Turbulence-resistant space-to-ground laser communication method and system based on wavelength-polarization reconfigurable synthetic aperture reception

Abstract

The application relates to the technical field of optical wireless communication, and discloses a wavelength-polarization reconfigurable synthetic aperture receiving-based anti-turbulence space-ground laser communication method and system. The method utilizes wavelength incoherence and polarization orthogonality to construct incoherent multi-beam transmitting signals at a transmitting end, uniformly takes receiving branch signals corresponding to different wavelengths, different polarizations, different apertures and different paths as candidate synthetic aperture units at a receiving end, constructs a candidate beam combination state set, and performs incoherent combination after an optimal beam combination is identified by a global fusion enhanced multi-branch network. The method and system can suppress light intensity flicker and power fluctuation caused by atmospheric turbulence without relying on active phase synchronization, reduce communication interruption probability, and improve the reliability of a space-downlink laser communication link.

Description

Technical Field

[0001] This application relates to the field of optical wireless communication technology, and in particular to an anti-turbulent star-to-ground laser communication technology based on wavelength-polarization reconfigurable synthetic aperture receiver. Background Technology

[0002] With the large-scale deployment of low-Earth orbit (LEO) satellite constellations, the continuous construction of integrated space-air-ground networks, and the rapid growth in demand for high-speed space information transmission, satellite-to-ground laser communication has become an important technical route for high-throughput satellite-to-ground downlink data transmission due to its advantages such as large communication capacity, abundant spectrum resources, strong resistance to electromagnetic interference, and good beam directionality. Compared with traditional radio frequency (RF) communication, satellite-to-ground laser communication can achieve high-speed data transmission under narrower beam conditions. Typical application scenarios include: high-resolution remote sensing satellites transmitting massive amounts of observation data back to ground gateway stations in real time via downlink laser links; LEO broadband satellite internet constellations transmitting large-capacity data to multiple ground gateway stations via laser-fed links; deep space probes transmitting scientific exploration data back to Earth-side ground terminals via optical communication links; and multi-link fusion or multi-path parallel downlink scenarios formed when multiple on-orbit satellites and multiple ground stations are networked collaboratively.

[0003] However, in the aforementioned typical application scenarios, satellite-to-ground laser communication links are susceptible to atmospheric turbulence, cloud cover, aerosol scattering, and pointing errors when traversing the atmosphere. Atmospheric turbulence causes beam drift, beam spread, intensity flicker, and wavefront distortion, resulting in random fluctuations in the optical power at the receiver. This leads to increased bit error rate, decreased link stability, and in severe cases, communication outages. For long-distance satellite-to-ground downlinks, satellite motion, dynamic changes in propagation distance, and time-varying atmospheric channel characteristics further exacerbate the difficulty of reliable link establishment. Especially in scenarios with multiple links coexisting, such as multiple satellites to the same ground station or multiple satellites to multiple ground stations, the equivalent transmission distance of different links and the atmospheric turbulence conditions they pass through vary significantly, and the degree of signal damage to each receiving branch often exhibits a marked imbalance.

[0004] Existing anti-turbulence methods mainly include structured beam emission, adaptive optics compensation, wavefront mode compensation, spatial diversity reception, multi-aperture reception, and coherent multibeamforming. However, existing technologies still have several technical problems in the specific application scenarios mentioned above: structured beam emission is difficult to maintain the optical field under long-distance transmission conditions, and the system implementation is complex; adaptive optics and wavefront mode compensation are limited by system size, cost, real-time control capability, and the number of controllable or processable modes; although spatial diversity and multi-aperture reception can reduce the influence of light intensity scintillation by utilizing the aperture averaging effect, they mostly adopt fixed combination methods, making it difficult to flexibly switch between different receiving channels according to the real-time turbulence state, and their merging gain is limited when the correlation of multiple signals is strong or some receiving branches are severely damaged; optical phased arrays and coherent multibeamforming require high phase synchronization accuracy of multiple optical fields, and the time-varying phase distortion introduced by atmospheric turbulence and path length changes in long-distance satellite-to-ground links makes it difficult to stably achieve active phase synchronization and real-time coherent synthesis. Therefore, there is an urgent need for a new technical solution that can reconfigure and incoherently combine multiple optical signals without relying on active phase synchronization, by utilizing wavelength incoherence, polarization orthogonality, and multi-aperture receiving information. This would reduce light intensity flicker and communication interruption probability caused by atmospheric turbulence, and improve the reliability of satellite-to-ground downlink laser communication links. Summary of the Invention

[0005] The purpose of this application is to provide an anti-turbulence star-to-ground laser communication method and system based on wavelength-polarization reconfigurable synthetic aperture receiver, so as to solve the problems mentioned in the background art.

[0006] This application provides an anti-turbulence star-to-ground laser communication method based on wavelength-polarization reconfigurable synthetic aperture receiver, comprising the following steps: S1: The satellite communication terminal modulates the downlink data signal onto at least two optical carriers of different wavelengths and performs polarization orthogonal modulation on the optical carriers to form an incoherent multi-beam transmission signal with wavelength and polarization dimensions; S2: The non-coherent multi-beam transmission signal in S1 is directed by a beam pointing device, so that the non-coherent multi-beam transmission signal is transmitted downlink along the satellite-to-ground laser communication link; S3: The ground station receiver receives the incoherent multi-beam transmitted signal described in S2 after being transmitted through atmospheric turbulence via a multi-aperture receiver. It then performs polarization separation and wavelength separation on the optical signal received by each receiving aperture in sequence to obtain N receiving branch signals. Each receiving branch signal is characterized by wavelength identifier, polarization identifier, aperture identifier, and path identifier and serves as a candidate synthetic aperture unit. S4: Based on the N candidate synthetic aperture units obtained in S3, the data processing center constructs a candidate beam combination state set. This set includes all combinations from the N candidate synthetic aperture units where m beams are selected for incoherent merging. The total number of candidate beam combination states is: in This represents the total number of candidate synthetic aperture elements. This indicates the number of candidate synthetic aperture elements participating in incoherent merging. Indicates from Choose the path The number of road combinations This represents the total number of candidate beam combination states; S5: Input the N received branch signals obtained in S3 into the global fusion enhanced multi-branch network. The global fusion enhanced multi-branch network includes N+1 branches, wherein the first N branches receive the corresponding received branch signals and extract the independent features of each path through the sub-network, and the N+1th branch receives the weighted fusion data of the N received branch signals. The joint features between the N received branch signals are extracted using a two-dimensional convolutional kernel. The independent features and the joint features are then concatenated and passed through a hidden layer and an output layer. Finally, the optimal beam combination under the current turbulence conditions is identified and output from the candidate beam combination state set described in S4. Indicates the first The path receives branch signals. Indicates the first The weighting coefficients corresponding to the branch signals received by the path. ; S6: Perform incoherent combining on the received branch signals corresponding to the optimal beam combination output by S5 to recover the downlink data signal.

[0007] In a preferred embodiment, the satellite communication terminal in S1 separates at least two incoherent wavelengths using a wavelength demultiplexer, and then achieves orthogonal polarization output of X-polarization and Y-polarization for each optical carrier using a polarization controller; the beam pointing device in S2 includes a voice coil motor-driven fast steering mirror array, a micromirror array antenna, or an optical scanner, and the beam pointing device achieves multi-directional common pointing of the incoherent multi-beam transmitted signal and completes far-field alignment.

[0008] In a preferred embodiment, the multi-aperture receiving device in S3 includes a dual-aperture receiving antenna, a multi-aperture receiving array, or a distributed receiving array composed of multiple ground stations. The ground station receiving end uses a half-wave plate, a polarization beam splitter, or a polarization separator to achieve spatial orthogonal polarization separation of the optical signal received by each receiving aperture, and uses a wavelength selection switch, an optical filter, or a wavelength demultiplexer to separate the polarization-separated optical signal into different wavelength channels, thereby obtaining the N receiving branch signals. The optical signal corresponding to each receiving aperture carries different aperture identifiers and path identifiers, which serve as the basis for constructing the candidate synthetic aperture unit in S4.

[0009] In a preferred embodiment, before inputting the N received branch signals into the global fusion enhanced multi-branch network in S5, each received branch signal is first subjected to IQ-path orthogonalization, dispersion compensation, constant modulus algorithm, frequency offset recovery, and phase offset recovery digital signal processing in sequence to obtain constellation point data as the nth branch signal in the global fusion enhanced multi-branch network. Road input data ,in .

[0010] In a preferred embodiment, the first N branches of the globally fused enhanced multi-branch network in S5 each include a sliding window module and a cascaded sub-network. The window length of the sliding window module is M, and the sub-network is used to extract the independent features from the corresponding received branch signals within the sliding window range. The (N+1)th branch includes a cascaded two-dimensional convolution module and a flattening module. The kernel size of the two-dimensional convolution module is... The flattening module is used to convert the output of the two-dimensional convolution module into a flattening module. The dimensions are flattened to M; the independent features and the joint features are concatenated via a Concat layer and then input into the hidden layer and the output layer.

[0011] In a preferred embodiment, when the globally fused enhanced multi-branch network in S5 identifies and outputs the optimal beam combination from the set of candidate beam combination states, the network determines the optimal beam combination based on the channel quality, turbulence intensity, received power, bit error rate, or Q factor of each candidate synthetic aperture unit; wherein the turbulence intensity is characterized by the scintillation index of the received branch signal, and the scintillation index is defined as: in Indicates the first The flicker index of the branch signal received by the path. Indicates the first The received light intensity of the branch signal in the path. This represents the variance of the received light intensity. and These represent the time mean of the received light intensity and the time mean of the square of the received light intensity, respectively; the larger the scintillation index, the stronger the power fluctuation of the corresponding received branch signal caused by atmospheric turbulence.

[0012] In a preferred embodiment, the number of received branch signals corresponding to the optimal beam combination. One to Any integer value in the range allows the optimal beam combination to include two, three, or more receiving branch signals, or it may exclude receiving branch signals that are severely affected by atmospheric turbulence; when At that time, the candidate beam combination state set contains a total of The candidate combination states, and the number of received branch signals corresponding to the optimal beam combination. Taken from 1, 2, 3 or 4; the method is applied to single-link transmission scenarios from a single satellite to a single ground station, multi-link fusion transmission scenarios from multiple satellites to a single ground station, or multi-link networking transmission scenarios from multiple satellites to multiple ground stations, in order to suppress light intensity flicker, power fluctuations and communication interruptions of the satellite-to-ground laser communication link caused by atmospheric turbulence, foggy weather or link fading.

[0013] This application also provides an anti-turbulence star-to-ground laser communication system based on wavelength-polarization reconfigurable synthetic aperture receiver, including: The transmitting module is configured to modulate satellite-to-ground downlink data signals onto optical carriers of at least two different wavelengths and to perform polarization orthogonal modulation on the optical carriers to form an incoherent multi-beam transmission signal with wavelength and polarization dimensions. The beam pointing module is configured to control the pointing of the incoherent multi-beam transmission signal output by the transmitting module, so that the incoherent multi-beam transmission signal is transmitted downlink along the satellite-to-ground laser communication link. The multi-aperture receiving module is configured to receive the incoherent multi-beam transmitted signal after transmission through atmospheric turbulence, and output the optical signal corresponding to each receiving aperture. The wavelength-polarization separation module is configured to sequentially perform polarization separation and wavelength separation on the optical signal output by the multi-aperture receiving module to obtain N receiving branch signals. Each receiving branch signal is characterized by wavelength identifier, polarization identifier, aperture identifier and path identifier and serves as a candidate synthetic aperture unit. A reconfigurable synthetic aperture receiver module is configured to construct a candidate beam combination state set based on the N candidate synthetic aperture units. The candidate beam combination state set contains all combinations in which m beams are selected from the N candidate synthetic aperture units for incoherent merging. The total number of candidate beam combination states is The received branch signals corresponding to the optimal beam combination output by the global fusion enhanced multi-branch network module are noncoherently combined to recover the satellite-to-ground downlink data signal. The globally fused enhanced multi-branch network module includes N+1 branches. The first N branches receive the corresponding received branch signals and extract independent features of each path through sub-networks. The N+1th branch receives the weighted fused data of the N received branch signals. The joint features between the N received branch signals are extracted using a two-dimensional convolutional kernel. The independent features and the joint features are then concatenated and passed through a hidden layer and an output layer. The optimal beam combination under the current turbulent conditions is identified and output from the candidate beam combination state set and sent to the reconfigurable synthetic aperture receiving module. Indicates the first The path receives branch signals. Indicates the first The weighting coefficients corresponding to the branch signals received by the path. .

