A method and device for predicting physical information of atmospheric turbulence for satellite-ground link laser communication

Abstract

The application discloses a kind of atmospheric turbulence physical information prediction method and equipment for star-ground link laser communication, comprising: the atmospheric turbulence region of star-ground link is divided into several turbulence layers, and the phase screen and amplitude screen model of turbulence region is constructed;Atmospheric turbulence physical information is set to different values, based on phase screen model and amplitude screen model, using complex domain simulation method, corresponding atmospheric turbulence effect data is simulated, each group of atmospheric turbulence effect data is used as training sample, corresponding atmospheric turbulence physical information is used as label, to form a sample pair;PINN model is constructed, and a plurality of sample pairs are used for training, atmospheric turbulence effect data is collected in real time at the receiving end of star-ground laser communication system, input into the PINN model trained, and real-time atmospheric turbulence physical information is predicted.The application can accurately characterize the dynamic characteristics of atmospheric turbulence of star-ground link, realize precision simulation, and can predict turbulence parameters in real time.

Description

Technical Field

[0001] This invention relates to satellite-to-ground laser communication technology, and more particularly to a method and device for predicting atmospheric turbulence physical information for satellite-to-ground link laser communication. Background Technology

[0002] Space-to-ground laser communication, as an efficient information transmission method between satellites and ground optical stations, has advantages such as high transmission rate, small divergence angle, and good security, and can meet the needs of secure transmission of massive amounts of data. However, when lasers propagate in the atmosphere, they are significantly affected by atmospheric turbulence, which causes turbulence effects such as phase fluctuations, light intensity flicker, beam drift, and beam spread. These turbulence effects seriously affect the acquisition, alignment, and tracking performance of the communication system, and may even lead to data loss.

[0003] To mitigate the effects of turbulence, adaptive optics (AO) systems are widely used for real-time compensation of wavefront distortion. However, under strong turbulence conditions, AO systems alone are insufficient to achieve optimal correction, especially in the absence of prior assessment or prediction of turbulence intensity, resulting in low system resource utilization efficiency. Current methods for simulating and predicting atmospheric turbulence have the following limitations:

[0004] First, existing phase screen simulations only consider wavefront phase distortion and are therefore inaccurate.

[0005] Second, existing numerical simulations are mostly focused on near-ground horizontal links, with a lack of experimental data on satellite-to-ground links. Furthermore, the simulations only consider phase distortion and ignore amplitude distortion, making it difficult to fully characterize the dynamic evolution of turbulence and resulting in inaccurate simulation accuracy.

[0006] Third, existing technologies are mostly focused on forward numerical simulation of atmospheric turbulence. However, in actual satellite-to-ground laser communication engineering applications, the atmospheric environment of satellite-to-ground links is complex, and the physical parameters of atmospheric turbulence are often unknown and difficult to measure in real time using sounding rockets or weather balloons. On-site measurement of turbulence parameters requires a lot of manpower and material resources, which limits feasibility and economy.

[0007] Therefore, there is an urgent need for a method that can accurately characterize the dynamic characteristics of atmospheric turbulence in satellite-to-ground links, achieve accurate simulation, and predict turbulence parameters in real time, so as to provide reliable support for adaptive compensation of satellite-to-ground laser communication systems. Summary of the Invention

[0008] To address the problems existing in the prior art, the purpose of this invention is to provide a more accurate method and device for predicting atmospheric turbulence physical information for satellite-to-ground link laser communication.

[0009] To achieve the above-mentioned objectives, the present invention provides the following technical solution.

[0010] A method for predicting atmospheric turbulence physical information for satellite-to-ground link laser communication includes the following steps:

[0011] (1) Divide the atmospheric turbulence region of the satellite-to-ground link into several turbulence layers;

[0012] (2) Combining the turbulence spectrum index and generalized refractive index structure parameters of each turbulence layer, different subharmonic orders and turbulence scale dynamic correction factors are set for different turbulence layers. The spatially weighted complex Gaussian random number matrix is ​​subjected to layered power spectrum filtering, inverse Fourier transform and layered multi-scale harmonic compensation to obtain an anisotropic phase screen model of each turbulence layer that satisfies the Nyquist criterion; wherein, the turbulence scale dynamic correction factor is an exponential function of the ratio of the turbulence inner and outer scales of the turbulence layer;

