time delay related quantity of the occultation direct signal according to the signal power and the signal additional phase, and then carrying out subsequent work according to the acquired first time delay related quantity and the acquired second time delay related quantity.
S120, generating a first sample training set with a first delay correlation quantity of a first sample label; the first sample tag includes the actual evaporation waveguide parameters.
Specifically, after the first delay related quantity as set forth in S110 is obtained, a first sample training set with the first delay related quantity of the first sample tag may be generated, where the first sample tag includes the real evaporation waveguide parameter. The embodiment of the invention does not limit the generation mode of the first sample training set with the first time delay related quantity of the first sample label. In particular, a schematic diagram of the propagation of GNSS sea surface scattering signals at the evaporation waveguide is shown in fig. 3. Wherein the input may be a code delay correlation of the GNSS sea surface scattering signal, availableAnd (3) representing.
It should be noted that, the evaporation waveguide is an atmospheric waveguide phenomenon formed by evaporation of water vapor on the sea surface, the height is usually below 40 meters, the probability of occurrence of the evaporation waveguide on the sea surface is very high, and the atmospheric correction refractive index of the evaporation waveguide can be described by a three-section function modelAnd height/>The functional relationship between them is as follows:
；
Wherein, in the functional relation, five parameters M 0、c1、c2、h1 and h 2 are related together, and the parameter M 0 is expressed in height The corrected refractive index at z 0 is the sea surface roughness length, which is an approximate expression of sea surface roughness fluctuation caused by stormy waves, and is typically z 0 =1.5/>M; c 1 is the waveguide substrate modified refractive index gradient; h 1 is the thickness of the waveguide substrate; c 2 is the gradient of the suspended layer; h 2 is the suspended layer thickness. In particular, in the functional relationship described above, the third stratification is considered to be a standard atmosphere, with a gradient of 0.118m -1, under the condition z > h 1+h2. Evaporative waveguide intensity/>I.e. the modified refractive index at z 0 with/>The difference between the refractive indices is corrected, and the atmospheric waveguide refractive index profile described by the model contains parameters which can describe the information of the height, the intensity and the like of the evaporation waveguide.
Based on the obtained first delay related quantity, a first sample training set with a first delay related quantity of a first sample tag may be generated, wherein the first sample tag comprises a real evaporation waveguide parameter. Further, the subsequent model training work can be carried out according to the generated first sample training set.
S130, generating a second sample training set with a second delay correlation quantity of a second sample label; the second sample tag includes real surface floating waveguide parameters.
Specifically, after the second delay related quantity as set forth in S110 is obtained, a second sample training set with a second delay related quantity of a second sample tag may be generated, where the second sample tag includes a real surface suspended waveguide parameter. The embodiment of the invention does not limit the generation mode of the second sample training set with the second time delay correlation quantity of the second sample label.
The difference from the evaporation waveguide inversion in S120 is that in the surface suspended waveguide inversion, the surface waveguide and suspended waveguide can be inverted using the GNSS occultation direct signal power delay related quantity, the input is the GNSS occultation direct signal power delay related quantity, availableAnd (3) representing.
The atmospheric waveguide can be divided into an evaporation waveguide, a surface waveguide and a suspended waveguide according to the formation of the atmospheric waveguide; the lower boundary of the surface waveguide or the waveguide substrate is connected with the ground surface and is positioned at the middle low altitude of 0-300m, the condition of weather with more stable atmosphere is generally present, at the moment, a more stable temperature inversion layer exists in the lower atmosphere, and the humidity is gradually decreased along with the height; the suspended waveguide is an atmospheric waveguide with a suspended lower boundary, and is generally located at a high altitude above 300 meters, and an inverted temperature layer exists at the high altitude to form the atmospheric waveguide. The atmospheric correction refractive index profile describing the surface waveguide and the suspended waveguide can be expressed by using more parameters, and the adopted multi-parameter atmospheric refractive index profile is shown in fig. 4, and the functional relation expression is shown in the following formula:
；
Wherein ten parameters of M 0、M1、Md、c0、c1、c2、d、zb、zd and z t are related in the functional relation expression, wherein the parameter M 0 is a corrected refractive index at a height of z 0, and z 0 is a sea surface roughness length, and a typical value is z 0 =1.5 M; m 1、c0 and d are parameters describing the evaporation waveguide; c 1 is the graded index of the mixed layer, the variation range is [ -1,0.4]; z d is the evaporation waveguide height; z b is the substrate height of the surface waveguide or suspended waveguide; z t is the flying layer height of the waveguide; /(I)The thickness of the suspension layer is the thickness; m d is the reverse difference of the inversion layer; c 2 is the modified refractive index gradient of the last atmosphere, which can be considered as standard atmosphere, with a typical value of 0.118m -1; intensity of the evaporation waveguide/>I.e., the difference between the modified refractive index at z 0 and the modified refractive index at the evaporation waveguide height z d; waveguide strength of surface waveguide or suspended waveguideI.e. the difference between the modified refractive index at z b and the modified refractive index at z t. The model describes the refractive index profile of the atmospheric waveguide, contains parameters which can describe the information of the height, the intensity and the like of the waveguide, and can simultaneously represent various atmospheric waveguides such as surface suspended waveguides and the like.
Wherein (a) in fig. 4 is used to describe the atmospheric waveguide modified refractive index profile of the vaporization waveguide, and contains parameters that can describe information about the vaporization waveguide height, intensity, etc. (b) The atmospheric waveguide modified refractive index profile used for describing the surface waveguide and the suspended waveguide contains parameters which can describe information such as the heights of the surface waveguide and the suspended waveguide.
Based on the obtained second delay related quantity, a second sample training set with a second delay related quantity of a second sample tag can be generated, wherein the second sample tag comprises a real surface floating waveguide parameter. And further, the subsequent model training work can be carried out according to the generated second sample training set.
S140, inputting the first sample training set into a preset multi-head attention mechanism model for model training to obtain predicted evaporation waveguide parameters output by the model, and performing model training according to the predicted evaporation waveguide parameters and the real evaporation waveguide parameters until a preset first model training ending condition is met to obtain an evaporation waveguide parameter prediction model for inverting the evaporation waveguide parameters in the atmospheric waveguide parameters.
Specifically, when the first sample training set is input into a preset multi-head attention mechanism model to perform model training, predicted evaporation waveguide parameters output by the model can be obtained, model training is performed according to the obtained predicted evaporation waveguide parameters and real evaporation waveguide parameters until a preset first model training ending condition is met, and then an evaporation waveguide parameter prediction model can be obtained and used for inverting the evaporation waveguide parameters in the atmospheric waveguide parameters. In particular, the construction of the preset multi-head attention mechanism model can be based on the existing multi-head attention mechanism model, or can be a multi-head attention mechanism model built by experience customization according to the model construction of related technicians, and the construction mode of the preset multi-head attention mechanism model is not limited in the embodiment of the invention. For the preset first model training ending condition, the model training ending condition may be a model training ending condition set by a related technician, and the embodiment of the invention does not limit the standard of the first model training ending condition.
Where the attention mechanism may be a function that maps a query vector and a set of key-value pairs to an output, essentially a linear weighting of key-value pairs, good at modeling long-term dependent sequences. The self-attention mechanism is a variation of the attention mechanism, and the input query vector and the key value vector are the characteristics, so that the relation among the characteristics can be well learned. The Query (Q), key (Key, K), value (V) and output of the self-attention mechanism are vectors, the output being a weighted sum of V. Wherein the weights assigned to each V are calculated from the query vector and the corresponding keys, the dot product of the query vector and all keys is calculated first, and then each dot product vector is divided by the scale factorScaling,/>The weight value of each value is obtained through the Softmax function as the dimension of the feature, and is shown in the following formula (1). The self-attention function finally linearly weights the value vector with a weight value as shown in the following equation (2):
（1）；
where x i represents the input of the Softmax function.