[0014] In a preferred embodiment, the transmitting module includes a wavelength demultiplexer and a polarization controller. The wavelength demultiplexer is used to separate at least two incoherent wavelengths, and the polarization controller is used to achieve orthogonal polarization output of X-polarization and Y-polarization for each optical carrier. The beam pointing module includes a voice coil motor-driven fast steering mirror array, a micromirror array antenna, or an optical scanner. The multi-aperture receiving module includes a dual-aperture receiving antenna, a multi-aperture receiving array, or a distributed receiving array composed of multiple ground stations. The wavelength-polarization separation module includes a half-wave plate, a polarization beam splitter, or a polarization separation device, as well as a wavelength selection switch, an optical filter, or a wavelength demultiplexer. The first N branches of the globally fused enhanced multi-branch network module each include a sliding window module and a sub-network, and the N+1th branch includes a convolution kernel with a size of [missing information]. The two-dimensional convolutional module and the output of the two-dimensional convolutional module are... The flattening is an M-dimensional flattening module. The independent features and the joint features are concatenated by a Concat layer and then input into the hidden layer and the output layer.

[0015] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0016] According to embodiments of this application, by modulating downlink data signals onto at least two optical carriers of different wavelengths at a satellite communication terminal and performing polarization orthogonal modulation on the optical carriers to form an incoherent multi-beam transmission signal with wavelength and polarization dimensions, when the incoherent multi-beam transmission signal spatially overlaps in the far-field receiving plane, since the different wavelength channels are incoherent and the different polarization channels are orthogonally separable, the strong phase correlation interference term caused by traditional coherent multi-beam transmission is no longer generated. The synthesized light intensity obtained at the receiving end can be regarded as an incoherent superposition of the light intensities of each beam. This reduces the impact of coherent interference on beam separation, channel estimation, and signal combining from the source. It also enables the entire scheme to complete multi-beam combining without relying on active phase synchronization and real-time coherent synthesis. This avoids the high requirements for active phase synchronization accuracy imposed by time-varying phase distortion introduced by satellite motion, dynamic changes in propagation distance, and time-varying atmospheric channel characteristics in long-distance satellite-to-ground links. It also reduces the dependence on complex adaptive optics hardware such as optical phased arrays, wavefront sensors, and deformable mirrors, simplifying the overall implementation complexity of satellite-to-ground laser communication links.

[0017] Furthermore, by setting up a wavelength demultiplexer, a polarization controller, and a voice coil motor-driven fast steering mirror array, micromirror array antenna, or optical scanner, the satellite communication terminal can complete multi-wavelength separation and orthogonal polarization output in both the wavelength and polarization dimensions. It can also achieve multi-directional independent pointing of multiple wavelength-polarized light beams and complete far-field alignment through a beam pointing device. This allows beams of different wavelengths and polarizations to be independently guided to the ground receiver, forming a far-field superposition structure with resolvable multi-beam distribution on the receiving plane. This provides a physical guarantee from the transmitting side for the subsequent receiver to identify and separate each incoherent beam under multi-aperture and multi-path conditions.

[0018] At the ground station receiver, a multi-aperture receiving device, consisting of a dual-aperture receiving antenna, a multi-aperture receiving array, or a distributed multi-ground station receiving array, is used in sequence with a half-wave plate, a polarization beam splitter, a wavelength selective switch, an optical filter, or a wavelength demultiplexer. This process sequentially performs spatial polarization orthogonal separation and different wavelength channel separation on the optical signal received by each receiving aperture, resulting in the final obtained... The received branch signals can all be jointly characterized by the four elements of wavelength identifier, polarization identifier, aperture identifier and path identifier, thereby substantially expanding the physical meaning of "synthetic aperture" from the traditional single spatial dimension to a composite multi-dimensional dimension of wavelength-polarization-aperture-path. On the one hand, this enriches the diversity degrees of freedom that can participate in merging, so that received branch signals from different satellites to the same ground station, the same satellite to different ground stations, or the same beam through different atmospheric regions can all be included in a unified synthetic aperture framework for processing. On the other hand, it provides a complete and diverse set of equivalent aperture units as selection objects for subsequent reconfigurable selection actions.

[0019] Through the The candidate synthetic aperture unit constructs a set of candidate beam combination states covering all feasible subsets, thus increasing the number of received branch signal paths involved in the merging process. exist to The range of values ​​can be arbitrarily changed, and it is allowed to actively exclude a signal from the optimal beam combination when the signal of a certain receiving branch is severely affected by atmospheric turbulence. This breaks away from the constraints of the fixed combination method used in traditional multi-aperture diversity reception and avoids the drag on the overall combining gain by the severely damaged receiving branch signal. The optimal combining performance does not necessarily increase monotonically with the increase of the number of receiving branch signals participating in the combining, but rather matches the current channel state of each candidate synthetic aperture unit. Thus, it can take into account the robustness of combining gain and interruption probability in multi-link fusion transmission scenarios with different turbulence conditions, different equivalent transmission distances, and different link qualities.

[0020] In the above Before the received branch signals are input into the globally fused enhanced multi-branch network, each received branch signal undergoes sequential digital signal processing, including IQ orthogonalization, dispersion compensation, constant modulus algorithm, frequency offset recovery, and phase offset recovery, to obtain high-quality constellation point data for the network. Road input data This allows the feature extraction and combination state recognition of the network to be based on reliable constellation data that has already undergone linear damage and frequency phase offset correction, which is beneficial for the network to maintain robust optimal beam combination recognition capability under different receiving power and different turbulence conditions.

[0021] More importantly, by leveraging the previous Each branch extracts independent features from the received branch signal via a sliding window module and a sub-network. Branches via The two-dimensional convolution module and the flattening module are described in Weighted fusion data of branch signals received by the path Multiple joint correlation features are extracted along the branch direction. The independent features and the joint features are then concatenated through a Concat layer and sent to a specific topology configuration in the hidden layer and output layer. This allows the globally fused enhanced multi-branch network to simultaneously utilize independent information such as the constellation distribution, power fluctuation, and error distribution of each received branch signal when identifying the optimal beam combination, as well as joint information such as correlation changes and synchronicity differences caused by turbulence between different wavelength channels, different polarization channels, different apertures, or different paths. The independent feature branches and joint feature branches complement each other structurally. Compared with the traditional maximum ratio combining method that only relies on instantaneous channel gain for weighting, it can achieve higher combining gain and lower bit error rate in the operating range of low received power, strong nonlinear noise, or high correlation of multiple received branch signals.

[0022] By analyzing the channel quality, turbulence intensity, received power, bit error rate, and other parameters of each candidate synthetic aperture unit... Multiple indicators, such as factors, are used as the basis for identifying the optimal beam combination, with the scintillation index being the most important. The intensity of power fluctuations in the received branch signals caused by atmospheric turbulence is characterized, enabling the selection of the optimal beam combination to simultaneously reflect the instantaneous channel gain and the statistically significant degree of turbulence impairment. When one or more received branch signals exhibit strong flicker or deep fading, the network can proactively exclude these degraded branches from the optimal beam combination based on the correlation differences between the branches. This allows for real-time reconfigurable configuration of the synthetic aperture receiver structure by dynamically updating the optimal beam combination as atmospheric turbulence changes over time.

[0023] After determining the optimal beam combination, the method only performs incoherent combining of the corresponding received branch signals. The combining process is based on intensity superposition without active phase synchronization, giving the method and system good robustness to time-varying phase distortions introduced by satellite motion, propagation distance changes, and atmospheric channel dynamics in the satellite-to-ground downlink. Simultaneously, due to the large number of received branch signals participating in the combining... Available to The values ​​can be any values ​​between the given values, and the total number of candidate combinations is . This enables the method and system to be uniformly adapted to single-link transmission scenarios from a single satellite to a single ground station, multi-link fusion transmission scenarios from multiple satellites to a single ground station, and multi-link networking transmission scenarios from multiple satellites to multiple ground stations. It has good suppression capabilities for light intensity flicker, power fluctuations, and communication interruptions caused by atmospheric turbulence, cloud and fog weather, or link fading, which helps to improve the reliability and flexible networking capabilities of satellite-to-ground downlink laser communication links in various typical networking scenarios.

[0024] Overall, the transmitting module, beam pointing module, multi-aperture receiving module, wavelength-polarization separation module, reconfigurable synthetic aperture receiving module, and globally fused enhanced multi-branch network module are combined in the same system according to the above-mentioned cooperative relationship, and the corresponding computer programs are carried on a computer-readable storage medium. This makes the technical solution provided by this application feasible at the method, system, and storage medium levels, facilitating the deployment of corresponding modules or programs at different physical locations such as satellite communication terminals, ground station receiving terminals, and connected data processing centers to form a complete anti-turbulence satellite-to-ground laser communication link. The above-mentioned processes are not independent of each other, but constitute a cooperative chain composed of physical domain incoherence construction, multi-dimensional candidate synthetic aperture unit definition, candidate beam combination state set construction, networked optimal combination identification, and incoherent merging of selected branches. It is the overall effect of this cooperative chain that enables the method and system to obtain multi-dimensional, reconfigurable, and adaptive suppression capabilities against atmospheric turbulence without relying on active phase synchronization and complex adaptive optics hardware.

[0025] Furthermore, for satellite communication terminals, the satellite-to-ground downlink data signal to be transmitted is generated and modulated onto at least two optical carriers of different wavelengths. Through polarization modulation, the different optical carriers are further positioned in at least two mutually orthogonal polarization states, thus forming an incoherent multi-beam transmission signal containing both wavelength and polarization dimensions. Different wavelength channels are incoherent due to their different optical frequencies, while orthogonal polarization channels are distinguishable due to their orthogonal polarization states. Therefore, when multiple beams spatially overlap in the far-field receiving plane, the strong phase-correlation interference term caused by traditional coherent multi-beam transmission is no longer generated. The light intensity obtained at the receiving end can be considered as an incoherent superposition of the light intensities of each beam, thereby reducing the impact of coherent interference on beam separation, channel estimation, and signal combining.

[0026] For multi-beam downlink transmission, the transmitting end uses a beam pointing device to control the direction of the incoherent multi-beam transmitted signal, enabling the multiple optical signals to be transmitted along the satellite-to-ground laser communication link. The beam pointing device may include a voice coil motor-driven fast-steering mirror array, a micromirror array, an optical scanner, or other devices capable of multi-beam pointing control. Through this device, beams of different wavelengths and polarizations can be guided to the ground receiver, forming a multi-beam distribution in the far-field receiving region.

[0027] For the ground station receiver, a multi-aperture receiving device is used to receive multiple optical signals transmitted through atmospheric turbulence. This multi-aperture receiving device can include a dual-aperture receiving antenna, a multi-aperture receiving array, or a distributed receiving array composed of multiple ground stations. The optical signals received by each receiving aperture carry different aperture and path information, which can serve as candidate aperture units for subsequent synthetic aperture reception. Further, the ground receiver performs polarization separation and wavelength separation on the received multiple optical signals. For example, orthogonal polarization component separation can be achieved using a half-wave plate, polarization beam splitter, or polarization separation device, and different wavelength channels can be separated using a wavelength selection switch, optical filter, or wavelength demultiplexer. After the above processing, received branch signals in multiple wavelength-polarization dimensions are obtained.

[0028] For reconfigurable wavelength-polarization synthetic aperture, this application uses the received branch signals corresponding to different wavelength channels, different polarization channels, and different receiving apertures as selectable aperture units for the synthetic aperture. These selectable aperture units can be flexibly combined according to the current channel state.

[0029] The term "synthetic aperture" refers to the fact that multiple received signals obtained from different wavelengths and polarizations are not fixedly combined, but rather participate in the construction of an equivalent aperture as multiple selectable receiving units. By selecting and fusing these receiving units, the system can form different equivalent receiving aperture structures.

[0030] The term "reconfigurable" refers to the system's ability to select the optimal combination from multiple candidate states based on the degree to which each received signal is affected by atmospheric turbulence. For example, when there are N wavelength-polarization synthetic aperture units, the system can select m of them for incoherent merging, where... Its candidate combination number can be expressed as: The total number of combinations supported by the reconfigurable wavelength-polarization synthetic aperture receiver is: in, This represents the total number of candidate synthetic aperture combinations. This indicates the number of wavelength-polarization composite aperture units involved in the merging process. When... At this time, the system can support 15 candidate combination methods from one to four channels. This mechanism allows the receiver to select two, three, four or more signals to combine based on different turbulence conditions, instead of always using a fixed all-combining method.