[0013] (3) Construct an amplitude screen model using the log-normal probability density function distribution and the Gamma-Gamma probability density function distribution as distribution functions;

[0014] (4) Set different values ​​for atmospheric turbulence physical information, and use the complex domain simulation method to simulate and generate phase screens and amplitude screens of each turbulence layer when the atmospheric turbulence physical information is different, based on the phase screen model and amplitude screen model, and obtain the corresponding atmospheric turbulence effect data. Use each set of atmospheric turbulence effect data as training samples and the corresponding atmospheric turbulence physical information as labels to form a sample pair; where atmospheric turbulence physical information includes generalized refractive index structure parameters, power spectrum power law, turbulence inner scale, and turbulence outer scale, and atmospheric turbulence effect data includes beam spread, beam drift, optical field amplitude distortion and phase distortion data;

[0015] (5) Construct a PINN model and train it using several sample pairs. The PINN model includes an input layer, several hidden layers and an output layer. The input layer inputs atmospheric turbulence effect data and the output layer outputs real-time atmospheric turbulence physical information.

[0016] (6) Input the atmospheric turbulence effect data collected in real time by the receiver of the satellite-to-ground laser communication system into the trained PINN model to predict the real-time atmospheric turbulence physical information.

[0017] Furthermore, the anisotropic phase screen model is specifically as follows:

[0018]

[0019]

[0020] In the formula, Let be the phase screen at the coordinate position (x, y) of the d-th turbulent layer. Let i be a complex Gaussian random number matrix, where i is the imaginary unit. , These are the space wavenumbers along the x and y directions, respectively. For wavenumber vectors, Let be the normalization constant of the d-th turbulent layer. The d-th turbulent layer is based on generalized refractive index structural parameters. Spatial weighting factor, This is the dynamic correction factor for the turbulence scale of the d-th turbulent layer. denoted as the Non-Kolmogorov power spectrum of the d-th turbulent layer, r is the position vector in the plane, and N represents the number of turbulent layers.

[0021] Furthermore, the turbulence-scale dynamic correction factor is specifically as follows:

[0022]

[0023] in, , These represent the inner and outer scales of turbulence in the d-th turbulent layer, respectively. is the turbulence spectrum index of the d-th turbulent layer.

[0024] Furthermore, the Non-Kolmogorov spectral power spectrum is specifically as follows:

[0025]

[0026]

[0027] In the formula, This represents the anisotropic atmospheric turbulence parameters of the d-th turbulent layer along the x and y directions. The power spectrum of the d-th layer is power-law. , Power spectrum power law The function, These are the space wavenumbers along the z-direction, For the d-th turbulent layer, the external scale correlation wavenumber of the turbulence, denoted as the intra-scale correlation wavenumber of the d-th turbulent layer.

[0028] Furthermore, the method for generating the amplitude screen model is as follows:

[0029] The probability density function distribution in the following formula generates the amplitude screen:

[0030]

[0031]

[0032]

[0033] In the formula, The probability density function representing the amplitude screen. Indicates light intensity. Indicates the flicker factor. This represents the average light intensity received at the point. K represents the normalized light intensity. * () denotes a modified Bessel function of the second kind of order *. This represents the Rytov variance. and It is about The function, Represents the atmospheric structure constant. The threshold value represents the atmospheric structure constant, and L represents the link distance for laser transmission. =2π / λ is the light wave number, and λ is the light wavelength.

[0034] Furthermore, the complex domain simulation method specifically involves: using a multi-layer phase screen step-by-step transmission method, each region corresponds to a phase screen and an amplitude screen, the complex amplitude of the incident Gaussian beam is Fourier transformed, multiplied by the angular spectral transfer function to achieve vacuum transmission, and then superimposed with the phase screen and amplitude screen, and the transmitted light field distribution is obtained by inverse Fourier transform, and then the transmitted atmospheric turbulence effect data is generated through simulation.

[0035] Furthermore, the PINN model is trained using the following loss function:

[0036]

[0037]

[0038]

[0039] In the formula, Indicates the total loss. Indicates data loss. Represents physical constraint loss. , This represents the weighting coefficient, and N represents the number of sample pairs. This represents the label of the j-th training sample. This represents the output of the PINN model when the j-th training sample is input; Represents the gradient operator. Wavelength, For laser electric field, The mean is zero, representing random fluctuations in the refractive index caused by atmospheric turbulence for laser transmission.