（2）；
In particular, in order to enhance the feature encoding capability, a multi-head self-attention mechanism is required, in which multiple feature subspaces can be used, the aforementioned scaling dot product self-attention function can be used for calculating each subspace, an input query vector Q, a key K and a value V vector can be obtained by performing linear transformation on an input feature, and then outputs of different heads are spliced on a channel, wherein a structural schematic diagram of the multi-head self-attention mechanism is shown in fig. 5 in detail. This allows the model to focus on information from subspace representations at different locations.
In addition, in order to enhance the feature expression capability and modeling capability, the coding layer based on the multi-head self-attention mechanism in the model consists of multiple layers, each layer consists of the multi-head self-attention mechanism and a fully-connected feedforward network (FeedForward Neural Network), a jump connection and a layer normalization function are used around each sub-layer, the jump connection is used for learning network residual errors, the layer normalization normalizes the features in the channel dimension, and the jump connection and the layer normalization function can avoid gradient disappearance problems. The output of each sub-layer can thus be found as follows:
；
wherein, Is a function of the sub-layer implementation itself; /(I)Is a layer normalization function. To facilitate a jump connection, the output dimension d of all sub-layers in the model, as well as the embedded layer, is a power exponent of 2, which may be 64, 128, 256, 512, for example. When the input is normalized by the Sigmoid function, the selected expression of the Sigmoid function is shown in the following formula:
；
The finally constructed inversion model based on the self-attention layer is shown in fig. 6, wherein N is the number of layers of the multi-head self-attention layer. It should be noted that, in the embodiment of the present invention, the number of layers N of the multi-head self-focusing layer is not limited, and may be set according to the experience of the related technician.
In summary, when training the model according to the predicted evaporation waveguide parameter and the real evaporation waveguide parameter, if a preset first model training ending condition is met, an evaporation waveguide parameter prediction model can be obtained so as to facilitate inversion of the evaporation waveguide parameter in the atmospheric waveguide parameter. In particular, when inversion is performed to obtain parameters describing the evaporation waveguide, the code delay amount of the sea surface scattering GNSS signal after receiving and processing can be extracted based on a learnable model of an established self-attention mechanism, the code delay is input into the self-attention mechanism model after being subjected to position coding, the multi-parameter atmosphere correction refractive index profile parameters describing the evaporation waveguide can be output, the error of the calculated atmosphere correction refractive index parameters and the actual atmosphere correction refractive index parameters is minimized through a gradient descent optimization algorithm, the self-attention model parameters are updated in an iterative mode, and finally the parameters describing the evaporation waveguide are obtained through inversion.
Optionally, the multi-head attention mechanism model is composed of a plurality of multi-head self-attention mechanism coding layers; the multi-head self-attention mechanism coding layer consists of a self-attention network layer and a fully-connected feedforward network layer; correspondingly, inputting the first sample training set into a preset multi-head attention mechanism model for model training to obtain a predicted evaporation waveguide parameter output by the model, wherein the method comprises the following steps: normalizing the first time delay correlation quantity in the first sample training set to obtain a normalized first time delay correlation quantity; inputting the normalized first delay related quantity to a self-attention network layer for feature extraction to obtain a delay feature vector, and carrying out layer normalization processing on the delay feature vector to obtain normalized feature parameters; inputting the normalized characteristic parameters and the normalized first delay related quantity to a fully-connected feedforward network layer for characteristic extraction to obtain a characteristic extraction result; and obtaining the predicted evaporation waveguide parameters output by the model according to the characteristic extraction result of the multi-head self-attention mechanism coding layer.
In particular, the multi-head attention mechanism model may be constructed by a multi-layer multi-head self-attention mechanism coding layer, wherein the multi-head self-attention mechanism coding layer may be constructed by a self-attention network layer and a fully-connected feedforward network layer. When the first sample training set is input into a preset multi-head attention mechanism model for model training, a predicted evaporation waveguide parameter output by the model can be obtained, and in a specific implementation manner of one embodiment, the specific manner of the method can be that first time delay correlation quantity in the first sample training set is normalized, and a corresponding normalized first time delay correlation quantity can be obtained; inputting the obtained normalized first delay related quantity to a self-attention network layer for feature extraction to obtain a delay feature vector, and carrying out layer normalization processing on the delay feature vector to obtain a corresponding normalized feature parameter; after the normalized characteristic parameters and the normalized first delay related quantity are input to a fully-connected feedforward network layer for characteristic extraction, a corresponding characteristic extraction result can be obtained; finally, according to the characteristic extraction result of the multi-head self-attention mechanism coding layer, the predicted evaporation waveguide parameters output by the model can be obtained. The embodiment of the invention does not limit the acquisition mode of predicting the evaporation waveguide parameters.
When the first sample training set is input into a preset multi-head attention mechanism model for model training, the code delay related quantity of the actual GNSS sea surface scattering signal can be obtainedInputting a self-attention model, coding features by using a self-attention mechanism inversion model formed by a plurality of multi-head self-attention layers, namely, based on the self-attention mechanism model, performing position coding on the code delay amount of an input scattering signal to extract time sequence related information, and finally obtaining an atmospheric correction refractive index profile parameter vector/>, which is output by the model, through linear transformationThat is, five parameters M 0、c1、c2、h1 and h 2 describing the evaporation waveguide; the atmospheric correction refractive index profile parameter and the actual atmospheric refractive index profile parameter/>, of the evaporation waveguide in the training data set are obtained through calculationCalculating the error, and obtaining the actual atmospheric refractive index parameter by measurement. The Euclidean distance is selected as the optional error function to measure the error, and the selection of the error function is not limited in the embodiment of the invention.
；
That is, the sample training set is input into a preset multi-head attention mechanism model for model training, the predicted evaporation waveguide parameters output by the model are obtained, model iterative training is carried out according to the predicted evaporation waveguide parameters and the real evaporation waveguide parameters until the model training ending condition is met, and the evaporation waveguide parameter prediction model is obtained. The predicted evaporation waveguide parameters of the output of the model training process areThe real evaporation waveguide parameter is/>According to/>And/>Determining error/>The training is iterated continuously to minimize the error and the model training is ended when a certain value is reached. Wherein the multi-head attention mechanism model is based on/>Output predicted evaporation waveguide parameter is/>The flow of (2) is as follows:
(1) Delaying a code Normalizing by using a Sigmoid function;
(2) Inputting the normalized vector into a multi-layer multi-head self-attention layer for feature coding;
(3) Mapping the encoded vector into a five-dimensional output vector through linear transformation Parameters M 0、c1、c2、h1 and h 2 describing the evaporation waveguide are included.
Iterative updating of the self-attention layer model parameters using a stochastic gradient descent algorithm minimizes errors, resulting in a self-attention mechanism inversion model that can be used to invert parameters describing the evaporation waveguide. The evaporation waveguide parameters can be inferred and inverted by using a self-attention mechanism inversion model obtained through training, and the code delay correlation quantity of GNSS sea surface scattering signals is inputParameters M 0、c1、c2、h1 and h 2 describing the evaporation waveguide are output.
When the evaporation waveguide parameters are inverted by establishing the self-attention mechanism model, the self-attention mechanism can better process time sequences which are dependent for a long time compared with a cyclic neural network, and the most advanced performance is achieved in machine translation and voice recognition.
S150, inputting the second sample training set into a preset multi-head attention mechanism model for model training to obtain predicted surface suspended waveguide parameters output by the model, and performing model training according to the predicted surface suspended waveguide parameters and the real surface suspended waveguide parameters until a preset second model training ending condition is met to obtain a surface suspended waveguide parameter prediction model for inverting the surface suspended waveguide parameters in the atmospheric waveguide parameters.