[0031] To achieve the wavelength-polarization synthetic aperture characteristics, a globally fused enhanced multi-branch network is further introduced. DSP processing is performed on the received signals corresponding to different wavelengths, polarizations, and apertures, including IQ-path orthogonalization, dispersion compensation, constant mode algorithm, frequency offset recovery, and phase offset recovery, thereby obtaining high-quality constellation points. The processed multi-channel data is then input into a signal fusion network for weighted fusion. This fusion network is a multi-branch neural network used to automatically learn the merging weights of each data channel through training, avoiding the subjectivity and limitations of manually setting weights. The multi-branch neural network includes N+1 branches. The first N branches receive the corresponding N processed data channels and extract the independent features of each channel through sub-networks; the N+1th branch receives the weighted fusion result of the N processed data channels and extracts the joint features between the multiple data channels through a two-dimensional convolutional kernel. The output features of the N+1 branches are concatenated and then input to the output layer through a hidden layer to obtain the fused signal output. This structure can simultaneously learn the independent features of each signal and the correlation features between multiple signals, thereby improving the signal fusion effect. Compared to the traditional maximum ratio combining algorithm, this method can improve the signal-to-noise ratio of the combined signal and reduce the bit error rate.

[0032] Furthermore, this network is used for feature extraction, combination state identification, and optimal combination selection of multiple received branch signals. Specifically, received signals from different wavelengths, polarizations, apertures, or paths are treated as multiple independent branches of the network. The network's fusion characteristics are used to freely combine different branches to identify the current optimal beam combination state. Based on this combination state, the system selects the corresponding received branch signals for incoherent combining, completing a reconfigurable wavelength-polarization synthetic aperture receiver. This receiver treats each received signal as a selectable synthetic aperture unit and adaptively selects the optimal one or more received signals for incoherent combining based on atmospheric turbulence-induced light intensity flicker, power fluctuations, bit error rate changes, or channel quality changes. Thus, the receiver avoids the performance degradation problem of fixed combining methods under strong turbulence or channel imbalance conditions, achieving flexible suppression of the effects of atmospheric turbulence and reducing the probability of communication interruption in the space-to-ground laser communication link.

[0033] The specification of this application contains numerous technical features distributed across various technical solutions. Listing all possible combinations of these technical features (i.e., technical solutions) would make the specification excessively lengthy. To avoid this problem, the various technical features disclosed in the above-described invention, the various technical features disclosed in the following embodiments and examples, and the various technical features disclosed in the accompanying drawings can be freely combined to form various new technical solutions (all of which are considered to have been described in this specification), unless such a combination of technical features is technically infeasible. For example, one example discloses feature A+B+C, and another example discloses feature A+B+D+E. Features C and D are equivalent technical means that serve the same function, and technically only one needs to be used; they cannot be used simultaneously. Feature E can technically be combined with feature C. Therefore, the solution A+B+C+D should not be considered as described because it is technically infeasible, while the solution A+B+C+E should be considered as described. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of a reconfigurable wavelength-polarization synthetic aperture receiver based on a wavelength-polarization reconfigurable synthetic aperture receiver according to an embodiment of this application for anti-turbulent star-to-ground laser communication.

[0035] Figure 2 This is a theoretical comparison block diagram of incoherent multibeam transmission and coherent multibeam transmission according to embodiments of this application.

[0036] Figure 3 This is a schematic diagram of the structure of a globally fused enhanced multi-branch network according to an embodiment of this application.

[0037] Figure 4 This is a schematic diagram of an experimental setup for an incoherent four-beam atmospheric turbulence transmission system based on a wavelength-polarization synthetic aperture receiver, according to an embodiment of this application.

[0038] Figure 5 Based on the embodiments of this application A schematic diagram of the test results for wavelength-polarized incoherent four-beam transmission of a voice coil motor driven fast-steering mirror array.

[0039] Figure 6 This is a schematic diagram showing the comparative test results of incoherent four-beam combining performance based on a wavelength-polarization synthetic aperture receiver according to an embodiment of this application.

[0040] Figure 7 This is a schematic diagram of the test results of the incoherent four-beam atmospheric turbulence transmission performance based on a wavelength-polarization synthetic aperture receiver according to an embodiment of this application.

[0041] Figure 8 This is a schematic diagram of the test results of multipath multibeam transmission based on a reconfigurable wavelength-polarization synthetic aperture receiver according to an embodiment of this application.

[0042] Figure 9 This is a schematic flowchart of an anti-turbulence star-to-ground laser communication method based on wavelength-polarization reconfigurable synthetic aperture receiver according to an embodiment of this application. Detailed Implementation

[0043] In the following description, many technical details are presented to help the reader better understand this application. However, those skilled in the art will understand that the technical solutions claimed in this application can be implemented even without these technical details and various variations and modifications based on the following embodiments.

[0044] Explanation of some concepts: Incoherent multi-beam transmission signals refer to multiple transmitted optical signals formed by at least two optical carriers of different wavelengths carrying downlink data signals and undergoing polarization orthogonal modulation. The different wavelength channels are incoherent with each other due to their different optical frequencies, and the different polarization channels can be distinguished from each other due to their orthogonal polarization states. Therefore, when the multiple optical signals spatially overlap in the far-field receiving plane, they no longer generate strong phase-correlated coherent interference terms.

[0045] Synthetic aperture, in this application, refers to the equivalent receiving aperture structure formed by unifying the receiving branch signals corresponding to different wavelength channels, different polarization channels, different receiving apertures, and different transmission paths as optional aperture units. Its physical meaning is that the receiving branch signals in multiple diversity dimensions constitute a set of receiving apertures composed of multi-dimensional diversity degrees of freedom in an equivalent sense. It is different from the traditional synthetic receiving structure composed of multiple spatially distributed physical apertures, and also different from the synthetic aperture defined in synthetic aperture radar by synthesizing the scanning results of the same physical aperture at different times by a moving platform.

[0046] A candidate synthetic aperture unit (SAR) is a received branch signal characterized by a quadruple of wavelength identifier, polarization identifier, aperture identifier, and path identifier. It is the smallest equivalent aperture unit that constitutes the synthetic aperture and participates in the reconfigurable combination selection.

[0047] Reconfigurability means that the receiving end is no longer limited to a fixed full merging method, but allows it to switch from... In the candidate synthetic aperture unit of the path, The system adaptively selects the receiving branch signals to participate in incoherent combining with any integer number of channels. When a receiving branch signal is severely affected by atmospheric turbulence, the selected optimal beam combination may not include that receiving branch signal, thus enabling the equivalent receiving aperture structure to switch online as turbulence conditions change.

[0048] The candidate beam combination state set refers to the set of states derived from the above. In the candidate synthetic aperture unit of the road, according to all feasible The set consisting of all combinations obtained by selecting values ​​has a total number of values. .

[0049] Optimal beam combination refers to the beam combination selected by a globally fused enhanced multi-branch network from the set of candidate beam combination states, based on the channel quality, turbulence intensity, received power, bit error rate, or other parameters of each candidate synthetic aperture unit under the current turbulence conditions. The set of received branch signals identified and output by the factor for use in incoherent combining.

[0050] Globally integrated enhanced multi-branch networks refer to networks that include A neural network structure with parallel branches, where the first branch... Each branch receives the corresponding path's received branch signal and extracts the independent features of each path through the sub-network. Each branch receives the The weighted fusion data of the branch signals received by the path is extracted by a two-dimensional convolutional kernel. The joint features between the received branch signals are spliced ​​together with the independent features and the joint features, and then output as an optimal beam combination through a hidden layer and an output layer.

[0051] Independent features refer to the front end of the globally fused enhanced multi-branch network. The features extracted by each subnetwork within a branch from the corresponding received branch signal within a sliding window range, used to characterize the statistical properties of that single received branch signal.

[0052] Joint features refer to the first characteristic of the globally fused enhanced multi-branch network. The two-dimensional convolutional module in each branch extracts the data from the weighted fused data along the inter-branch direction, which is used to characterize the data. The correlation between the received branch signals and the characteristics of multipath joint damage caused by atmospheric turbulence.

[0053] Incoherent combining refers to a combining method that is based on the superposition of the light intensity or power of the received branch signals and does not require active phase synchronization of the light fields of each path. This is different from coherent combining, which is based on phase control to achieve multi-beam synthesis in the optical domain.

[0054] The scintillation index is the ratio of the variance of the received light intensity of the receiving branch signal to the square of its mean. It is used to quantify the intensity of power fluctuations caused by atmospheric turbulence. The larger the value, the more severe the influence of atmospheric turbulence on the corresponding receiving branch signal.

[0055] Aperture identification refers to the identification information used to identify which receiving aperture receives the receiving branch signal at the ground station receiving end.

[0056] Path identification refers to the identification information used to identify the transmission path through which the received branch signal passes. The transmission path includes, but is not limited to, different downlinks from different satellites to the same ground station, different downlinks from the same satellite to different ground stations, or different equivalent links corresponding to the same beam passing through different atmospheric regions.

[0057] The following section further describes this application with reference to specific experimental methods and results.

[0058] In this embodiment, an incoherent four-beam satellite-to-ground laser communication experimental system based on wavelength-polarization reconfigurable synthetic aperture receiver is constructed to verify the link transmission and signal combining performance of the method under atmospheric turbulence conditions. Figure 4 As shown, the experimental system includes a transmitter, a beam pointing module, a free-space optical communication equivalent link module, an atmospheric turbulence simulation module, a receiver, and a signal processing module.

[0059] At the transmitting end, an arbitrary waveform generator produces two data signals to be transmitted, which are then modulated onto optical carriers. After being amplified by an optical amplifier, the two modulated optical signals are adjusted to mutually orthogonal polarization states by a polarization controller, thereby forming an incoherent four-beam transmission signal containing dual-wavelength and dual-polarization information.

[0060] In the beam pointing module, the incoherent four-beam transmitted signal is collimated by a two-dimensional transmitting antenna array and then enters a voice coil motor-driven fast-steering mirror array. The fast-steering mirror array is used to adjust the exit direction of each beam, enabling far-field pointing and alignment of the dual-wavelength, dual-polarization beams. This step allows multiple beams of light to enter the equivalent free-space optical communication link and form a collectable multi-beam spot at the receiving end.

[0061] In the free-space optical communication equivalent link module, the equivalent link includes a first off-axis Newtonian collimator, a second off-axis Newtonian collimator, a beam splitter, a CCD camera, and an adjustable optical attenuator. The first collimator is used to map the emitted light field into a far-field light spot distribution, and the CCD camera is used to observe the state of the far-field light spot; the adjustable optical attenuator is used to adjust the optical power at the receiving end to simulate the link loss under different equivalent transmission distances; the second collimator is used to re-collimate the optical signal after power adjustment and output it as parallel light.

[0062] In the atmospheric turbulence simulation module, a rotatable phase plate is set in the transmission path to simulate wavefront distortion, beam drift, beam spread, and intensity flicker caused by atmospheric turbulence on beam propagation. By changing the phase plate parameters or rotation state, atmospheric turbulence channels with different intensities and random distributions can be generated.

[0063] At the receiving end, the optical signal, after transmission through the equivalent link, enters the dual-aperture receiving antenna. Each receiving aperture separates the orthogonal polarization components using a half-wave plate and a polarization beam splitter, resulting in X-polarization and Y-polarization receiving branches. Subsequently, a wavelength demultiplexer is used to separate the different wavelength channels, thereby obtaining receiving branch signals in multiple wavelength-polarization-aperture dimensions.

[0064] In the signal processing module, the signals from each receiving branch are converted into electrical signals after optical amplification, polarization mixing, photoelectric detection, and transimpedance amplification. These electrical signals then enter the digital signal processing module for IQ demultiplexing, equalization, frequency offset estimation, carrier phase recovery, and bit error rate calculation to obtain the communication quality characteristics of each branch. Finally, the four receiving branch signals are input into a globally fused enhanced multi-branch network. This network performs feature extraction, combination state identification, and optimal combination selection for the received signals corresponding to different wavelengths, polarizations, and apertures. Depending on the degree of damage to each signal under current turbulent conditions, the network can select two, three, or four signals for incoherent merging to form a reconfigurable wavelength-polarization synthetic aperture reception result, reducing the probability of communication interruptions caused by atmospheric turbulence and restoring satellite-to-ground downlink data signals.