[0040] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method described above.

[0041] A computer-readable storage medium having a computer program / instructions stored thereon, which, when executed by a processor, implements the above-described method.

[0042] A computer program product includes a computer program / instructions that, when executed by a processor, implement the above-described method.

[0043] Compared with the prior art, the beneficial effects of this invention are:

[0044] 1. This invention employs a power spectrum inversion method with hierarchical multi-scale differentiated harmonic compensation to construct an anisotropic phase screen model for a three-layer turbulent region. This model simulates a more accurate phase screen.

[0045] 2. By using the complex domain simulation method to simultaneously generate amplitude and phase screens, high-precision simulation of amplitude and phase distortion is achieved, which makes up for the deficiency of existing simulations that ignore amplitude distortion and provides comprehensive data support for the analysis of laser transmission effects.

[0046] 3. The training of the PINN model combines data-driven and physical constraints. It can achieve real-time prediction of turbulence parameters with only a small amount of training data. It has strong generalization ability and high prediction accuracy, which solves the problems of high cost of field measurement and lack of physical meaning in traditional data-driven models.

[0047] 4. The prediction results of this invention can directly provide real-time turbulence parameter references for satellite-to-ground laser communication adaptive optics systems, effectively improving the acquisition, alignment and tracking performance of communication systems in complex turbulent environments, reducing the risk of data loss, and have important engineering application value. Attached Figure Description

[0048] Figure 1 This is a flowchart illustrating the atmospheric turbulence physical information prediction method for satellite-to-ground link laser communication provided in an embodiment of the present invention.

[0049] Figure 2 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0050] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0051] Example 1

[0052] This invention provides a method for predicting atmospheric turbulence physical information for satellite-to-ground laser communication, comprising the following steps:

[0053] S101. Divide the atmospheric turbulence region of the satellite-to-ground link into several turbulence layers.

[0054] Specifically, based on the altitude distribution characteristics of atmospheric turbulence in the satellite-to-ground link, the signal transmission path is divided into three turbulent layers: the boundary layer (0~2km), the troposphere (2~10km), and the stratosphere (10~20km).

[0055] S102. Combining the turbulence spectral index and generalized refractive index structure parameters of each turbulent layer, differentiated subharmonic orders and turbulence scale dynamic correction factors are set for different turbulent layers. The spatially weighted complex Gaussian random number matrix is ​​subjected to layered power spectrum filtering, inverse Fourier transform and layered multi-scale harmonic compensation to obtain an anisotropic phase screen model of each turbulent layer that satisfies the Nyquist criterion.

[0056] The anisotropic phase screen model is specifically as follows:

[0057]

[0058]

[0059] In the formula, Let be the phase screen at the coordinate position (x, y) of the d-th turbulent layer. Let i be a complex Gaussian random number matrix, where i is the imaginary unit. , These are the space wavenumbers along the x and y directions, respectively. For wavenumber vectors, Let be the normalization constant of the d-th turbulent layer. The d-th turbulent layer is based on generalized refractive index structural parameters. Spatial weighting factor, This is the dynamic correction factor for the turbulence scale of the d-th turbulent layer. The power spectrum of the d-th turbulent layer is the Non-Kolmogorov spectrum, r is the position vector in the plane, and N represents the number of turbulent layers. In this embodiment, N=3. d=1,2,3 represent the boundary layer, troposphere, and stratosphere, respectively.

[0060]

[0061] in, , These represent the inner and outer scales of turbulence in the d-th turbulent layer, respectively. is the turbulence spectrum index of the d-th turbulent layer.

[0062] Each turbulent layer uses an independent Non-Kolmogorov power spectrum model.

[0063]

[0064]

[0065]

[0066]

[0067] In the formula, This represents the anisotropic atmospheric turbulence parameters of the d-th turbulent layer along the x and y directions. The power spectrum of the d-th layer is power-law. , Power spectrum power law The function, For the d-th turbulent layer, the external scale correlation wavenumber of the turbulence, Let d be the intra-scale correlation wavenumber of the turbulent layer. (·) is the gamma function. At that time, the power spectrum changes from anisotropic to isotropic.