Specifically, when the second sample training set is input into a preset multi-head attention mechanism model to perform model training, predicted surface suspended waveguide parameters output by the model can be obtained, model training is performed according to the predicted surface suspended waveguide parameters and real surface suspended waveguide parameters until a preset second model training ending condition is met, a surface suspended waveguide parameter prediction model can be obtained, and the model can be used for inverting the surface suspended waveguide parameters in the atmospheric waveguide parameters.
In particular, based on a learnable model of an established self-attention mechanism, after extracting the power delay related quantity of the GNSS occultation direct signal after receiving and processing, the power delay related quantity can be input into the self-attention mechanism model after being subjected to position coding, the multi-parameter atmospheric correction refractive index profile parameters describing the surface suspended waveguide are output, the error of the obtained atmospheric correction refractive index parameters and the actual atmospheric correction refractive index parameters is minimized through a gradient descent optimization algorithm, the self-attention model parameters are updated in an iterative mode, and finally the parameters describing the surface suspended waveguide are obtained through inversion.
The second sample training set is input into a preset multi-head attention mechanism model for model training, and when the predicted surface suspended waveguide parameters output by the model are obtained, the power delay related quantity of the actual GNSS occultation direct signal can be obtainedAfter position coding, the self-attention layer is input, a self-attention mechanism inversion model formed by a plurality of layers of multi-head self-attention layers is output to code the characteristics, namely, based on the self-attention mechanism model, the model performs position coding on the power and time delay related quantity of the input scattering signal to extract time sequence related information, and finally, the atmospheric correction refractive index profile parameter vector/>, which is output by the model, is obtained through linear transformationThat is, ten parameters M 0、M1、Md、c0、c1、c2、d、zb、zd and z t are described; the atmospheric correction refractive index profile parameter and the actual atmospheric refractive index profile parameter/>, of the surface suspended waveguide in the training data set are obtained through calculationThe error is calculated, the Euclidean distance is selected as the optional error function to measure the error, and the error function selection is not limited in the embodiment of the invention.
；
Iterative updating of the self-attention layer model parameters using a stochastic gradient descent algorithm minimizes errors, resulting in a self-attention mechanism inversion model that can be used to invert parameters describing surface suspended waveguides. The surface suspended waveguide parameters can be inverted by using a self-attention mechanism inversion model obtained through training, the power delay correlation quantity of GNSS occultation direct signal is input, and the parameters M 0、M1、Md、c0、c1、c2、d、zb、zd and z t describing the surface suspended waveguide are output.
According to the embodiment of the invention, the navigation position information of a Global Navigation Satellite System (GNSS) satellite under a geocentric rectangular coordinate system, the receiving position information of a GNSS scattered signal receiving antenna and the reflecting position information of a sea surface specular reflection point are obtained; determining a first delay related quantity of sea surface scattering signals according to the navigation position information, the receiving position information and the reflection position information; acquiring signal power of GNSS occultation; determining the signal additional phase of the GNSS occultation according to the clock error and the position distance between the GNSS occultation signal receiver and the signal transmitter; and determining a second delay related quantity of the occultation direct signal according to the signal power and the signal additional phase. Further, a first sample training set with a first delay related quantity of the first sample tag can be generated; the first sample tag comprises a real evaporation waveguide parameter and generates a second sample training set with a second delay related quantity of the second sample tag; the second sample tag includes real surface floating waveguide parameters. Inputting a first sample training set into a preset multi-head attention mechanism model for model training to obtain predicted evaporation waveguide parameters output by the model, and performing model training according to the predicted evaporation waveguide parameters and the real evaporation waveguide parameters until a preset first model training ending condition is met to obtain an evaporation waveguide parameter prediction model for inverting the evaporation waveguide parameters in the atmospheric waveguide parameters; inputting the second sample training set into a preset multi-head attention mechanism model for model training to obtain predicted surface suspended waveguide parameters output by the model, and performing model training according to the predicted surface suspended waveguide parameters and the real surface suspended waveguide parameters until the preset second model training ending condition is met to obtain a surface suspended waveguide parameter prediction model for inverting the surface suspended waveguide parameters in the atmospheric waveguide parameters. According to the technical scheme provided by the embodiment of the invention, the atmospheric waveguide parameters are inverted through the preset multi-head attention mechanism model, so that the performance and accuracy of atmospheric waveguide inversion are further improved.
Example two
Fig. 7 is a flowchart of an atmospheric waveguide parameter inversion method according to a second embodiment of the present invention, where the method may be applicable to inversion of atmospheric waveguide parameters, and the method may be performed by an atmospheric waveguide parameter inversion apparatus, and the atmospheric waveguide parameter inversion apparatus may be implemented in hardware and/or software, and the atmospheric waveguide parameter inversion apparatus may be configured in an electronic device, and the electronic device may be a terminal device or a server, and the embodiment of the present invention is not limited thereto.
As shown in fig. 7, the method for inverting atmospheric waveguide parameters provided by the embodiment of the invention specifically includes the following steps:
S210, determining a first to-be-measured delay related quantity of a sea surface scattering signal to be measured; and determining a second time delay related quantity to be detected of the occultation direct signal to be detected.
Specifically, in response to the requirement of inverting the atmospheric waveguide parameter, a first delay related quantity to be measured of the sea surface scattering signal to be measured and a second delay related quantity to be measured of the occultation direct signal to be measured can be determined. The method for determining the first time delay related quantity to be measured of the sea surface scattering signal to be measured and the method for determining the second time delay related quantity to be measured of the occultation signal to be measured are not limited.
After the first time delay correlation quantity to be measured and the second time delay correlation quantity to be measured are obtained, follow-up work is conducted based on the obtained first time delay correlation quantity to be measured and the second time delay correlation quantity to be measured of the occultation direct signal to be measured.
S220, inputting the first delay related quantity to be measured into an evaporation waveguide parameter prediction model to obtain evaporation waveguide parameters output by the model.
S230, inputting the second delay related quantity to be measured into the surface suspended waveguide parameter prediction model to obtain the surface suspended waveguide parameter output by the model.
Specifically, the first delay related quantity to be measured is input into the evaporation waveguide parameter prediction model, so that the evaporation waveguide parameter output by the model can be obtained. And inputting the second delay related quantity to be measured into the surface suspended waveguide parameter prediction model to obtain the surface suspended waveguide parameter output by the model. The method for acquiring the evaporation waveguide parameters output by the model and the method for acquiring the surface suspended waveguide parameters output by the model are not limited. And further carrying out subsequent processing work according to the obtained evaporation waveguide parameters and the surface suspended waveguide parameters output by the model.
S240, generating atmospheric waveguide parameters including evaporation waveguide parameters and surface suspended waveguide parameters.
Specifically, after the evaporation waveguide parameters and the surface suspended waveguide parameters output by the model are obtained, the corresponding atmospheric waveguide parameters including the evaporation waveguide parameters and the surface suspended waveguide parameters can be generated.
Optionally, after generating the atmospheric waveguide parameter including the evaporation waveguide parameter and the surface-suspended waveguide parameter, the method further includes: determining equipment related information of the frequency equipment to be tested; according to the evaporation waveguide parameters, determining a first mapping relation between the atmospheric correction refractive index of the evaporation waveguide and the sea surface vertical height based on a preset evaporation waveguide mapping relation model; determining a second mapping relation between the atmospheric correction refractive index of the surface suspended waveguide and the sea surface vertical height based on a preset surface suspended waveguide mapping relation model according to the surface suspended waveguide parameters; and determining the equipment furthest detection distance of the frequency equipment to be detected according to the equipment related information, the first mapping relation and the second mapping relation.