[0065] The following section provides a detailed explanation of some of the technologies used in this implementation case.

[0066] Voice coil motor driven fast-reflecting mirror array: This fast-reflecting mirror array enables rapid scanning and pointing, providing support for beamforming in satellite-to-ground downlink communication. The array has 2×2 elements, a 40mm diameter element, and an RMS value ≤1 / 20λ (λ=632.8nm). It is suitable for multi-directional pointing of four incoherent collimated beams with dual polarization and dual wavelengths.

[0067] Free-space optical communication equivalent link: The equivalent link can flexibly adjust different equivalent transmission distances. The Newtonian collimator has a focal length of 2.5m and an aperture of 250mm. The adjustable attenuator has an attenuation range of 0dB to 45dB. The equivalent distance calibration method adopts the Airy disk far-field diffraction theory. When parallel light rays pass through a disk with a diameter of... When the aperture is circular, after traveling the required propagation distance, the far-field distribution of the Airy disk can be described as follows: in This represents the total transmit power of Airyban. and These represent the diameter of the aperture and the transmission distance, respectively. Approximately 84% of the light energy of the Airy disk beam is concentrated at the central bright spot, and its theoretical optical power density can be expressed as: In this embodiment, the laser wavelength is set to 1550nm, the aperture diameter to 20mm, and the output power to 490mW. The theoretical far-field power density at different distances is calculated through simulation. For the receiver, an optical power detection device with a 100mm diameter can be used to measure the optical power density. The adjustable optical attenuator is adjusted to match the measured value with the theoretical prediction value at a given equivalent distance. Multiple measurements are taken and the attenuation value of the adjustable attenuator at the corresponding equivalent distance is recorded to complete the distance calibration.

[0068] Atmospheric turbulence simulation device: The turbulence phase plate is a transparent optical element whose surface refractive index distribution is designed according to specific Kolmogorov statistical laws, capable of simulating the phase distortion caused by a light beam passing through turbulent atmosphere. The parameters of the three turbulence phase plates selected in this embodiment are... They are 3mm, 2mm, and 1.5mm respectively. Atmospheric coherence length (a characteristic parameter characterizing the intensity of atmospheric turbulence) The smaller the value, the stronger the turbulence. Note the phase plate. The parameter is not equivalent to the Friedel-Crafts coherence length measured in an outdoor atmospheric channel; here it is used only to represent the relative turbulence intensity, corresponding to strong, moderate, and weak turbulence, respectively. Scintillation index ( It is often used to quantify the intensity of power fluctuations caused by turbulence. It involves calculating the received optical power. The ratio of the variance to the square of its mean, i.e. The scintillation index is obtained to characterize the intensity of power fluctuations caused by turbulence. The larger the scintillation index, the stronger the power fluctuations caused by turbulence.

[0069] Furthermore, by driving the turbulent phase plate to rotate using a motor, the temporal evolution characteristics of turbulence can be simulated, i.e., the change of phase distortion caused by turbulence over time. After passing through the rotating phase plate, the collimated optical signal is introduced with wavefront distortion, resulting in random fluctuations in the beam intensity distribution and random drift of the beam centroid position, thus realistically simulating the impact of atmospheric turbulence on the satellite-to-ground transmission link.

[0070] The experimental results and performance analysis are presented below. (1) Wavelength-polarized four-beam incoherent transmission performance based on voice coil motor fast reflector array: The 2×2 fast-direction mirror array comprises four independent micromirror units, each used to adjust the emission direction of its corresponding beam. The optical pointing angles of the four micromirror units were tested in the experiment. The results show that the optical pointing range of the fast-direction mirror array can reach ±23.6° in the X-axis direction and ±16.8° in the Y-axis direction. Figure 5 As shown in (a).

[0071] Furthermore, to evaluate the impact of the pointing error of the micromirror array on the far-field superposition performance, the relationship between the normalized coupled optical power and the pointing error was measured. The results in Figure (5) show that as the pointing error increases, the coupled optical power gradually decreases, and the tracking error of each micromirror unit is less than 6 μrad. The fast-steering mirror array can meet the accuracy requirements of incoherent multi-beam far-field alignment and superposition.

[0072] In the far-field superposition performance test, an off-axis Newtonian collimator with a diameter of 250mm and a focal length of 2.5m was used to evaluate the effect of multi-beam far-field superposition. The stability of the four-beam far-field superposition was obtained by continuously measuring the coupled optical power after superposition of four incoherent beams. When multiple beams have the same wavelength and non-orthogonal polarization states, the received optical power is affected by coherent interference, resulting in random fluctuations. In contrast, when a decorrelated beam is used and placed in orthogonal polarization states, the received optical power can achieve incoherent superposition, with a superposition efficiency of 91.9%. Figure 5 (c) Furthermore, the isolation between the wavelength channel and the polarization channel was tested, and the results showed that the isolation was greater than 27 dB, with the wavelength channel showing better isolation than the polarization channel. Higher wavelength-polarization isolation is beneficial for subsequent separation, selection, and fusion of different wavelengths and polarization channels.

[0073] (2) Wavelength-polarization synthesized aperture receiving performance of incoherent four-beam transmission system: In this embodiment, the wavelength-polarization synthetic aperture receiving performance of the incoherent four-beam transmission system is experimentally verified. The four beams are respectively... , , and This corresponds to four wavelength-polarized beams formed by optical carriers with frequencies of 193.3 THz and 193.35 THz in X-polarization and Y-polarization states. Using received optical power as a variable, the Q-factor variation of the 16-QAM signal was tested under single-beam transmission and under two-beam, three-beam, and four-beam merging conditions. Figure 6 (a) The results show that at a bit error rate threshold of 3.8 × 10⁻⁶, -3Compared with a single-input single-output system, the two-beam, three-beam, and four-beam combining based on the global fusion enhanced multi-branch network achieve Q-factor gains of 5.5dB, 7.3dB, and 8.4dB, respectively.

[0074] Furthermore, the correlation between the four beams and the combining gain under different beam counts were tested. At a received optical power of -44 dBm, the correlation coefficient between the four beams remained low, indicating that different wavelength-polarization channels can provide effective diversity reception conditions. After combining using GFE-MBN-2, GFE-MBN-3, and GFE-MBN-4 methods, the Q factors of the combined signals reached 3.99 dB, 5.48 dB, and 6.66 dB, respectively.

[0075] Compared to traditional maximum ratio combining methods, the global fusion enhanced multi-branch network achieves higher combining gain, and its advantages become more pronounced when the received optical power decreases. This is because the network can learn the nonlinear noise characteristics in the system and extract correlation features between multiple received signals through the (N+1)th fusion branch, thereby improving the performance of multi-beam signal fusion and decision-making.

[0076] (3) Atmospheric turbulence suppression performance of incoherent wavelength-polarization synthetic aperture receiver: In this embodiment, the wavelength-polarization reconfigurable synthetic aperture (SAP) receiving performance of incoherent multibeam transmission under different atmospheric turbulence intensities is experimentally verified. In the experiment, the Q-factor, received optical power, bit error rate, and communication interruption probability of the 16-QAM signal are used as performance evaluation indicators. Comparative tests are conducted on single-beam reception, two-beam combining, three-beam combining, four-beam traditional maximum ratio combining, and four-beam combining based on a globally fused enhanced multi-branch network. Figure 7 As shown.

[0077] exist Under weak turbulence conditions, the received signal was measured multiple times consecutively, with each measurement interval of 1.5 seconds. Experimental results show that light intensity flicker caused by atmospheric turbulence leads to random fluctuations in the received optical power, thereby increasing the probability of communication interruption. In this embodiment, the mean received optical power was -35.5 dBm, and its 3σ power fluctuation range was 2.1 dB. A bit error rate threshold of 3.8 × 10⁻⁶ was used. -3 The corresponding Q-factor value of 8.52 dB was used as the criterion for communication interruption. The communication interruption probabilities for the four diversity receiving beams were 74.26%, 72.27%, 70.29%, and 76.24%, respectively. After employing four-beam combining based on a globally fused enhanced multi-branch network, the maximum gain improvement reached 3.4 dB compared to single-beam receiving. As the number of beams participating in combining increased, the combining gain gradually increased, and no communication interruption was observed under these experimental conditions.

[0078] Furthermore, the interruption probability suppression capability of the reconfigurable wavelength-polarization synthetic aperture receiving method was verified under stronger turbulence conditions. When the received optical power fluctuation ranges were 8.4 dB and 21.4 dB, the corresponding turbulence intensities were... and The four-beam combining based on the globally fused enhanced multi-branch network achieved combining gains of 4.5 dB and 5.2 dB, respectively. Experimental results show that different wavelength channels and orthogonal polarization channels have a weak correlation with the response to atmospheric turbulence. Therefore, multi-beam incoherent combining can reduce the risk of communication interruption caused by severe turbulence affecting a single path.

[0079] (4) Reconfigurable combinatorial performance of incoherent wavelength-polarization synthesized aperture receivers: This study verifies the flexible merging capability of the reconfigurable wavelength-oscillation synthetic aperture receiving method in a multi-satellite, multi-link scenario with different orbits. To simulate a multi-link transmission scenario from multiple satellites to a ground station indoors, this embodiment constructs four equivalent free-space optical links using bidirectional dual-terminal free-space optical communication equivalent links, denoted as Link 1, Link 2, Link 3, and Link 4, respectively. The equivalent transmission distances of the four links are set to 650km, 600km, 550km, and 500km, respectively. Simultaneously, turbulent phase plates with different coherence length parameters are used to simulate the atmospheric turbulence affecting each link, resulting in varying degrees of light intensity flicker and received power fluctuations. Figure 8 The results in (a) and (b) show that the scintillation indices of the four links are 0.028, 0.259, 0.198, and 0.854, respectively, indicating that the power fluctuation of each link varies significantly under different equivalent transmission distances and different turbulence conditions. The post-turbulence light spot observed by the CCD camera also shows that the light spot expansion and dispersion become more pronounced as the turbulence intensity increases.

[0080] Furthermore, reconfigurable synthetic aperture receiver verification was performed on four equivalent free-space optical links, such as... Figure 8 As shown in (c), when four candidate links exist, the system can support all two-link, three-link, and four-link combinations, for a total of 11 candidate merging configurations. Since the four links differ in equivalent transmission distance and turbulence intensity, the beam combining gain does not necessarily increase monotonically with the increase in the number of merged links. Therefore, the method described in this application does not fixedly select more links for merging, but rather adaptively selects the optimal combination based on the current channel quality of each link.

[0081] Experimental results show that link 4 is relatively less affected by turbulence, and combinations including link 4 can significantly reduce the probability of communication interruption. In contrast, the combination [1,2,3], which does not include link 4, uses more merged links, but its interruption probability is still higher than combinations such as [2,4] and [3,4] that include link 4. This indicates that in multi-link, multi-beam satellite-to-ground laser communication scenarios, the optimal merging performance depends not only on the number of links involved in merging, but also on the degree of turbulence damage to each link and the channel quality.

[0082] The above experiments show that the reconfigurable wavelength-oscillating synthetic aperture receiving method proposed in this application can adaptively select the optimal merging configuration based on the propagation distance, turbulence intensity, received power distribution and communication quality differences of multiple links, thereby reducing the probability of communication interruption and improving the reliability and flexible networking capability of satellite-to-ground laser communication links in multi-satellite, multi-path and multi-beam transmission scenarios.

[0083] To make this application easier to understand, the following description, in conjunction with the accompanying drawings, provides a more detailed account of the anti-turbulence star-to-ground laser communication method and system based on wavelength-polarization reconfigurable synthetic aperture receiver. It should be noted that the following embodiments are merely illustrative of the technical concept of this application and not intended to limit its scope of protection; any equivalent substitutions, combinations, or improvements made by those skilled in the art without departing from the technical concept of this application should fall within the scope of protection of this application.