[0068] The generalized refractive index structure parameters at various altitudes were calculated using the Hufnagel-Valley 5 / 7 model, specifically as follows:

[0069]

[0070]

[0071] in Here, h represents the generalized refractive index structure parameter, v represents the root mean square wind speed (typically 21 m / s), and A represents the near-surface refractive index structure constant (typically 1.7 × 10⁻¹). 4 m⁻² / ³).

[0072] S103. Construct an amplitude screen model using the log-normal probability density function distribution and the Gamma-Gamma probability density function distribution as distribution functions.

[0073] The amplitude screen is generated using the probability density function distribution shown in the following formula, thereby simulating the signal amplitude distortion during atmospheric transmission of the beam, specifically:

[0074]

[0075]

[0076]

[0077] In the formula, Let represent the probability density function of the amplitude screen. The above equation represents the log-normal probability density function distribution, and the below equation represents the Gamma-Gamma probability density function distribution. Indicates light intensity. Indicates the flicker factor. This represents the average light intensity received at the point. K represents the normalized light intensity. * () denotes a modified Bessel function of the second kind of order *. This represents the Rytov variance. and It is about The function, Represents the atmospheric structure constant. This represents the threshold value for the atmospheric structure constant, specifically 1x10. -16 m -2 / 3 L is the link distance for laser transmission. =2π / λ is the light wave number, and λ is the light wavelength.

[0078] S104. Set different values ​​for atmospheric turbulence physical information, and use the complex domain simulation method to simulate and generate phase screens and amplitude screens for each turbulent region pair when the atmospheric turbulence physical information has different values, based on the phase screen model and amplitude screen model, and obtain the corresponding atmospheric turbulence effect data. Use each set of atmospheric turbulence effect data as training samples and the corresponding atmospheric turbulence physical information as labels to form a sample pair.

[0079] Among them, atmospheric turbulence physical information includes generalized refractive index structure parameters, power law of power spectrum, turbulence internal scale, turbulence external scale, and atmospheric turbulence effect data including beam spread, beam drift, optical field amplitude distortion, and phase distortion data.

[0080] The complex domain simulation method is as follows: by using a multi-layer phase screen step-by-step transmission method, each region corresponds to a phase screen and an amplitude screen. The complex amplitude of the incident Gaussian beam is Fourier transformed and multiplied by the angular spectrum transfer function to achieve vacuum transmission. Then, it is superimposed with the phase screen and the amplitude screen, and the transmitted light field distribution is obtained by inverse Fourier transform. Then, the atmospheric turbulence effect data after transmission is generated through simulation.

[0081] Specifically, in this embodiment, the simulation parameters are set as follows: laser wavelength λ = 0.5 μm, initial beam radius W0 = 0.05 m, zenith angle θ = 0°; the turbulence model adopts the Hufnagel-Valley 5 / 7 model, and the near-surface refractive index structure constant A = 1.7 × 10⁻¹ 4m⁻² / ³, root mean square wind speed v=21m / s; turbulent inner scale l0=0.001m, outer scale L0=50m; number of sampling points N=512, grid spacing at the transmitter δ1=0.0035m, grid spacing at the receiver δ0=0.005m, telescope aperture at the observation surface 0.5m; total transmission height 20km, divided into three segments: boundary layer (0~2km), troposphere (2~10km), and stratosphere (10~20km), with one phase screen set in each segment, for a total of 3 phase screens.

[0082] In terms of hardware deployment, considering the long satellite link distance and complex turbulent environment, the system adopts a GPU / CPU heterogeneous computing architecture to balance computing power and energy consumption. The software environment uses the Windows 11 operating system, which offers good compatibility with the heterogeneous computing architecture and long-term operational stability, effectively reducing the risk of system crashes. The core software uses Python 3.11 and PyTorch 2.2.0, providing efficient GPU acceleration support for deep learning model training and inference. Combined with the CUDA 12.3 acceleration library, it can fully unleash the hardware's computing power. Regarding the setting of scene parameters, the root mean square wind speed references the average wind speed characteristics of the upper atmosphere in mid-latitude regions, ensuring the realism of the parameter settings. The selection of the inner and outer scales of turbulence is based on the vortex characteristics of turbulence at different altitudes along the satellite-to-ground link. The boundary layer turbulent vortices are smaller in scale and densely distributed, while those in the stratosphere are relatively sparse. Choosing appropriate scale parameters can improve the physical realism of the phase screen and amplitude screen simulations. The configuration of the number of sampling points and grid spacing strikes a balance between simulation accuracy and computational efficiency. 1024 sampling points ensure the capture of small-scale turbulent vortices without increasing computational latency due to excessive data volume. The division of the three-segment turbulence region fully follows the altitude distribution pattern of atmospheric turbulence in the satellite-to-ground link. Each segment is equipped with an independent phase screen and amplitude screen, which can accurately characterize the unique characteristics of turbulence at different altitudes.