Specifically, after generating the atmospheric waveguide parameters including the evaporation waveguide parameters and the surface suspended waveguide parameters, the device related information of the to-be-measured frequency device can be determined, and according to the evaporation waveguide parameters, the first mapping relation between the atmospheric correction refractive index of the evaporation waveguide and the sea surface vertical height can be determined based on a preset evaporation waveguide mapping relation model. And determining a second mapping relation between the atmosphere correction refractive index of the surface suspended waveguide and the sea surface vertical height based on a preset surface suspended waveguide mapping relation model according to the surface suspended waveguide parameters. And according to the equipment related information, the first mapping relation and the second mapping relation, determining the equipment furthest detection distance of the frequency equipment to be detected. The embodiment of the invention does not limit the determination modes of the first mapping relation, the second mapping relation and the furthest detection distance.
The method for determining the equipment furthest detection distance of the frequency equipment to be detected according to the equipment related information, the first mapping relation and the second mapping relation comprises the following steps: determining a fusion correction parameter according to the first mapping relation and the second mapping relation; according to the fusion correction parameters, determining the propagation loss of the frequency equipment to be tested; and determining the equipment furthest detection distance of the frequency equipment to be detected according to the propagation loss and the equipment related information.
Specifically, when determining the furthest detection distance of the equipment of the to-be-detected frequency equipment according to the equipment related information, the first mapping relation and the second mapping relation, the selected determination mode may be to determine the fusion correction parameter according to the first mapping relation and the second mapping relation, and determine the propagation loss of the to-be-detected frequency equipment. And according to the propagation loss and the equipment related information, the equipment furthest detection distance of the frequency equipment to be detected can be determined. The embodiment of the invention does not limit the determination modes of fusion correction parameters, propagation loss and equipment related information.
Particularly, after the atmospheric waveguide parameters including the evaporation waveguide parameters and the surface suspended waveguide parameters are obtained, the detection distance of the frequency-using equipment such as the radar can be predicted, and the prediction mode selected when the detection distance is predicted is not limited. For example, the prediction mode may be that the evaporation waveguide and the surface suspended waveguide are fused first, and the evaporation waveguide section is kept unchanged during fusion, and the evaporation waveguide section below 40 m and the surface suspended waveguide section above 40 m are spliced and fused. In addition, when the parabolic equation or the collective optics is used for calculating the field distribution of the electromagnetic wave, information such as the detection distance of the frequency-using device can be obtained, and the specific steps can be as follows: (1) Receiving and processing GNSS scattered signals and GNSS occultation direct signals; (2) Inverting the evaporation waveguide based on the self-attention mechanism model by using the obtained GNSS scattering signal, and inverting the surface waveguide and the suspended waveguide based on the self-attention mechanism model by using the obtained GNSS occultation direct signal; (3) And fusing the obtained evaporation waveguide with the atmospheric correction refractive index profiles of the surface waveguide and the suspended waveguide, and predicting the detection distance of the frequency equipment by using the obtained atmospheric waveguide information.
In an alternative implementation, illustrated by way of example in calculating the radar detection distance, the step fourier Jie Ruxia of the selected parabolic equation is shown as:
；
Wherein m (x, z) is the atmospheric corrected refractive index at space (x, z), i.e. the fused atmospheric refractive index profile; Wherein/> Is the propagation angle; f and F -1 are Fourier transformation and inverse transformation, so that the propagation loss L of the radar electromagnetic wave can be calculated, and the calculation formula of the propagation loss is as follows:
；
the method for determining the equipment furthest detection distance of the frequency equipment to be detected according to the propagation loss and the equipment related information comprises the following steps: according to the propagation loss and the equipment related information, determining the received signal power of the to-be-tested frequency equipment; if the received signal power is not smaller than the preset minimum received signal threshold, determining the equipment furthest detection distance of the to-be-detected frequency equipment according to the equipment related information.
Specifically, when determining the furthest detection distance of the equipment of the frequency equipment to be tested according to the propagation loss and the equipment related information, the determination mode may be that the received signal power of the frequency equipment to be tested is determined according to the propagation loss and the equipment related information, and when the received signal power is not less than the preset minimum received signal threshold value, the furthest detection distance of the equipment of the frequency equipment to be tested is determined according to the equipment related information, and it is to be noted that the value of the preset minimum received signal threshold value is not limited in the embodiment of the invention.
In an implementation manner of an alternative embodiment, if the selected frequency device to be tested is a radar, the following formula can be referred to for the expression manner of the radar equation:
；
wherein, S min is a radar received signal power which is more than or equal to S; /(I)For the transmit signal power; g is the antenna gain; /(I)Radar scattering area (Radar Cross Section, RCS) for the target; /(I)Is the system loss; l is the propagation loss of the wave at the distance of R.
Therefore, based on the information obtained by the above formula, the furthest detection distance R max of the radar can be determined, and the calculation formula for the furthest detection distance R max of the radar is as follows:
；
Wherein A c is the radar effective aperture area; the minimum detectable received signal for a radar receiver.
S250, respectively generating an evaporation waveguide parameter prediction model and a surface suspension waveguide parameter prediction model by adopting an atmospheric waveguide parameter inversion model training method as described above.
Specifically, the evaporation waveguide parameter prediction model and the surface suspended waveguide parameter prediction model can be respectively generated by adopting the atmospheric waveguide parameter inversion model training method, and the generation modes of the evaporation waveguide parameter prediction model and the surface suspended waveguide parameter prediction model are not limited in the embodiment of the invention.
In particular, after the evaporation waveguide parameter prediction model is obtained, the code delay amount of the GNSS sea surface scattering signal to be detected can be obtained. After the delay related quantity to be measured is input into the trained evaporation waveguide parameter prediction model, the evaporation waveguide related parameter output by the model can be obtained, and then the relation between the atmosphere correction refractive index and the vertical height of the evaporation waveguide, namely the relation between the atmosphere correction refractive index and the vertical height from the sea surface of the evaporation waveguide, is determined based on the function model of the evaporation waveguide according to the evaporation waveguide related parameter.
In addition, after the surface suspended waveguide parameter prediction model is obtained, the power and the phase delay of the to-be-detected GNSS occultation direct signal to be detected can be obtained, the related quantity of the to-be-detected delay is determined, and after the related quantity of the to-be-detected delay is input into the surface suspended waveguide parameter prediction model which is obtained through training, the related parameters of the surface suspended waveguide output by the model can be obtained. According to the acquired related parameters of the surface suspended waveguide, the relation between the atmosphere correction refractive index and the vertical height of the surface suspended waveguide can be determined based on a function model of the surface suspended waveguide.