[0084] In summary, the technical difficulties this application seeks to overcome are not isolated within any single dimension of "wavelength diversity," "polarization diversity," or "multi-aperture reception." Rather, they focus on a specific technical contradiction in long-distance satellite-to-ground downlink laser communication scenarios: the highly heterogeneous degree of turbulence damage to signals from different receiving branches due to the coupling of multiple factors such as atmospheric turbulence, cloud cover, aerosol scattering, and pointing errors. Specifically, when some receiving branch signals are already in a state of deep fading or severe light flicker, the traditional approach of simply merging all signals or monotonically increasing the number of branches may actually lower the overall merging gain and increase the probability of communication interruption due to the incorporation of degraded branches. To address this specific contradiction, this application does not simply stack multiple wavelengths, multiple polarizations, multiple apertures, and neural network merging techniques side-by-side, but rather organically reorganizes them at the technical concept level: the transmitter relies on the incoherent multi-beam transmission signal constructed by the combined properties of wavelength incoherence and polarization orthogonality (e.g., , , , Multi-wavelength-polarized optical carriers, see [link / reference] Figure 1 , Figure 2This transforms far-field superposition from coherent interference to intensity addition, eliminating the rigid dependence on active phase synchronization and real-time coherent synthesis at the source; the receiving end then converts the received branch signals corresponding to different wavelength channels, different polarization channels, different receiving apertures, and different transmission paths. ( The unified abstraction is to define candidate synthetic aperture units jointly characterized by wavelength identifier, polarization identifier, aperture identifier, and path identifier. This substantially expands the physical meaning of "synthetic aperture" from a traditional single spatial dimension to a composite multi-dimensional dimension of wavelength-polarization-aperture-path. Based on this, a system covering all feasible subset selections with a total number of [number missing] is constructed. The set of candidate beam combination states; then, from which... A globally fused enhanced multi-branch network with parallel branches (see [reference]). Figure 3 Extract the received branch signals from each path using a dual-path method. The single-path independent characteristics and in weighted fusion data Upper Classic Two-dimensional convolutional kernels extract multi-path joint correlation features along the inter-branch direction, identify and output the optimal beam combination under the current turbulence conditions from the candidate beam combination state set, and allow active exclusion from the optimal beam combination when certain receiving branch signals are severely impaired (see [reference]). Figures 6 to 8 The aforementioned technical features are not separate functional additions, but rather interdependent and mutually presupposing: incoherent multi-beam transmission provides the physical basis for the separable reception of multi-dimensional candidate synthetic aperture units; the unified representation of multi-dimensional candidate synthetic aperture units provides an equivalent aperture unit source for the complete construction of the candidate beam combination state set; and the candidate beam combination state set provides a complete search space for the multi-branch network to identify the optimal beam combination. The specific topology of the branch, in turn, provides dual characteristics for the identification and decision-making of non-fixed subset selection and incoherent merging, ultimately forming a closed-loop anti-turbulence receiving mechanism that does not rely on active phase synchronization, can tolerate degraded branches, and can reconstruct the receiving aperture structure online according to the turbulence state. This achieves specific technical effects that cannot be directly derived from a single existing means compared to fixed merging methods.

[0085] Furthermore, through long-term and in-depth research, the inventors of this application have discovered that in long-distance satellite-to-ground laser communication scenarios such as downlink backhaul of high-resolution remote sensing satellites, downlink from multiple gateway stations in low-orbit broadband internet constellations, downlink relay from deep space probes to Earth, and collaborative networking of multiple satellites and multiple ground stations, the adverse effects of atmospheric turbulence, cloud and fog weather, aerosol scattering, and pointing errors on the satellite-to-ground downlink laser communication link are far more than just superficial issues such as "received power fluctuations" or "increased bit error rates." If we trace the technical roots further, we can find that there is a strong intrinsic correlation between several technical problems existing in the anti-turbulence receiver side of the current technology, and some of these problems have been neglected by traditional anti-turbulence solutions for a long time.

[0086] After a thorough analysis of the principles of existing anti-turbulence technologies, the inventors discovered that, firstly, while space diversity schemes, represented by multi-aperture receivers, can suppress light intensity flicker by utilizing the aperture averaging effect in a statistical sense, they often combine the outputs of all receiver apertures in a fixed combination manner. In multi-link fusion transmission scenarios involving multiple satellites in different orbits, multiple ground stations, or different equivalent transmission distances, the degree of turbulence damage to the signals of each receiver branch is not uniformly distributed. One or several receiver branch signals may be in a state of deep fading or severe light intensity flicker due to the high intensity of turbulence in the atmospheric regions they pass through. In this case, if the fixed method of combining all signals is used, the severely damaged branch will not only fail to make an effective contribution to the overall combining gain, but may also lower the overall quality of the combined decision signal due to the power fluctuations and error components it introduces. Through analysis of numerous real-world link transmission scenarios, the inventors further realized that the essence of this drag phenomenon cannot be alleviated simply by increasing the number of receiving apertures. Rather, it is precisely the inherent structural limitation of the fixed combining method itself when facing multiple receiving branch signals with significantly different degrees of damage. Existing solutions mostly use the number of receiving apertures as the only adjustable dimension, while ignoring the equivalent status of wavelength channels, polarization channels, and transmission paths in the construction of diversity degrees of freedom. As a result, the overall combining gain is negatively affected when degraded branches are introduced.

[0087] Secondly, through observation and research on a large number of actual link states, the inventors discovered that the diversity gain of electrical domain combining usually depends on the uncorrelation between the fading processes corresponding to different propagation paths. However, when multiple received branch signals exhibit strong correlation due to sharing atmospheric path segments or being affected by the same meteorological event, traditional combining methods such as maximum ratio combining and equal gain combining, which rely solely on instantaneous channel gain, are significantly limited in their combining effect because they cannot fully capture the joint impairment characteristics between multiple received branch signals. Therefore, the inventors creatively realized that only by simultaneously incorporating the independent characteristics of each received branch signal and the joint characteristics between multiple received branch signals into the basis of combining decision-making can the aforementioned combining gain bottleneck be overcome.

[0088] More importantly, after a thorough analysis of the implementation conditions of optical domain combining schemes such as optical phased arrays and coherent multibeam combining, the inventors realized that although such schemes can achieve multibeam combining in the optical domain through phase control, the time-varying phase distortion introduced by satellite motion, changes in propagation distance, and the time-varying characteristics of atmospheric channels in long-distance satellite-to-ground links poses severe challenges to the real-time performance and stability of any active phase synchronization control loop. Therefore, even though such coherent combining schemes theoretically have high combining efficiency, they are difficult to reliably meet the comprehensive requirements of real-time performance, reliability, and system complexity of satellite-to-ground downlinks at the engineering implementation level.

[0089] Through comparative reflection on the above-mentioned problems, the inventors creatively realized that in order to address the imbalance damage caused by heterogeneous turbulence without relying on active phase synchronization, it is necessary to simultaneously possess separable and identifiable multidimensional diversity degrees of freedom in the physical domain, and simultaneously possess reconfigurable and adaptive optimal combination identification capabilities in the algorithmic domain. Both are indispensable. If multidimensional diversity is introduced only in the physical domain while all branches are still merged in a fixed manner, the drag on the overall gain by degraded branches cannot be resolved. If more refined merging weights are introduced only in the algorithmic domain but the diversity degrees of freedom are insufficient or cannot be flexibly combined, there is a lack of selectable equivalent aperture unit sets as choices.

[0090] Based on the above in-depth research, the inventors of this application propose an innovative technical solution. Its core lies in constructing an incoherent multi-beam transmission structure in the physical domain using wavelength incoherence and polarization orthogonality. The received branch signals are jointly characterized by a four-dimensional identifier—wavelength, polarization, aperture, and path—and used as candidate synthetic aperture units. Based on this, a set of candidate beam combination states is constructed, and these states are then configured with specific... The globally fused enhanced multi-branch network of the branch topology identifies and outputs the optimal beam combination under the current turbulence conditions within this set, and then performs incoherent combining only the selected receiving branch signals. This achieves flexible suppression of atmospheric turbulence without relying on active phase synchronization, reducing the probability of communication interruption in the satellite-to-ground laser communication link. The implementation process of the method and system described in this application is explained in detail below through specific embodiments.

[0091] Specific Implementation Example 1: Anti-turbulent satellite-to-ground laser communication method based on wavelength-polarization reconfigurable synthetic aperture receiver This embodiment provides an anti-turbulence space-to-ground laser communication method based on wavelength-polarization reconfigurable synthetic aperture receiver. The overall technical concept of this method is as follows: at the transmitting end, an incoherent multi-beam transmission structure is constructed using wavelength incoherence and polarization orthogonality, so that when multiple beams are superimposed in the far field, no strong phase-correlated coherent interference terms are generated, thus avoiding dependence on active phase synchronization and real-time coherent synthesis from the source. At the receiving end, the receiving branch signals corresponding to different wavelength channels, different polarization channels, different receiving apertures, and different transmission paths are uniformly regarded as equivalent selectable aperture units of the synthetic aperture. A globally fused enhanced multi-branch network identifies and outputs the optimal beam combination under the current turbulence conditions from all candidate beam combination states. Then, the selected receiving branch signals are incoherently combined, thereby suppressing light intensity flicker and communication interruptions caused by atmospheric turbulence without relying on active phase synchronization. Specifically, this method may include the following steps.

[0092] Step 100: The satellite communication terminal modulates the downlink data signal onto at least two optical carriers of different wavelengths and performs polarization orthogonal modulation on the optical carriers to form an incoherent multi-beam transmission signal with wavelength and polarization dimensions.

[0093] Specifically, this step may include the following sub-steps. Step 110: At least two optical carriers of different wavelengths are provided by a multi-wavelength light source in the satellite communication terminal, the at least two wavelengths being spaced sufficiently to be mutually incoherent in optical frequency. For example, two or more independently distributed feedback semiconductor lasers can be used as the light source, or a tunable laser can be used in conjunction with a wavelength selection device. Step 120: After the downlink data signal to be transmitted is generated by an arbitrary waveform generator, it is modulated onto optical carriers of different wavelengths. The modulation can adopt intensity modulation, QPSK, high-order quadrature amplitude modulation, etc. Step 130: After the modulated optical signals are amplified by an optical amplifier, at least two incoherent wavelengths are separated by a wavelength demultiplexer, and the separated optical signals are sent to polarization controllers respectively. The polarization controllers implement orthogonal polarization outputs of X polarization and Y polarization for each optical carrier, so that different optical carriers are in at least two mutually orthogonal polarization states.

[0094] More specifically, for example, when there are two wavelengths, each equipped with two orthogonal polarization states (X and Y), the transmitter forms a total of four optical signals, which can be denoted as follows: , , and The four optical signals constitute an incoherent multi-beam transmission signal with wavelength and polarization dimensions. It should be noted that the values ​​of the above-mentioned number of wavelengths and polarizations are merely examples, and this application does not limit their specific quantities. Those skilled in the art can flexibly expand the number of wavelengths and polarizations according to actual needs.

[0095] Furthermore, when the incoherent multi-beam transmitted signals spatially overlap on the far-field receiving plane, the different wavelength channels are incoherent due to their different optical frequencies, and the orthogonal polarization channels are distinguishable due to their orthogonal polarization states. Therefore, the strong phase-correlation interference term caused by traditional coherent multi-beam transmission is no longer generated during the far-field superposition process. The combined light intensity obtained at the receiving end can be considered as an incoherent superposition of the light intensities of each beam, and can be expressed as: in, Indicates far-field coordinates Synthetic light intensity observed at the location; Indicates at wavelength and polarization state The amplitude of the light field at the far-field coordinates below the channel; and These represent summations over all different wavelengths and polarization states involved in the transmission. It should be noted that, as shown in Equation 1, the incoherent multi-beam transmitted signal no longer exhibits the strong phase-correlation interference term caused by traditional coherent multi-beam transmission in the far field. This reduces the impact of coherent interference on beam splitting, channel estimation, and signal combining, and provides a physical basis for subsequent reconfigurable incoherent combining without relying on active phase synchronization.

[0096] Step 200: The non-coherent multi-beam transmission signal described in step 100 is directed by a beam pointing device, so that the non-coherent multi-beam transmission signal is transmitted downlink along the satellite-to-ground laser communication link.

[0097] Furthermore, the beam pointing device may include a voice coil motor-driven fast steering mirror array, a micromirror array antenna, or an optical scanner, or other devices capable of multi-beam pointing control. For example, this embodiment may employ a 2×2 voice coil motor-driven fast steering mirror array, which includes four independent micromirror units. Each micromirror unit is used to adjust the emission direction of its corresponding beam, allowing beams of different wavelengths and polarizations to be independently guided to the ground receiver, forming a multi-beam distribution in the far-field receiving area and achieving far-field alignment. Optionally, the incoherent multi-beam transmission signal may be collimated into a parallel beam by a two-dimensional transmitting antenna array before entering the fast steering mirror array, thereby improving the stability of multi-beam far-field superposition. Further optionally, when performing far-field alignment, the beam pointing device may also be combined with a CCD camera, a four-quadrant detector, or similar far-field monitoring devices to provide real-time feedback and closed-loop correction of the pointing position of each wavelength-polarized beam in the far field, thereby reducing the impact of pointing errors on far-field superposition efficiency.