[0083] S105. Construct the PINN model and train it using several sample pairs.

[0084] The PINN model includes an input layer (4 neurons, namely beam spread, beam drift, optical field amplitude distortion and phase distortion data), several hidden layers (specifically 6 hidden layers, each with 64 neurons, and the activation function is ReLU) and an output layer (4 neurons, specifically generalized refractive index structure parameters, power spectrum power law, turbulence inner scale, and turbulence outer scale).

[0085] The training data consists of 1,000 sets of atmospheric turbulence effect data and corresponding atmospheric turbulence effect data transmitted via simulated satellite-to-ground link laser transmission, of which 800 sets are the training set and 200 sets are the validation set.

[0086] The following loss function is used for training:

[0087]

[0088]

[0089]

[0090] In the formula, Indicates the total loss. This represents data loss, which is the mean squared error between the predicted and actual values. The physical constraint loss is represented by the stochastic scalar Helmholtz equation, which describes the relationship between laser characteristic parameters and atmospheric turbulence. , This represents the weighting coefficient, and N represents the number of sample pairs. This represents the label of the j-th training sample. This represents the output of the PINN model when the j-th training sample is input; Represents the gradient operator. Wavelength, For laser electric field, The mean is zero, representing random fluctuations in the refractive index caused by atmospheric turbulence for laser transmission.

[0091] Automatic differentiation techniques were employed to calculate the spatial and temporal derivatives of the physical equations. Simulated atmospheric turbulence effect data were used as input, atmospheric turbulence physics information was used as output, and atmospheric turbulence control equations were embedded as soft constraints to train a turbulence parameter prediction model—the PINN model. Training and optimization were performed using the Adam optimizer with a learning rate of 0.001 and 1000 iterations. Model validation resulted in a phase structure function fitting error ≤3%, a scintillation index prediction error ≤5%, and a turbulence parameter prediction error ≤8% on the validation set, meeting engineering application requirements.

[0092] Furthermore, while the pre-trained model possesses a certain generalization ability, the turbulence characteristics of different satellite-to-ground links vary. Online fine-tuning allows the model to quickly adapt to the turbulence patterns of the current link, reducing prediction errors caused by scene migration. The weights of data loss and physical constraint loss in the loss function are set at 0.7:0.3. Data loss ensures the consistency between the model's predictions and the actual turbulence parameters, while physical constraint loss guarantees that the predictions conform to the basic physical laws of atmospheric turbulence, avoiding outliers that contradict physical principles.

[0093] S106. The receiving end of the satellite-to-ground laser communication system collects atmospheric turbulence effect data in real time, inputs it into the trained PINN model, and predicts the real-time atmospheric turbulence physical information.

[0094] Before making predictions, parameters such as phase amplitude variation, phase structure function, light intensity scintillation index, and beam spread should be theoretically verified.

[0095] After receiving preprocessed observation data (atmospheric turbulence effect data), the PINN model rapidly completes prediction calculations through GPU parallel computing, with a single prediction time controlled within 100ms, meeting the real-time compensation requirements of satellite-to-ground laser communication. The prediction frequency is synchronized with the sensor sampling frequency to ensure that each set of observation data is processed in a timely manner. The core advantage of GPU parallel computing lies in its ability to process multiple sets of observation data simultaneously, fully utilizing hardware computing power and avoiding the accumulation of latency caused by serial processing of single data. During the prediction process, the PINN model automatically reads the fine-tuned model parameters and, combined with the real-time input observation features, quickly outputs the core turbulence parameters.