According to the embodiment of the invention, the first delay related quantity to be measured is input into the evaporation waveguide parameter prediction model, so that the evaporation waveguide parameter output by the model can be obtained; and inputting the second delay related quantity to be measured into the surface suspended waveguide parameter prediction model to obtain the surface suspended waveguide parameter output by the model. The method can further generate atmospheric waveguide parameters including evaporation waveguide parameters and surface suspended waveguide parameters, and then can further determine equipment related information of the to-be-tested frequency equipment; according to the evaporation waveguide parameters, determining a first mapping relation between the atmospheric correction refractive index of the evaporation waveguide and the sea surface vertical height based on a preset evaporation waveguide mapping relation model; determining a second mapping relation between the atmospheric correction refractive index of the surface suspended waveguide and the sea surface vertical height based on a preset surface suspended waveguide mapping relation model according to the surface suspended waveguide parameters; determining a fusion correction parameter according to the first mapping relation and the second mapping relation, and determining the propagation loss of the frequency equipment to be tested; according to the propagation loss and the equipment related information, determining the received signal power of the to-be-tested frequency equipment; if the received signal power is not smaller than the preset minimum received signal threshold, determining the equipment furthest detection distance of the to-be-detected frequency equipment according to the equipment related information. The evaporation waveguide parameter prediction model and the surface suspended waveguide parameter prediction model are respectively generated by adopting an atmospheric waveguide parameter inversion model training method as described above. The technical scheme provided by the embodiment of the invention can effectively utilize GNSS sea surface scattering signals to invert the evaporation waveguide, and utilize GNSS occultation direct signal to invert the surface waveguide and the suspended waveguide to realize the monitoring full coverage of all kinds of atmospheric waveguides, and meanwhile, the system belongs to passive monitoring without actively transmitting signals. During inversion, an inversion model based on a self-attention mechanism is established to invert the atmospheric wave guide, and the self-attention mechanism is very effective for modeling of time sequence information and feature extraction, so that the method is relatively effective for extracting power delay correlation quantity. The self-attention mechanism is led to the field of atmospheric waveguide inversion, so that the performance and accuracy of atmospheric waveguide inversion are further improved, and the technical thought is expanded, and therefore, the GNSS atmospheric waveguide parameter inversion method based on the self-attention mechanism has great practical value and scientific research value.
Example III
Fig. 8 is a schematic structural diagram of an atmospheric waveguide parameter inversion model training device according to a third embodiment of the present invention. As shown in fig. 8, the apparatus for training an atmospheric waveguide parameter inversion model includes: a correlation determination module 310, a first training set determination module 320, a second training set determination module 330, an evaporation parameter obtaining module 340, and a dangling parameter obtaining module 350. Wherein:
a correlation determination module 310, configured to determine a first time delay correlation of the sea surface scattering signal and determine a second time delay correlation of the occultation direct signal;
A first training set determination module 320 for generating a first sample training set with a first delay related quantity of a first sample tag; the first sample tag includes real evaporation waveguide parameters; and
A second training set determining module 330, configured to generate a second sample training set with a second delay related quantity of a second sample label; the second sample tag comprises real surface suspended waveguide parameters;
the evaporation parameter obtaining module 340 is configured to input the first sample training set into a preset multi-head attention mechanism model for model training, obtain a predicted evaporation waveguide parameter output by the model, and perform model training according to the predicted evaporation waveguide parameter and the real evaporation waveguide parameter until a preset first model training end condition is met, obtain an evaporation waveguide parameter prediction model, and be used for inverting the evaporation waveguide parameter in the atmospheric waveguide parameter; and
The suspension parameter obtaining module 350 is configured to input the second sample training set into a preset multi-head attention mechanism model for model training, obtain a predicted surface suspension waveguide parameter output by the model, and perform model training according to the predicted surface suspension waveguide parameter and the real surface suspension waveguide parameter until a preset second model training end condition is met, obtain a surface suspension waveguide parameter prediction model, and be used for inverting the surface suspension waveguide parameter in the atmospheric waveguide parameters.
According to the embodiment of the invention, the first time delay related quantity of the sea surface scattering and scattering signal is determined, and the second time delay related quantity of the occultation direct signal is determined; generating a first sample training set with a first delay related quantity of a first sample tag; the first sample tag includes real evaporation waveguide parameters; and generating a second sample training set with a second delay related quantity of a second sample label; the second sample tag comprises real surface suspended waveguide parameters; inputting the first sample training set into a preset multi-head attention mechanism model for model training to obtain predicted evaporation waveguide parameters output by the model, and performing model training according to the predicted evaporation waveguide parameters and the real evaporation waveguide parameters until a preset first model training ending condition is met to obtain an evaporation waveguide parameter prediction model for inverting the evaporation waveguide parameters in the atmospheric waveguide parameters; and inputting the second sample training set into a preset multi-head attention mechanism model for model training to obtain predicted surface suspended waveguide parameters output by the model, and performing model training according to the predicted surface suspended waveguide parameters and the real surface suspended waveguide parameters until a preset second model training ending condition is met to obtain a surface suspended waveguide parameter prediction model for inverting the surface suspended waveguide parameters in the atmospheric waveguide parameters. According to the technical scheme provided by the embodiment of the invention, the atmospheric waveguide parameters are inverted through the preset multi-head attention mechanism model, so that the performance and accuracy of atmospheric waveguide inversion are further improved.
Optionally, the correlation amount determining module 310 includes:
The first information acquisition unit is used for acquiring navigation position information of a Global Navigation Satellite System (GNSS) satellite, receiving position information of a GNSS scattered signal receiving antenna and reflection position information of a sea surface specular reflection point under a geocentric rectangular coordinate system;
And the first quantity determining unit is used for determining a first time delay related quantity of the sea surface scattering and scattering signals according to the navigation position information, the receiving position information and the reflection position information.
Optionally, the correlation amount determining module 310 includes:
The signal power acquisition unit is used for acquiring the signal power of the GNSS occultation;
the additional phase determining unit is used for determining the additional phase of the GNSS occultation signal according to the clock error and the position distance between the GNSS occultation signal receiver and the signal transmitter;
And a second quantity determining unit for determining a second delay related quantity of the occultation direct signal according to the signal power and the signal additional phase.
Optionally, the additional phase determining unit includes:
The first delay subunit is used for determining a first phase delay of the GNSS occultation signal in the atmosphere central layer according to the clock error and the position distance between the GNSS occultation signal receiver and the signal transmitter;
The second delay subunit is used for determining second phase delay generated by the GNSS occultation signal in the ionosphere according to the clock error and the position distance between the signal receiver and the signal transmitter of the GNSS reference star;
And the phase determining subunit is used for determining the signal additional phase of the GNSS occultation according to the first phase delay and the second phase delay.
Optionally, the multi-head attention mechanism model is composed of a plurality of multi-head self-attention mechanism coding layers; the multi-head self-attention mechanism coding layer consists of a self-attention network layer and a fully-connected feedforward network layer;
Correspondingly, the evaporation parameter obtaining module 340 includes a predicted evaporation parameter obtaining unit, including:
The normalization processing subunit is used for carrying out normalization processing on the first time delay correlation quantity in the first sample training set to obtain a normalized first time delay correlation quantity;
the characteristic parameter obtaining subunit is used for inputting the normalized first delay related quantity to the self-attention network layer for characteristic extraction to obtain a delay characteristic vector, and carrying out layer normalization processing on the delay characteristic vector to obtain a normalized characteristic parameter;
the characteristic extraction subunit is used for inputting the normalized characteristic parameters and the normalized first delay related quantity into the fully-connected feedforward network layer to perform characteristic extraction, so as to obtain a characteristic extraction result;
And the prediction parameter obtaining unit is used for obtaining the prediction evaporation waveguide parameters output by the model according to the characteristic extraction result of the multi-head self-attention mechanism coding layer.
The atmospheric waveguide parameter inversion model training device provided by the embodiment of the invention can execute the atmospheric waveguide parameter inversion model training method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.
Example IV
Fig. 9 is a schematic structural diagram of an atmospheric waveguide parameter inversion apparatus according to a fourth embodiment of the present invention. As shown in fig. 9, the atmospheric waveguide parameter inversion apparatus includes: the device comprises a correlation to be measured determining module 410, an evaporation parameter output module 420, a suspension parameter output module 430, a waveguide parameter generating module 440 and a prediction model generating module 450. Wherein:
the correlation to be measured determining module 410 is configured to determine a first correlation to be measured of a sea surface scattering signal to be measured; determining a second time delay related quantity to be detected of the occultation direct signal to be detected;
The evaporation parameter output module 420 is configured to input the first delay related quantity to be measured to an evaporation waveguide parameter prediction model, so as to obtain an evaporation waveguide parameter output by the model; and
The suspension parameter output module 430 is configured to input a second delay related quantity to be measured to the surface suspension waveguide parameter prediction model, so as to obtain a surface suspension waveguide parameter output by the model;
A waveguide parameter generation module 440 for generating atmospheric waveguide parameters including an evaporation waveguide parameter and a surface-suspended waveguide parameter;
The prediction model generating module 450 is configured to generate the evaporation waveguide parameter prediction model and the surface suspended waveguide parameter prediction model by using an atmospheric waveguide parameter inversion model training device as described above.