[0098] Step 300: The ground station receiver receives the incoherent multi-beam transmitted signal after atmospheric turbulence through a multi-aperture receiver device, and sequentially performs polarization separation and wavelength separation on the optical signal received by each receiving aperture to obtain N receiving branch signals; each receiving branch signal is characterized by wavelength identifier, polarization identifier, aperture identifier and path identifier and serves as a candidate synthetic aperture unit.

[0099] Specifically, this step may include the following sub-steps. Step 310: The ground station receiver receives multiple optical signals transmitted through atmospheric turbulence using a multi-aperture receiving device. The multi-aperture receiving device may include a dual-aperture receiving antenna, a multi-aperture receiving array, or a distributed receiving array composed of multiple ground stations. The optical signal corresponding to each receiving aperture carries different aperture identifiers and path identifiers, wherein the aperture identifier is used to identify which receiving aperture receives the optical signal at the ground station receiver, and the path identifier is used to identify the transmission path traversed by the optical signal. For example, the path includes, but is not limited to, different downlinks from different satellites to the same ground station, different downlinks from the same satellite to different ground stations, or different equivalent links corresponding to the same beam passing through different atmospheric regions. Step 320: The optical signal received by each receiving aperture is spatially orthogonally polarized using a half-wave plate and a polarization beam splitter to obtain an X-polarization branch and a Y-polarization branch; optionally, the polarization separation can also be achieved using other polarization beam splitters or polarization separation devices. Step 330: Each polarization branch after polarization separation is further separated into different wavelength channels using a wavelength selection switch, optical filter, or wavelength demultiplexer, ultimately obtaining N receiving branch signals. It should be noted that this embodiment does not impose a restriction on the order of polarization separation and wavelength separation; wavelength separation can be performed first, followed by polarization separation, and the resulting receiving branch signals are physically equivalent.

[0100] For example, when a ground station uses a dual-aperture receiver, and the transmitter uses dual wavelengths, each wavelength having two orthogonal polarization states (X and Y), each receiving aperture can obtain 4 receiving branch signals, and the two receiving apertures can obtain a total of 8 receiving branch signals, i.e. It should be noted that the aforementioned The value of is determined by the number of wavelengths at the transmitting end, the number of polarizations, the number of receiving apertures, and the number of paths that the receiving end can resolve. This application does not limit . The specific value to be taken.

[0101] Step 340: Define each received branch signal obtained through the above sub-steps as a candidate synthetic aperture unit (SAP unit). Each candidate SAP unit can be composed of a quadruple. Common characteristics, among which Indicates wavelength identification, Indicates polarization marking, Indicates aperture marking, The path identifier is used to form N candidate synthetic aperture units, which are characterized by a four-dimensional identifier of wavelength, polarization, aperture, and path.

[0102] It should be noted that the "synthetic aperture" described in this application is not a synthetic receiving structure consisting solely of multiple spatially distributed receiving apertures in the traditional sense. Rather, it refers to unifying the receiving branch signals corresponding to the aforementioned quadruple as selectable aperture units for the synthetic aperture. By selecting and fusing these aperture units, the system can form different equivalent receiving aperture structures, thereby extending the synthetic aperture, which is traditionally composed solely of spatial dimensions, to a composite multi-dimensional synthetic aperture involving wavelength, polarization, aperture, and path. Furthermore, the "synthetic aperture" described in this application also differs from the "synthetic aperture" described in synthetic aperture radar. Its physical meaning lies in the fact that the receiving branch signals of multiple dimensions together constitute a set of receiving apertures composed of multi-dimensional diversity degrees of freedom in an equivalent sense, rather than synthesizing the scanning results of the same physical aperture at different times by a moving platform.

[0103] Step 400: Based on the N candidate synthetic aperture units obtained in step 300, the data processing center constructs a candidate beam combination state set; the candidate beam combination state set includes those selected from the N candidate synthetic aperture units. All combinations of paths that undergo incoherent merging, among which .

[0104] Specifically, selecting from N candidate synthetic aperture units The number of combinations corresponding to incoherent merging of paths can be expressed as: in, The coefficient of a binomial in combinatorics represents the sum of the coefficients of a number from the number of elements in a number. Selecting from candidate synthetic aperture units The number of path combinations. Based on this, the total number of candidate beam combination states. Available for all feasible The summation of the values ​​yields: in, This represents the total number of candidate synthetic aperture elements; This indicates the number of candidate synthetic aperture units participating in incoherent merging; This represents the total number of all candidate beam combination states supported by the reconfigurable wavelength-polarization synthetic aperture receiver.

[0105] For example, when At that time, we can obtain from Formula 3 That is, the candidate beam combination state set contains a total of 15 candidate combination states, covering all possible combinations from one to four beams. For example, the 15 candidate combination states include 4 single-beam reception, 6 two-beam merging, 4 three-beam merging, and 1 four-beam merging.

[0106] It should be noted that the term "reconfigurable" in this application has two meanings: first, the number of received branch signals involved in the merging process. In 1 to The optimal beam combination can be arbitrarily variable; secondly, when a certain receiving branch signal is severely affected by atmospheric turbulence, the optimal beam combination may not include that receiving branch signal, that is, it allows for the active exclusion of degraded receiving branch signals from the candidate combinations, rather than mechanically increasing the number of combining paths as the number of receiving branch signals increases. This mechanism frees the receiver from being limited to a fixed full combining method, but enables it to combine from single-path, two-path, three-path and even full beams under different turbulence conditions. The optimal combination is flexibly selected in the path, providing a complete candidate state space for the identification of the optimal beam combination in step 500 below.

[0107] Step 500: Input the N receiving branch signals obtained in step 300 into the global fusion enhanced multi-branch network. The global fusion enhanced multi-branch network identifies and outputs the optimal beam combination under the current turbulence conditions from the candidate beam combination state set described in step 400.

[0108] Specifically, this step may include the following sub-steps. Step 510: Perform digital signal processing on each received branch signal sequentially. The digital signal processing may include IQ-path orthogonalization, dispersion compensation, constant modulus algorithm, frequency offset recovery, and phase offset recovery, thereby obtaining constellation point data corresponding to each received branch signal, which serves as the first constellation point data in the global fusion enhanced multi-branch network. Road input data ,in Optionally, before the digital signal processing, the received branch signals may be subjected to operations such as optical amplification, polarization mixing, photodetection, and transimpedance amplification in sequence to convert the optical signals into electrical signals that are easy to process digitally.

[0109] Step 520, the globally fused enhanced multi-branch network includes There are 1 parallel branch. Among them, the first Each branch receives its corresponding input data. And extract independent features from each path through sub-networks; each front Each branch includes a sliding window module and the cascaded sub-network thereon, wherein the window length of the sliding window module is... The sub-network is used to extract independent features from the received branch signals of the corresponding path within the sliding window range. The independent features are used to characterize the statistical properties of the single received branch signal itself, including but not limited to constellation point distribution, power fluctuation, and error distribution.

[0110] Step 530, the first Each branch receives the weighted fused data of the N received branch signals: in, Indicates the first The data sequence of the branch receiving signal is input into the global fusion enhanced multi-branch network after digital signal processing; Indicates the first The path receives the branch signal at the first The weighting coefficients corresponding to each branch; This represents the global fused data obtained after weighted fusion of the N received branch signals; .

[0111] Furthermore, the first Each branch includes cascaded two-dimensional convolutional modules and flattening modules. The two-dimensional convolutional modules use a size of... Two-dimensional convolutional kernels are used to process the weighted fused data. Convolution operations are performed to extract joint features between the N received branch signals along the inter-branch direction. These joint features characterize the correlation between the received branch signals and the multi-path joint impairment features caused by atmospheric turbulence. The flattening module is used to convert the output of the two-dimensional convolution module into a single output. Flattened 3D feature vector, so as to facilitate the comparison with the previous The independent features output from each branch are concatenated.

[0112] Step 540, the former The independent features extracted from the branch and the first branch The joint features extracted from each branch are concatenated by a Concat layer to obtain a fused feature vector containing independent and joint features. The fused feature vector is then input into a hidden layer and an output layer. The hidden layer is used to perform nonlinear mapping of the features, and the output layer is used to identify and output the optimal beam combination under the current turbulence conditions from the set of candidate beam combination states.

[0113] Furthermore, when identifying and outputting the optimal beam combination, the global fusion enhanced multi-branch network uses criteria including the channel quality, turbulence intensity, received power, bit error rate, or Q-factor of each candidate synthetic aperture unit. The turbulence intensity can be characterized by the scintillation index of the received branch signal, which is defined as: in, Indicates the first The flicker index of the branch signal received by the path; Indicates the first The received light intensity of the branch signal; This represents the variance of the received light intensity; and These represent the time mean of the received light intensity and the time mean of the square of the received light intensity, respectively. It should be noted that a larger scintillation index indicates stronger power fluctuations in the corresponding received branch signal caused by atmospheric turbulence; conversely, a smaller index indicates weaker power fluctuations.

[0114] More specifically, the globally fused enhanced multi-branch network, through the previous Each branch simultaneously learns the independent characteristics of the received branch signals, such as the independent constellation distribution, power fluctuation characteristics, and error distribution characteristics of each branch; at the same time, through the first branch... Each branch learns the joint characteristics between the signals from multiple receiving branches, such as correlation variations between different wavelength channels, different polarization channels, different apertures, or different paths, and synchronicity differences caused by turbulence. Due to the... The two-dimensional convolutional kernels used in each branch can extract features along the inter-branch direction on the weighted fused data, so the features learned are different from those of the previous branches. The independent features learned by each branch complement each other.

[0115] For example, when a receiving branch signal experiences severe light intensity flicker or deep power fading due to atmospheric turbulence, the first... The joint features learned by each branch can identify asynchronous changes between the received branch signal and other received branch signals, thereby assisting the globally fused enhanced multi-branch network in selecting the optimal beam combination from the candidate beam combination state set that does not contain the degraded received branch signal. It should be noted that, compared to the traditional maximum ratio combining algorithm which relies solely on instantaneous channel gain for combining, the globally fused enhanced multi-branch network achieves higher combining gain by utilizing both independent and joint features. Furthermore, its advantages are more pronounced in scenarios with low received optical power, strong nonlinear noise, or high correlation among multiple received branch signals.

[0116] Optionally, the global fusion enhanced multi-branch network can be pre-trained based on received data collected offline under different turbulence intensities, so that it learns the mapping relationship between each candidate combination state and the channel quality index under different turbulence conditions, thereby enabling it to identify the optimal beam combination in real time during actual reception.

[0117] Step 600 involves incoherently combining the received branch signals corresponding to the optimal beam combination output in step 500 to recover the downlink data signal. This incoherent combining is based on intensity superposition and does not require active phase synchronization; therefore, it can adapt to time-varying phase distortions caused by satellite motion, changes in propagation distance, and dynamic changes in atmospheric channels in the satellite-to-ground downlink.

[0118] It should be noted that the optimal beam combination corresponds to the number of receiving branch signals. 1 to Any integer value in the set allows the optimal beam combination to include two, three, or more receiving branch signals, or it may exclude receiving branch signals that are severely affected by atmospheric turbulence. In other words, the reconfigurable synthetic aperture receiver described in this embodiment does not adhere to the criterion of "the more receiving branches participating in the merging, the better," but rather allows the selection of a smaller number of beam combinations without damaged branches from the candidate beam combination state set as the optimal beam combination, thereby avoiding the drag on the overall merging gain caused by severely damaged branches. For example, when the scintillation index of a certain receiving branch signal... When the Q-factor is high or correspondingly low, the global fusion enhanced multi-branch network can output a smaller number of path combinations that do not contain the signal of the receiving branch as the optimal beam combination from the candidate beam combination state set, thereby improving the overall combining performance. Optionally, the combining weights of the incoherent combining can also be output by the global fusion enhanced multi-branch network.

[0119] Furthermore, steps 500 and 600 can be run continuously online. As atmospheric turbulence changes over time, the degree of damage to the receiving branch signals corresponding to each candidate synthetic aperture unit also changes. The globally fused enhanced multi-branch network can dynamically update the optimal output beam combination, thereby realizing the real-time reconfigurable configuration of the synthetic aperture receiving structure.