[0096] The predicted real-time atmospheric turbulence physics information can be used to adjust the compensation strategy of the adaptive optics system, achieving accurate compensation for wavefront distortion. Specifically, adaptive optics compensation is performed through differentiated correction strategies. After receiving the turbulence parameters output by the prediction system, the adaptive optics compensation unit immediately activates the dynamic correction strategy to effectively correct for effects such as phase fluctuations and beam drift caused by turbulence. The core logic of the compensation strategy is based on the predicted turbulence parameters, such as adjusting the voltage of each driving unit. For boundary layer regions with high turbulence intensity, the correction weight of the corresponding driving units is appropriately increased to prioritize compensation for the main distortion components affecting communication quality. For stratospheric regions with weaker turbulence intensity, a relatively mild correction strategy is adopted to avoid energy loss due to over-correction. Through differentiated correction strategies, the service life of deformable mirrors can be improved while ensuring compensation effectiveness, meeting the economic requirements of engineering applications.

[0097] Example 2

[0098] Figure 2 This is a schematic diagram of the structure of a computer device provided by an embodiment of the present invention. The embodiment of the present invention provides services for the implementation of the method of the first embodiment of the present invention. Figure 2 As shown, the device may include: a memory 301 storing a computer-executable program; a processor 302 coupled to the memory 301; the processor 302 calls the computer-executable program stored in the memory 301 to perform the steps in the method described in Embodiment 1.

[0099] Memory 301 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) and / or cache memory. The device may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, memory 301 may be used to read and write non-removable, non-volatile magnetic media (commonly referred to as a "hard disk drive"). A program / utility having a set (at least one) of program modules may be stored, for example, in memory 301. Such program modules include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. The computer-executable program of the program modules typically performs the functions and / or methods described in the embodiments of the present invention.

[0100] The processor 302 executes various functional applications and data processing by running programs stored in the memory 301, such as implementing the method provided in Embodiment 1 of the present invention.

[0101] The code of a computer executable program can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages.

[0102] Example 3

[0103] This invention provides a storage medium containing a computer-executable program, which, when executed by a computer processor, is used to perform the method of Embodiment 1.

[0104] The storage medium of this invention can be any combination of one or more computer-readable media. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0105] Of course, the computer-executable program in the storage medium provided in the embodiments of the present invention is not limited to the above-described method operations, but can also perform related operations in the methods provided in any embodiment of the present invention.

[0106] Example 4

[0107] This invention also provides a computer program product, such as an app on a mobile phone or tablet, or an installer on a computer. This product includes a computer program / instructions that, when executed by a processor, implement the method described in Embodiment 1. The code for the computer-executable program used to perform the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as "C" or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0108] It should be noted that in this paper, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0109] It should be understood that the embodiments and descriptions above are only the principles, main features and advantages of the present invention. Various changes and modifications can be made to the present invention without departing from the spirit and scope of the invention, and all such changes and modifications fall within the protection scope of the present invention.