The method and the device determine the first delay related quantity to be measured of the sea surface scattering signal to be measured; determining a second time delay related quantity to be detected of the occultation direct signal to be detected; inputting the first delay related quantity to be measured into an evaporation waveguide parameter prediction model to obtain evaporation waveguide parameters output by the model; inputting a second delay related quantity to be detected into the surface suspended waveguide parameter prediction model to obtain surface suspended waveguide parameters output by the model; generating an atmospheric waveguide parameter comprising an evaporation waveguide parameter and a surface-suspended waveguide parameter; the evaporation waveguide parameter prediction model and the surface suspended waveguide parameter prediction model are respectively generated by adopting the atmospheric waveguide parameter inversion model training device. According to the technical scheme provided by the embodiment of the invention, the atmospheric waveguide parameters can be generated through the waveguide parameter prediction model, so that the performance and the accuracy of atmospheric waveguide inversion are further improved.
Optionally, after the waveguide parameter generating module 440, a furthest distance determining module is further included, including:
The device information determining unit is used for determining the device related information of the frequency device to be tested;
The first relation determining unit is used for determining a first mapping relation between the atmosphere correction refractive index of the evaporation waveguide and the sea surface vertical height based on a preset evaporation waveguide mapping relation model according to evaporation waveguide parameters; and
The second relation determining unit is used for determining a second mapping relation between the atmosphere correction refractive index of the surface suspended waveguide and the sea surface vertical height based on a preset surface suspended waveguide mapping relation model according to the surface suspended waveguide parameters;
And the detection distance determining unit is used for determining the equipment furthest detection distance of the frequency equipment to be detected according to the equipment related information, the first mapping relation and the second mapping relation.
Optionally, the detection distance determining unit includes:
the correction parameter determining subunit is used for determining a fusion correction parameter according to the first mapping relation and the second mapping relation;
the transmission loss determining subunit is used for determining the transmission loss of the frequency equipment to be tested according to the fusion correction parameters;
and the distance determining subunit is used for determining the equipment furthest detection distance of the frequency equipment to be detected according to the propagation loss and the equipment related information.
Optionally, the distance determining subunit is specifically configured to:
according to the propagation loss and the equipment related information, determining the received signal power of the to-be-tested frequency equipment;
If the received signal power is not smaller than the preset minimum received signal threshold, determining the equipment furthest detection distance of the to-be-detected frequency equipment according to the equipment related information.
The atmospheric waveguide parameter inversion device provided by the embodiment of the invention can execute the atmospheric waveguide parameter inversion method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.
Example five
Fig. 10 shows a schematic diagram of an electronic device 500 that may be used to implement an embodiment of the invention. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. Electronic equipment may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the inventions described and/or claimed herein.
As shown in fig. 10, the electronic device 500 includes at least one processor 510, and a Memory, such as a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), etc., communicatively connected to the at least one processor 510, wherein the Memory stores a computer program executable by the at least one processor, and the processor 510 can perform various suitable actions and processes according to the computer program stored in the ROM520 or the computer program loaded from the storage unit 580 into the RAM 530. In RAM530, various programs and data required for the operation of electronic device 500 may also be stored. The processor 510, ROM520, and RAM530 are connected to each other by a bus 540. An Input/Output (I/O) interface is also connected to bus 540.
Various components in electronic device 500 are connected to I/O interface 550, including: an input unit 560 such as a keyboard, a mouse, etc.; an output unit 570 such as various types of displays, speakers, and the like; a storage unit 580 such as a magnetic disk, an optical disk, or the like; and a communication unit 590 such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 590 allows the electronic device 500 to exchange information/data with other devices through a computer network such as the internet and/or various telecommunication networks.
Processor 510 may be a variety of general-purpose and/or special-purpose processing components having processing and computing capabilities. Some examples of processor 510 include, but are not limited to, a central Processing unit (Central Processing Unit, CPU), a graphics Processing unit (Graphics Processing Unit, GPU), various specialized artificial intelligence (ARTIFICIAL INTELLIGENCE, AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DIGITAL SIGNAL Processing, DSP), and any suitable processor, controller, microcontroller, etc. The processor 510 performs the various methods and processes described above, such as an atmospheric waveguide parametric inversion model training method or an atmospheric waveguide parametric inversion method.
In some embodiments, an atmospheric waveguide parametric inversion model training method or an atmospheric waveguide parametric inversion method may be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as storage unit 580. In some embodiments, part or all of the computer program may be loaded and/or installed onto electronic device 500 via ROM520 and/or communication unit 590. When the computer program is loaded into RAM530 and executed by processor 510, one or more of the steps of an atmospheric waveguide parametric inversion model training method or an atmospheric waveguide parametric inversion method described above may be performed. Alternatively, in other embodiments, processor 510 may be configured to perform an atmospheric waveguide parameter inversion model training method or an atmospheric waveguide parameter inversion method in any other suitable manner (e.g., by means of firmware).
Various implementations of the systems and techniques described here above can be implemented in digital electronic circuitry, integrated circuitry, field programmable gate array (Field Programmable GATE ARRAY, FPGA), application-specific integrated Circuit (ASIC), application-specific standard product (Application SPECIFIC STANDARD PARTS, ASSP), system-on-Chip (SoC), complex programmable logic device (Complex Programmable logic device, CPLD), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.
A computer program for carrying out methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions/acts specified in the flowchart and/or block diagram block or blocks to be implemented. The computer program may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.
In the context of the present invention, a computer-readable storage medium may be a tangible medium that can contain, or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, the computer readable storage medium may be a machine readable signal medium. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, RAM530, ROM520, an erasable programmable read-Only Memory (Erasable Programmable Read Only Memory, EPROM) or flash Memory, an optical fiber, a compact disc read-Only Memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having: a display device for displaying information to a user, for example, a Cathode Ray Tube (CRT) or a Liquid crystal display (Liquid CRYSTAL DISPLAY, LCD); and a keyboard and a pointing device (e.g., a mouse or a trackball) through which a user can provide input to the electronic device. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local area network (Local Area Network, LAN), wide area network (Wide Area Network, WAN), blockchain network, and the internet.
The computing system may include clients and servers. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical host and Virtual special server (Virtual PRIVATE SERVER, VPS) are overcome.
It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps described in the present invention may be performed in parallel, sequentially, or in a different order, so long as the desired results of the technical solution of the present invention are achieved, and the present invention is not limited herein.
The above embodiments do not limit the scope of the present invention. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of the present invention.