[0120] Furthermore, the method described in this embodiment can be applied to various typical network transmission scenarios. For example, the method can be applied to a single-link transmission scenario from a single satellite to a single ground station, where the N candidate synthetic aperture units are mainly characterized by wavelength dimension, polarization dimension, and receiving aperture dimension. The method can also be applied to a multi-link fusion transmission scenario from multiple satellites to a single ground station, where the path dimension in the N candidate synthetic aperture units can correspond to different downlinks from different satellites to the same ground station. The method can also be applied to a multi-link network transmission scenario from multiple satellites to multiple ground stations, where the aperture dimension and path dimension in the N candidate synthetic aperture units can be jointly provided by a distributed multi-ground station receiving array. In any of the above scenarios, the method can effectively suppress light intensity flicker, power fluctuations, and communication interruptions caused by atmospheric turbulence, cloud cover, or link fading through the reconfigurable wavelength-polarization synthetic aperture receiving mechanism.

[0121] For example, in multi-link fusion transmission scenarios, due to differences in the equivalent transmission distance and atmospheric turbulence conditions encountered by different links, the degree of turbulence damage to the candidate synthetic aperture units corresponding to each link may be significantly inconsistent. In this case, the method described in this embodiment does not mechanically select the combination with the most merging links. Instead, the global fusion enhanced multi-branch network selects the current optimal beam combination from the set of candidate beam combination states. For example, it selects a combination consisting only of two or three links with good channel quality, avoiding severely damaged links from dragging down the overall merging effect.

[0122] It should be noted that there is a synergistic relationship between the steps of the method: the introduction of wavelength incoherence and polarization orthogonality in step 100 lays the physical foundation for the construction of the wavelength-polarization-aperture-path multidimensional candidate synthetic aperture unit in step 300; the multidimensional candidate synthetic aperture unit obtained in step 300 provides the source of equivalent aperture units for the construction of the candidate beam combination state set in step 400; the candidate beam combination state set constructed in step 400 provides the search space for the optimal beam combination identification of the globally fused enhanced multi-branch network in step 500; the preceding steps in step 500... The branch and the first The synergy of the branches enables the globally fused enhanced multi-branch network to simultaneously capture both single-path independent features and multi-path joint features, thereby accurately identifying the optimal beam combination under the current turbulent conditions; the incoherent combining of the selected receiving branch signals in step 600 avoids the complexity of active phase synchronization. The synergistic effect of the above steps together constitutes the complete technical chain of the reconfigurable wavelength-polarization synthetic aperture receiving anti-turbulence method described in this application.

[0123] Specific Implementation Example 2: Anti-turbulence satellite-to-ground laser communication system based on wavelength-polarization reconfigurable synthetic aperture receiver This embodiment provides an anti-turbulence star-to-ground laser communication system based on wavelength-polarization reconfigurable synthetic aperture receiver, which can be used to implement the method described in Embodiment 1. The system mainly includes a transmitting module, a beam pointing module, a multi-aperture receiving module, a wavelength-polarization separation module, a reconfigurable synthetic aperture receiver module, and a globally fused enhanced multi-branch network module, which work in sequence. Each module is described in detail below.

[0124] The transmitting module is configured to modulate satellite-to-ground downlink data signals onto at least two optical carriers of different wavelengths and perform orthogonal polarization modulation on the optical carriers to form an incoherent multi-beam transmission signal with wavelength and polarization dimensions. Specifically, the transmitting module may include at least two light sources, a modulator, an optical amplifier, a wavelength demultiplexer, and a polarization controller. The wavelength demultiplexer is used to separate at least two incoherent wavelengths, spatially separating the optical carriers of different wavelengths; the polarization controller is used to achieve orthogonal polarization output of the X and Y polarizations of each optical carrier. Optionally, the transmitting module may also include an arbitrary waveform generator for generating the downlink data signal to be transmitted.

[0125] The beam pointing module is configured to control the pointing of the incoherent multi-beam transmission signal output by the transmitting module, enabling the incoherent multi-beam transmission signal to be transmitted downlink along the satellite-to-ground laser communication link. Specifically, the beam pointing module may include at least one of a voice coil motor-driven fast steering mirror array, a micromirror array antenna, or an optical scanner. For example, the beam pointing module may employ a 2×2 voice coil motor-driven fast steering mirror array, where each micromirror element of the fast steering mirror array corresponds to a wavelength-polarized light beam, used to achieve multi-directional independent pointing of the multi-beams and complete far-field alignment. Optionally, the beam pointing module may further include a two-dimensional transmitting antenna array located in front of it for collimating the incoherent multi-beam transmission signal; further optionally, the beam pointing module may also combine a CCD camera or a four-quadrant detector to perform closed-loop feedback and correction of the far-field pointing position of each wavelength-polarized beam.

[0126] The multi-aperture receiving module is configured to receive the incoherent multi-beam transmitted signal after transmission through atmospheric turbulence and output the optical signal corresponding to each receiving aperture. Specifically, the multi-aperture receiving module may include a dual-aperture receiving antenna, a multi-aperture receiving array, or a distributed receiving array composed of multiple ground stations. It should be noted that the receiving apertures included in the multi-aperture receiving module are spatially separated from each other, and the optical signals corresponding to each receiving aperture carry different aperture identifiers and path identifiers, thus providing equivalent aperture units with spatial diversity for subsequent reconfigurable synthetic aperture reception.

[0127] The wavelength-polarization separation module is configured to sequentially perform polarization separation and wavelength separation on the optical signal output from the multi-aperture receiving module, obtaining N receiving branch signals. Each receiving branch signal is characterized by a wavelength identifier, polarization identifier, aperture identifier, and path identifier, and serves as a candidate synthetic aperture unit. Specifically, the wavelength-polarization separation module may include at least one of a half-wave plate, a polarization beam splitter, or a polarization separation device, and at least one of a wavelength selection switch, an optical filter, or a wavelength demultiplexer. More specifically, the half-wave plate, polarization beam splitter, or polarization separation device is used to achieve spatially orthogonal polarization separation on the optical signal received by each receiving aperture; the wavelength selection switch, optical filter, or wavelength demultiplexer is used to achieve separation of different wavelength channels on the polarization-separated optical signal.

[0128] The reconfigurable synthetic aperture receiver module is configured to construct a candidate beam combination state set based on the N candidate synthetic aperture units; the candidate beam combination state set includes those selected from the N candidate synthetic aperture units. All combinations of paths that undergo incoherent merging, among which The total number of candidate beam combination states is expressed by Equation 3. The reconfigurable synthetic aperture receiver module is also configured to incoherently combine the received branch signals corresponding to the optimal beam combination output by the globally fused enhanced multi-branch network module to recover the satellite-to-ground downlink data signal.

[0129] The globally integrated enhanced multi-branch network module includes One branch. (Previous) Each branch receives its corresponding receiving branch signal and extracts independent features from each path through the sub-network; Each branch receives the weighted fusion data of the N received branch signals (see Formula 4) and extracts the joint features between the N received branch signals through a two-dimensional convolution kernel; the independent features and the joint features are concatenated by a Concat layer and then input into the hidden layer and the output layer, which identify and output the optimal beam combination under the current turbulence conditions from the candidate beam combination state set and send it to the reconfigurable synthetic aperture receiving module.

[0130] Furthermore, the front end of the globally fused enhanced multi-branch network module Each branch includes a sliding window module and a sub-network, wherein the window length of the sliding window module is... The first Each branch includes a convolutional kernel size of... The two-dimensional convolutional module and the output of the two-dimensional convolutional module are... Flattened The system may also include a flattening module. Optionally, the system may further include a digital signal processing module located between the wavelength-polarization separation module and the global fusion enhanced multi-branch network module. The digital signal processing module is configured to sequentially perform IQ-path orthogonalization, dispersion compensation, constant mode algorithm, frequency offset recovery, and phase offset recovery on each received branch signal, and use the obtained constellation point data as the first-order data in the global fusion enhanced multi-branch network module. Road input data .

[0131] It should be noted that the transmitting module and beam pointing module are typically deployed on the satellite communication terminal side; the multi-aperture receiving module, wavelength-polarization separation module, reconfigurable synthetic aperture receiving module, and globally fused enhanced multi-branch network module are typically deployed on the ground station receiving end and the connected data processing center side. The data processing center can be implemented by hardware with data processing capabilities, such as general-purpose processors, dedicated processors, digital signal processors, field-programmable gate arrays, or application-specific integrated circuits. The collaborative relationship between the modules in the system described in this embodiment is consistent with the collaborative relationship between the steps in the method described in Embodiment 1: the transmitting module and the beam pointing module jointly complete the generation and downlink transmission of incoherent multi-beams; the multi-aperture receiving module and the wavelength-polarization separation module jointly complete the construction of multi-dimensional candidate synthetic aperture units; the globally fused enhanced multi-branch network module and the reconfigurable synthetic aperture receiving module jointly complete the identification and incoherent merging of the optimal beam combination. The collaborative cooperation of the modules enables the system to suppress light intensity flicker, power fluctuations, and communication interruptions caused by atmospheric turbulence in the satellite-to-ground downlink without relying on active phase synchronization.

[0132] Specific Implementation Example 3: Computer-readable storage medium This embodiment provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements steps 100 to 600 of the method described in Embodiment 1. Exemplarily, the computer-readable storage medium may be a magnetic storage medium, an optical storage medium, a semiconductor storage medium, or other non-volatile storage medium capable of storing a computer program, including but not limited to hard disks, solid-state drives, flash memory, read-only memory, random access memory, magnetic disks, or optical disks. The processor executing the computer program may be a general-purpose processor, a special-purpose processor, a digital signal processor, a field-programmable gate array, or an application-specific integrated circuit (ASIC), or other devices with data processing capabilities.

[0133] Specific Implementation Example 4: Application in Typical Experimental Verification Scenarios To further illustrate the application effect of the method and system described in this application in a real-world scenario, this embodiment provides a typical experimental verification scenario. In this scenario, a dual-wavelength, dual-polarization, dual-aperture incoherent four-beam transmission configuration can be adopted, that is, forming at the transmitting end... , , , Four incoherent multi-beam transmitted signals are processed at the ground station receiver using a dual-aperture receiver and a wavelength-polarization separation module to obtain N received branch signals; when path dimension differences are not considered... A value of 4 is possible, corresponding to the total number of candidate beam combination states. When considering multiple equivalent free-space optical links, the number of candidate synthetic aperture units can be further expanded in the path dimension.

[0134] Furthermore, in this experimental verification scenario, a turbulent phase plate can be used to simulate the wavefront distortion, beam drift, beam spread, and intensity flicker caused by atmospheric turbulence on beam propagation, thereby verifying the anti-interruption performance of the reconfigurable wavelength-polarization synthetic aperture receiving method under different turbulence intensities. For example, turbulence channels of different intensities can be generated by changing the phase plate parameters or rotation state, and the received optical power, bit error rate, and other parameters can be used as indicators. The factors and the scintillation index defined by Formula 5 are used as performance evaluation indicators to compare and verify single-beam reception, two-beam combining, three-beam combining, four-beam traditional combining, and reconfigurable combining based on the global fusion enhanced multi-branch network described in this application.

[0135] For example, under weak turbulence conditions, since the degree of turbulence damage to the signals from each receiving branch is relatively small, the optimal beam combination output by the globally fused enhanced multi-branch network from the candidate beam combination state set can include all or most of the receiving branch signals; under moderate or strong turbulence conditions, if the scintillation index of some receiving branch signals is significantly higher than that of other receiving branch signals, or the corresponding... The factor is significantly low, and the global fusion enhanced multi-branch network can select the optimal beam combination from the candidate beam combination state set that does not contain these degraded receive branch signals. That is, the number of receive branch signals participating in merging when selecting the optimal beam combination is not necessarily equal to the total number of the N candidate synthetic aperture units.

[0136] More specifically, in a multi-link fusion transmission scenario where multiple equivalent free-space optical links coexist, due to differences in the equivalent transmission distance and turbulence conditions encountered by each link, the globally fused enhanced multi-branch network may select an optimal beam combination consisting of two links with better channel quality. The communication interruption probability corresponding to this optimal beam combination may be lower than that of a combined combination consisting of three links with poorer channel quality. It should be noted that this result indicates that the optimal combining performance of the reconfigurable wavelength-polarization synthetic aperture receiver does not necessarily increase monotonically with the increase in the number of receiving branch signals involved in the combining, but rather depends simultaneously on the degree of turbulence impairment and channel quality of each receiving branch signal; this characteristic is one of the key advantages of the reconfigurable mechanism described in this application compared to a fixed combining method.