Claims

1. A method for predicting atmospheric turbulence physical information for satellite-to-ground link laser communication, characterized in that, Includes the following steps: (1) Divide the atmospheric turbulence region of the satellite-to-ground link into several turbulence layers; (2) Combining the turbulence spectrum index and generalized refractive index structure parameters of each turbulence layer, different subharmonic orders and turbulence scale dynamic correction factors are set for different turbulence layers. The spatially weighted complex Gaussian random number matrix is ​​subjected to layered power spectrum filtering, inverse Fourier transform and layered multi-scale harmonic compensation to obtain an anisotropic phase screen model of each turbulence layer that satisfies the Nyquist criterion; wherein, the turbulence scale dynamic correction factor is an exponential function of the ratio of the turbulence inner and outer scales of the turbulence layer; (3) Construct an amplitude screen model using the log-normal probability density function distribution and the Gamma-Gamma probability density function distribution as distribution functions; (4) Set different values ​​for atmospheric turbulence physical information, and use the complex domain simulation method to simulate and generate phase screens and amplitude screens of each turbulence layer when the atmospheric turbulence physical information is different, based on the phase screen model and amplitude screen model, and obtain the corresponding atmospheric turbulence effect data. Use each set of atmospheric turbulence effect data as training samples and the corresponding atmospheric turbulence physical information as labels to form a sample pair; where atmospheric turbulence physical information includes generalized refractive index structure parameters, power spectrum power law, turbulence inner scale, and turbulence outer scale, and atmospheric turbulence effect data includes beam spread, beam drift, optical field amplitude distortion and phase distortion data; (5) Construct a PINN model and train it using several sample pairs. The PINN model includes an input layer, several hidden layers and an output layer. The input layer inputs atmospheric turbulence effect data and the output layer outputs real-time atmospheric turbulence physical information. (6) Input the atmospheric turbulence effect data collected in real time by the receiver of the satellite-to-ground laser communication system into the trained PINN model to predict the real-time atmospheric turbulence physical information; The amplitude screen model is generated by distributing the probability density function in the following formula: , , , In the formula, The probability density function representing the amplitude screen. Indicates light intensity. Indicates the flicker factor. This represents the average light intensity received at the point. K represents the normalized light intensity. * () denotes a modified Bessel function of the second kind of order *. This represents the Rytov variance. and It is about The function, Represents the atmospheric structure constant. The threshold value represents the atmospheric structure constant, and L represents the link distance for laser transmission. =2π / λ is the wavenumber, where λ is the wavelength of the light wave. (·) is the gamma function. For generalized refractive index structure parameters; The complex domain simulation method is as follows: by using a multi-layer phase screen step-by-step transmission method, each region corresponds to a phase screen and an amplitude screen. The complex amplitude of the incident Gaussian beam is Fourier transformed and multiplied by the angular spectrum transfer function to achieve vacuum transmission. Then, it is superimposed with the phase screen and the amplitude screen, and the transmitted light field distribution is obtained by inverse Fourier transform. Then, the atmospheric turbulence effect data after transmission is generated through simulation.

2. The method for predicting atmospheric turbulence physical information for satellite-to-ground link laser communication according to claim 1, characterized in that, The anisotropic phase screen model is specifically as follows: , , In the formula, Let be the phase screen at the coordinate position (x, y) of the d-th turbulent layer. Let i be a complex Gaussian random number matrix, where i is the imaginary unit. , These are the space wavenumbers along the x and y directions, respectively. For wavenumber vectors, Let be the normalization constant of the d-th turbulent layer. The d-th turbulent layer is based on generalized refractive index structural parameters. Spatial weighting factor, This is the dynamic correction factor for the turbulence scale of the d-th turbulent layer. denoted as the Non-Kolmogorov power spectrum of the d-th turbulent layer, r is the position vector in the plane, and N represents the number of turbulent layers.

3. The method for predicting atmospheric turbulence physical information for satellite-to-ground link laser communication according to claim 2, characterized in that, The turbulence-scale dynamic correction factor is specifically: , in, , These represent the inner and outer scales of turbulence in the d-th turbulent layer, respectively. is the turbulence spectrum index of the d-th turbulent layer.

4. The method for predicting atmospheric turbulence physical information for satellite-to-ground link laser communication according to claim 3, characterized in that, The specific power spectrum of the Non-Kolmogorov spectrum is as follows: , , In the formula, This represents the anisotropic atmospheric turbulence parameters of the d-th turbulent layer along the x and y directions. The power spectrum of the d-th layer is power-law. , Power spectrum power law The function, For the d-th turbulent layer, the external scale correlation wavenumber of the turbulence, denoted as the intra-scale correlation wavenumber of the d-th turbulent layer.

5. The method for predicting atmospheric turbulence physical information for satellite-to-ground link laser communication according to claim 1, characterized in that, The PINN model is trained using the following loss function: , , , In the formula, Indicates the total loss. Indicates data loss. Represents physical constraint loss. , This represents the weighting coefficient, and N represents the number of sample pairs. This represents the label of the j-th training sample. This represents the output of the PINN model when the j-th training sample is input; Represents the gradient operator. Wavelength, For laser electric field, The mean is zero, representing random fluctuations in the refractive index caused by atmospheric turbulence for laser transmission.

6. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: The processor executes the computer program to implement the method as described in any one of claims 1-5.

7. A computer-readable storage medium having a computer program / instructions stored thereon, characterized in that, The computer program / instructions, when executed by a processor, implement the method of any one of claims 1-5.

8. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, they implement the method of any one of claims 1-5.

Images