Claims (11)
1. The atmospheric waveguide parameter inversion model training method is characterized by comprising the following steps of:
Determining a first time delay related quantity of sea surface scattering signals and determining a second time delay related quantity of occultation direct light signals;
Generating a first sample training set with a first delay related quantity of a first sample tag; the first sample tag includes real evaporation waveguide parameters; and
Generating a second sample training set with a second delay related quantity of a second sample label; the second sample tag comprises real surface suspended waveguide parameters;
Inputting the first sample training set into a preset multi-head attention mechanism model for model training to obtain predicted evaporation waveguide parameters output by the model, and performing model training according to the predicted evaporation waveguide parameters and the real evaporation waveguide parameters until a preset first model training ending condition is met to obtain an evaporation waveguide parameter prediction model for inverting the evaporation waveguide parameters in the atmospheric waveguide parameters; and
Inputting the second sample training set into a preset multi-head attention mechanism model for model training to obtain predicted surface suspended waveguide parameters output by the model, and performing model training according to the predicted surface suspended waveguide parameters and the real surface suspended waveguide parameters until a preset second model training ending condition is met to obtain a surface suspended waveguide parameter prediction model for inverting the surface suspended waveguide parameters in the atmospheric waveguide parameters;
wherein said determining a first time delay related quantity of the sea surface scattering signal comprises:
Acquiring navigation position information of a Global Navigation Satellite System (GNSS) satellite, receiving position information of a GNSS scattered signal receiving antenna and reflecting position information of a sea surface specular reflection point under a geocentric rectangular coordinate system;
determining a first delay related quantity of sea surface scattering signals according to the navigation position information, the receiving position information and the reflection position information;
wherein determining the second delay related quantity of the occultation direct signal comprises:
acquiring signal power of GNSS occultation;
determining the signal additional phase of the GNSS occultation according to the clock error and the position distance between the GNSS occultation signal receiver and the signal transmitter;
And determining a second time delay related quantity of the occultation direct signal according to the signal power and the signal additional phase.
2. The method for training an atmospheric waveguide parametric inversion model according to claim 1, wherein determining the signal additional phase of the GNSS mask comprises:
determining a first phase delay of the GNSS occultation signal in the atmosphere central layer according to the clock error and the position distance between the GNSS occultation signal receiver and the signal transmitter;
Determining a second phase delay generated by the GNSS occultation signal in the ionosphere according to the clock error and the position distance between the signal receiver and the signal transmitter of the GNSS reference star;
And determining the signal additional phase of the GNSS occultation according to the first phase delay and the second phase delay.
3. The atmospheric waveguide parameter inversion model training method according to claim 1, wherein said multi-headed attention mechanism model is composed of a plurality of multi-headed self-attention mechanism coding layers; the multi-head self-attention mechanism coding layer consists of a self-attention network layer and a fully-connected feedforward network layer;
correspondingly, the step of inputting the first sample training set into a preset multi-head attention mechanism model for model training to obtain the predicted evaporation waveguide parameters output by the model comprises the following steps:
normalizing the first time delay related quantity in the first sample training set to obtain a normalized first time delay related quantity;
Inputting the normalized first delay related quantity to a self-attention network layer for feature extraction to obtain a delay feature vector, and carrying out layer normalization processing on the delay feature vector to obtain normalized feature parameters;
inputting the normalized characteristic parameters and the normalized first delay related quantity to the fully-connected feedforward network layer for characteristic extraction to obtain a characteristic extraction result;
And obtaining the predicted evaporation waveguide parameters output by the model according to the characteristic extraction result of the multi-head self-attention mechanism coding layer.
4. An atmospheric waveguide parameter inversion method, comprising:
Determining a first delay correlation quantity to be measured of a sea surface scattering signal to be measured; determining a second time delay related quantity to be detected of the occultation direct signal to be detected;
inputting the first delay related quantity to be measured into an evaporation waveguide parameter prediction model to obtain evaporation waveguide parameters output by the model; and
Inputting the second delay related quantity to be detected into a surface suspended waveguide parameter prediction model to obtain surface suspended waveguide parameters output by the model;
Generating an atmospheric waveguide parameter comprising the evaporation waveguide parameter and the surface suspended waveguide parameter;
wherein the evaporation waveguide parameter prediction model and the surface suspended waveguide parameter prediction model are respectively generated by adopting the atmospheric waveguide parameter inversion model training method according to any one of claims 1-3.
5. The method of atmospheric waveguide parameter inversion according to claim 4, further comprising, after said generating atmospheric waveguide parameters including said evaporation waveguide parameters and said surface-suspended waveguide parameters:
Determining equipment related information of the frequency equipment to be tested;
According to the evaporation waveguide parameters, determining a first mapping relation between the atmospheric correction refractive index of the evaporation waveguide and the sea surface vertical height based on a preset evaporation waveguide mapping relation model; and
Determining a second mapping relation between the atmosphere correction refractive index of the surface suspended waveguide and the sea surface vertical height based on a preset surface suspended waveguide mapping relation model according to the surface suspended waveguide parameters;
and determining the equipment furthest detection distance of the frequency equipment to be detected according to the equipment related information, the first mapping relation and the second mapping relation.
6. The method for atmospheric waveguide parameter inversion according to claim 5, wherein said determining the device furthest detection distance of the device to be tested according to the device-related information, the first mapping relationship and the second mapping relationship comprises:
determining a fusion correction parameter according to the first mapping relation and the second mapping relation;
Determining the propagation loss of the frequency equipment to be tested according to the fusion correction parameters;
And determining the equipment furthest detection distance of the frequency equipment to be detected according to the propagation loss and the equipment related information.
7. The method of atmospheric waveguide parameter inversion according to claim 6, wherein said determining the device furthest detection distance of the device to be tested according to the propagation loss and the device-related information comprises:
determining the received signal power of the frequency equipment to be tested according to the propagation loss and the equipment related information;
And if the received signal power is not smaller than a preset minimum received signal threshold value, determining the equipment furthest detection distance of the to-be-detected frequency equipment according to the equipment related information.
8. An atmospheric waveguide parameter inversion model training device, comprising:
the correlation quantity determining module is used for determining a first time delay correlation quantity of the sea surface scattering signal and determining a second time delay correlation quantity of the occultation direct signal;
A first training set determining module for generating a first sample training set with a first delay related quantity of a first sample tag; the first sample tag includes real evaporation waveguide parameters; and
A second training set determining module, configured to generate a second sample training set with a second delay related quantity of a second sample label; the second sample tag comprises real surface suspended waveguide parameters;
The evaporation parameter obtaining module is used for inputting the first sample training set into a preset multi-head attention mechanism model to carry out model training, obtaining predicted evaporation waveguide parameters output by the model, carrying out model training according to the predicted evaporation waveguide parameters and the real evaporation waveguide parameters until a preset first model training ending condition is met, obtaining an evaporation waveguide parameter prediction model, and inverting the evaporation waveguide parameters in the atmospheric waveguide parameters; and
The suspension parameter obtaining module is used for inputting the second sample training set into a preset multi-head attention mechanism model to perform model training to obtain predicted surface suspension waveguide parameters output by the model, performing model training according to the predicted surface suspension waveguide parameters and the real surface suspension waveguide parameters until a preset second model training ending condition is met to obtain a surface suspension waveguide parameter prediction model, and inverting the surface suspension waveguide parameters in the atmospheric waveguide parameters;
Wherein the correlation amount determination module includes:
The first information acquisition unit is used for acquiring navigation position information of a Global Navigation Satellite System (GNSS) satellite, receiving position information of a GNSS scattered signal receiving antenna and reflection position information of a sea surface specular reflection point under a geocentric rectangular coordinate system;
A first quantity determining unit configured to determine a first delay-related quantity of sea surface scattering signals according to the navigation position information, the reception position information, and the reflection position information;
wherein, the correlation amount determining module further comprises:
The signal power acquisition unit is used for acquiring the signal power of the GNSS occultation;
the additional phase determining unit is used for determining the additional phase of the GNSS occultation signal according to the clock error and the position distance between the GNSS occultation signal receiver and the signal transmitter;
And the second quantity determining unit is used for determining a second delay related quantity of the occultation direct signal according to the signal power and the signal additional phase.
9. An atmospheric waveguide parameter inversion apparatus, comprising:
The to-be-measured correlation quantity determining module is used for determining a first to-be-measured delay correlation quantity of the to-be-measured sea surface scattering signal; determining a second time delay related quantity to be detected of the occultation direct signal to be detected;
the evaporation parameter output module is used for inputting the first delay related quantity to be measured into an evaporation waveguide parameter prediction model to obtain evaporation waveguide parameters output by the model; and
The suspension parameter output module is used for inputting the second delay related quantity to be detected into the surface suspension waveguide parameter prediction model to obtain the surface suspension waveguide parameter output by the model;
The waveguide parameter generation module is used for generating atmospheric waveguide parameters comprising the evaporation waveguide parameters and the surface suspended waveguide parameters;
The prediction model generation module is used for generating the evaporation waveguide parameter prediction model and the surface suspension waveguide parameter prediction model by adopting the atmospheric waveguide parameter inversion model training device in claim 8.
10. An electronic device, comprising:
one or more processors;
A memory for storing one or more programs;
When the one or more programs are executed by the one or more processors, the one or more processors are caused to implement an atmospheric waveguide parameter inversion model training method or an atmospheric waveguide parameter inversion method as claimed in any one of claims 1-7.
11. A computer readable storage medium having stored thereon a computer program, which when executed by a processor implements an atmospheric waveguide parameter inversion model training method or an atmospheric waveguide parameter inversion method according to any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410070672.3A CN117592381B (en) | 2024-01-18 | 2024-01-18 | Atmospheric waveguide parameter inversion model training method, device, equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410070672.3A CN117592381B (en) | 2024-01-18 | 2024-01-18 | Atmospheric waveguide parameter inversion model training method, device, equipment and medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117592381A CN117592381A (en) | 2024-02-23 |
| CN117592381B true CN117592381B (en) | 2024-05-17 |
Family
ID=89922295
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410070672.3A Active CN117592381B (en) | 2024-01-18 | 2024-01-18 | Atmospheric waveguide parameter inversion model training method, device, equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117592381B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111819675A (en) * | 2018-03-20 | 2020-10-23 | 科磊股份有限公司 | Method and system for real-time measurement control |
| CN112017289A (en) * | 2020-08-31 | 2020-12-01 | 电子科技大学 | Well-seismic combined initial lithology model construction method based on deep learning |
| CN113588688A (en) * | 2021-06-18 | 2021-11-02 | 田斌 | Device and method for evaluating ocean atmospheric waveguide potential energy based on satellite system |
| CN114611415A (en) * | 2022-05-11 | 2022-06-10 | 中国海洋大学 | Beyond-visual-range propagation loss prediction method based on SL-TrellisNets network |
| CN114647812A (en) * | 2022-03-17 | 2022-06-21 | 中国科学院国家空间科学中心 | GNSS-R sea wind inversion method based on multi-dimensional feature excavation neural network |
| CN115062526A (en) * | 2022-02-22 | 2022-09-16 | 中国科学院自动化研究所 | Deep learning-based three-dimensional ionosphere electron concentration distribution model training method |
| CN116090352A (en) * | 2023-02-01 | 2023-05-09 | 北京华电力拓能源科技有限公司 | Full waveform inversion method based on gate cycle unit and attention mechanism |
| CN117240666A (en) * | 2023-08-21 | 2023-12-15 | 中国科学院声学研究所 | Neural network underwater acoustic channel estimation method based on cluster self-attention mechanism |
-
2024
- 2024-01-18 CN CN202410070672.3A patent/CN117592381B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111819675A (en) * | 2018-03-20 | 2020-10-23 | 科磊股份有限公司 | Method and system for real-time measurement control |
| CN112017289A (en) * | 2020-08-31 | 2020-12-01 | 电子科技大学 | Well-seismic combined initial lithology model construction method based on deep learning |
| CN113588688A (en) * | 2021-06-18 | 2021-11-02 | 田斌 | Device and method for evaluating ocean atmospheric waveguide potential energy based on satellite system |
| CN115062526A (en) * | 2022-02-22 | 2022-09-16 | 中国科学院自动化研究所 | Deep learning-based three-dimensional ionosphere electron concentration distribution model training method |
| CN114647812A (en) * | 2022-03-17 | 2022-06-21 | 中国科学院国家空间科学中心 | GNSS-R sea wind inversion method based on multi-dimensional feature excavation neural network |
| CN114611415A (en) * | 2022-05-11 | 2022-06-10 | 中国海洋大学 | Beyond-visual-range propagation loss prediction method based on SL-TrellisNets network |
| CN116090352A (en) * | 2023-02-01 | 2023-05-09 | 北京华电力拓能源科技有限公司 | Full waveform inversion method based on gate cycle unit and attention mechanism |
| CN117240666A (en) * | 2023-08-21 | 2023-12-15 | 中国科学院声学研究所 | Neural network underwater acoustic channel estimation method based on cluster self-attention mechanism |
Non-Patent Citations (2)
| Title |
|---|
| 基于优化滤波的水平非均匀蒸发波导反演算法;董道广;芮国胜;康健;田文飚;;火力与指挥控制;20160715(07);全文 * |
| 机载GNSS掩星观测反演大气折射率廓线研究;李潇;张训械;李烨;王烁;李兴国;吕铁军;;航天器工程;20110315(02);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117592381A (en) | 2024-02-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11035961B2 (en) | Systems and methods for real time kinematic satellite positioning | |
| CN109283562B (en) | Vehicle three-dimensional positioning method and device in Internet of vehicles | |
| CN103728643B (en) | With the Big Dipper three network RTK blur level single epoch fixing means frequently that wide lane retrains | |
| Chu et al. | Multimodal deep learning for heterogeneous GNSS-R data fusion and ocean wind speed retrieval | |
| CN113031031B (en) | GNSS signal accurate classification-based weighted positioning method in urban canyon | |
| US10324160B2 (en) | Geolocation of beyond LOS HF emitters | |
| JP6689396B2 (en) | Weather forecasting device, weather forecasting method, and program | |
| CN111221018A (en) | GNSS multi-source information fusion navigation method for inhibiting marine multipath | |
| US11573329B2 (en) | Modeling effects of structures on global-positioning system localization | |
| CN114900224B (en) | Method for aircraft handover target measurement and control satellite and related equipment | |
| CN109738902A (en) | A kind of autonomous acoustic navigation method of underwater high-speed target with high precision based on sync beacon mode | |
| Wang et al. | A new GNSS-R altimetry algorithm based on machine learning fusion model and feature optimization to improve the precision of sea surface height retrieval | |
| CN117592381B (en) | Atmospheric waveguide parameter inversion model training method, device, equipment and medium | |
| WO2023236643A1 (en) | Positioning method and apparatus, device and storage medium | |
| CN115616637B (en) | Urban complex environment navigation positioning method based on three-dimensional grid multipath modeling | |
| CN111708005A (en) | Underwater photon displacement correction and depth measurement method and device for single photon laser radar | |
| Ye et al. | Robust real‐time kinematic positioning method based on NLOS detection and multipath elimination in GNSS challenged environments | |
| CN114202104B (en) | Method for determining similarity degree of tropical cyclone path and storage medium | |
| Ma et al. | A Method for Establishing Tropospheric Atmospheric Refractivity Profile Model Based on Multiquadric RBF and k-means Clustering | |
| Graupe et al. | An automated framework for long-range acoustic positioning of autonomous underwater vehicles | |
| Ersan et al. | Map matching with kalman filter and location estimation | |
| Sun et al. | Incremental Learning for LOS/NLOS Classification of Global Navigation Satellite System | |
| CN116794646B (en) | Mixed high-frequency radar target tracking method based on variable decibel leaf reasoning | |
| CN113917509B (en) | Double-difference ambiguity fixing method, device and readable storage medium | |
| Mattes | Analysis of waveforms in the satellite altimetry by using neural networks |
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 |