[0137] The above embodiments have the following technical effects First, by combining wavelength incoherence and polarization orthogonality, the resulting incoherent multi-beam transmission signals no longer generate strongly phase-correlated coherent interference terms when superimposed in the far field, thus avoiding dependence on active phase synchronization and real-time coherent synthesis and reducing the implementation complexity of the space-to-ground laser communication link.

[0138] Second, by unifying the received branch signals corresponding to different wavelength channels, different polarization channels, different receiving apertures, and different transmission paths as candidate synthetic aperture units, the traditional spatial dimension of synthetic aperture is extended to a composite multi-dimensional synthetic aperture of wavelength-polarization-aperture-path, enriching the diversity degrees of freedom that can participate in merging.

[0139] Third, by constructing a total of The system generates a set of candidate beam combination states, from which the global fusion enhanced multi-branch network identifies and outputs the optimal beam combination under the current turbulence conditions. This allows the system to proactively exclude certain receiving branch signals from the optimal beam combination when they are severely affected by atmospheric turbulence, thus avoiding the dragging down of the overall merging effect by severely damaged receiving branch signals and helping to reduce the probability of communication interruption in the satellite-to-ground laser communication link.

[0140] Fourth, through the aforementioned globally fused enhanced multi-branch network The branch parallel architecture enables the network to simultaneously learn the independent characteristics of each received branch signal and the joint characteristics between multiple received branch signals. Compared with the traditional maximum ratio combining method that only relies on instantaneous channel gain, it is beneficial to further improve the combining gain and reduce the bit error rate under low received power and strong turbulence conditions. Moreover, its advantages are more obvious in scenarios where the correlation between multiple received branch signals is high.

[0141] Fifth, the method and system described do not rely on complex adaptive optics hardware such as optical phased arrays, wavefront sensors, and deformable mirrors. The overall structure is relatively simple and can adapt to various typical networking transmission scenarios such as single-satellite single-station, multi-satellite single-station, and multi-satellite multi-station. It has good suppression capabilities for light intensity flicker, power fluctuations, and communication interruptions caused by atmospheric turbulence, cloud and fog weather, and link fading, which helps to improve the reliability and flexible networking capabilities of satellite-to-ground downlink laser communication links.

[0142] To facilitate understanding and implementation of this application by those skilled in the art, several optional implementation details are provided below by way of example. It should be noted that the following content is only for illustrative purposes and is not intended to limit the scope of protection of this application. Inventors may select and supplement these details according to actual engineering needs.

[0143] The wavelength spacing between each optical carrier can be selected in the range of tens to hundreds of gigahertz to ensure that different optical carriers are incoherent in optical frequency. For example, each optical carrier can operate in the C-band commonly used in optical communication, and the frequency spacing between adjacent optical carriers can be on the order of tens of gigahertz. The polarization controller can set the polarization extinction ratio of the X-polarization to the Y-polarization of each optical carrier to be no less than 20 decibels to ensure the distinguishability between orthogonal polarizations. For example, the fast-steering mirror array can be configured with 2×2 micromirror units, each with a wide optical pointing range to adapt to multi-directional independent pointing of incoherent multi-beams, and with a small tracking error to meet the accuracy requirements of incoherent multi-beam far-field alignment and superposition. The RMS value of the surface shape of the fast-steering mirror array should meet high optical wavefront requirements to reduce additional wavefront distortion. The window length of the sliding window module... The sampling rate of each received branch signal and the time scale of the features to be extracted can be selected, for example, several tens to several hundred sampling points can be selected; in an optional implementation, the window length This allows the duration covered by the sliding window to match the characteristic time of power fluctuations caused by atmospheric turbulence, thereby facilitating the effective extraction of single-path independent features by the sub-network and the joint features of multiple paths by the two-dimensional convolution module.

[0144] The subnetwork may include at least one fully connected layer and at least one nonlinear activation layer; for example, the subnetwork may adopt a feature extraction structure consisting of a one-dimensional input layer, a nonlinear activation layer, and an output layer, wherein the dimension of its output independent feature vector is the same as that output by the flattening module. The feature vectors are matched to facilitate concatenation with the joint features through a Concat layer; the activation function can be a non-linear activation function such as sigmoid, ReLU, or Leaky ReLU.

[0145] For example, the first The two-dimensional convolutional module with each branch can be used The convolution kernel performs convolution along the inter-branch direction, and one or more convolution channels can be set; the convolution result is flattened by the flattening module. Flattened The network consists of a 3D feature vector; the hidden layer can be one or more fully connected layers, and the output layer uses a linear activation function. The output is a joint feature vector of N branches, which is concatenated with the independent feature vectors and input into the final fully connected neural network. This network ultimately identifies and outputs the optimal beam combination under the current turbulence conditions from the candidate beam combination state set. The globally fused enhanced multi-branch network can use received branch signals collected under different turbulence intensities and their corresponding channel quality labels as training samples. The labels may include received power, bit error rate, and other parameters. Indicators such as factors or scintillation index are used, and the candidate beam combination state that achieves the best performance of the merged decision signal is used as the optimal combination label; the training objective function can be selected to minimize the bit error rate or Factor maximization is the loss function for the optimization objective; during the training phase, gradient descent optimizers commonly used in this field, such as Adam and SGD, and corresponding learning rate scheduling strategies can be adopted to improve the robustness of the network in identifying the optimal beam combination under different turbulent conditions.

[0146] The weighting coefficients The weighting coefficients can be initialized at the start of training by taking uniform initial values ​​based on the number of signal channels in each received branch, or by normalizing them according to the initial received power; during training, the weighting coefficients... These parameters can be used as trainable parameters for the globally fused enhanced multi-branch network and updated together with other parameters in the sub-network and the two-dimensional convolutional module through backpropagation.

[0147] Optionally, the globally fused enhanced multi-branch network can output the corresponding confidence estimate while outputting the optimal beam combination. When the confidence is lower than a preset threshold or the network output is abnormal, the reconfigurable synthetic aperture receiving module can fall back to a preset alternative merging strategy. The alternative merging strategy includes any one or more combinations of maximum ratio merging, equal gain merging, selective merging, or incoherent merging of all received branch signals. Thus, the basic availability of the anti-turbulence space-to-ground laser communication method and system can still be guaranteed even when the network is abnormal or under extreme turbulence conditions where the network is not covered.

[0148] For example, in a single-satellite single-station scenario, the path identifier can correspond to the equivalent link of the same beam passing through different atmospheric regions; in a multi-satellite single-station scenario, the path identifier can correspond to different downlinks from different satellites to the same ground station; in a multi-satellite multi-station scenario, the path identifier can correspond to different downlinks between different satellites and different ground stations, in which case the aperture identifier is used to distinguish different receiving apertures within the same ground station; further optionally, each satellite participating in transmission and each ground station participating in reception can be assigned a unique number, and the triple of "satellite number - ground station number - receiving aperture number within the ground station" can be used as the joint representation of the aperture identifier and the path identifier, so as to ensure that the candidate synthetic aperture units corresponding to different links can be distinguished one by one in the global fusion enhanced multi-branch network.

[0149] For example, IQ path orthogonalization can be achieved using the Gram-Schmidt orthogonalization algorithm; dispersion compensation can be achieved using a frequency domain equalizer; the constant modulus algorithm can be used as a blind equalizer to jointly compensate for polarization demultiplexing and residual dispersion; frequency offset recovery can be achieved using the fourth power algorithm or the periodogram method; phase offset recovery can be achieved using the Viterbi-Viterbi algorithm or the blind phase search algorithm; the specific algorithms and parameters of each of the above steps can be flexibly selected according to the modulation format and system configuration.

[0150] Optionally, the method and system can be used not only to suppress light intensity flicker caused by atmospheric turbulence, but also to suppress link fading and power fluctuations caused by factors such as cloud and fog weather, aerosol scattering, rainfall attenuation, and pointing errors; in such scenarios, the globally fused enhanced multi-branch network still focuses on received power, bit error rate, and... The optimal beam combination is identified and output based on factors and other indicators, thereby maintaining the reliability of the satellite-to-ground laser communication link under different channel conditions. It should be noted that in this patent application, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. In this patent application, if it refers to performing an action according to an element, it means performing the action at least according to that element, including two cases: performing the action only according to that element, and performing the action according to that element and other elements. Expressions such as "multiple," "repeatedly," and "various" include two, two times, two kinds, and more than two, more than two times, and more than two kinds.

[0151] All documents mentioned in this application are considered to be incorporated in their entirety into the disclosure of this application so that they can serve as a basis for modifications if necessary. Furthermore, it should be understood that after reading the foregoing disclosure of this application, those skilled in the art can make various alterations or modifications to this application, and these equivalent forms also fall within the scope of protection claimed in this application.

Claims

1. A method for anti-turbulence satellite-to-ground laser communication based on wavelength-polarization reconfigurable synthetic aperture receiver, characterized in that, Includes the following steps: S1: The satellite communication terminal modulates the downlink data signal onto at least two optical carriers of different wavelengths and performs polarization orthogonal modulation on the optical carriers to form an incoherent multi-beam transmission signal with wavelength and polarization dimensions; S2: The non-coherent multi-beam transmission signal in S1 is directed by a beam pointing device, so that the non-coherent multi-beam transmission signal is transmitted downlink along the satellite-to-ground laser communication link; S3: The ground station receiver receives the incoherent multi-beam transmitted signal described in S2 after being transmitted through atmospheric turbulence via a multi-aperture receiver. It then performs polarization separation and wavelength separation on the optical signal received by each receiving aperture in sequence to obtain N receiving branch signals. Each receiving branch signal is characterized by wavelength identifier, polarization identifier, aperture identifier, and path identifier and serves as a candidate synthetic aperture unit. S4: Based on the N candidate synthetic aperture units obtained in S3, the data processing center constructs a candidate beam combination state set. This set includes all combinations from the N candidate synthetic aperture units where m beams are selected for incoherent merging. The total number of candidate beam combination states is: in This represents the total number of candidate synthetic aperture elements. This indicates the number of candidate synthetic aperture elements participating in incoherent merging. Indicates from Choose the path The number of road combinations This represents the total number of candidate beam combination states; S5: Input the N received branch signals obtained in S3 into the global fusion enhanced multi-branch network. The global fusion enhanced multi-branch network includes N+1 branches, wherein the first N branches receive the corresponding received branch signals and extract the independent features of each path through the sub-network, and the N+1th branch receives the weighted fusion data of the N received branch signals. The joint features between the N received branch signals are extracted using a two-dimensional convolutional kernel. The independent features and the joint features are then concatenated and passed through a hidden layer and an output layer. Finally, the optimal beam combination under the current turbulence conditions is identified and output from the candidate beam combination state set described in S4. Indicates the first The path receives branch signals. Indicates the first The weighting coefficients corresponding to the branch signals received by the path. ; S6: Perform incoherent combining on the received branch signals corresponding to the optimal beam combination output by S5 to recover the downlink data signal.

2. The method according to claim 1, characterized in that, The satellite communication terminal described in S1 separates at least two incoherent wavelengths using a wavelength demultiplexer, and then achieves orthogonal polarization output of X-polarization and Y-polarization for each optical carrier using a polarization controller. The beam pointing device described in S2 includes a voice coil motor-driven fast steering mirror array, a micromirror array antenna, or an optical scanner. The beam pointing device enables multi-directional common pointing of the incoherent multi-beam transmitted signal and completes far-field alignment.

3. The method according to claim 1, characterized in that, The multi-aperture receiving device described in S3 includes a dual-aperture receiving antenna, a multi-aperture receiving array, or a distributed receiving array composed of multiple ground stations. The ground station receiving end uses a half-wave plate, a polarization beam splitter, or a polarization separation device to achieve spatial orthogonal polarization separation of the optical signal received by each receiving aperture, and uses a wavelength selection switch, an optical filter, or a wavelength demultiplexer to separate the polarization-separated optical signal into different wavelength channels, thereby obtaining the N receiving branch signals. Each receiving aperture carries a different aperture identifier and path identifier, which serve as the basis for constructing the candidate synthetic aperture unit as described in S4.

4. The method according to claim 1, characterized in that, Before inputting the N received branch signals into the global fusion enhanced multi-branch network in S5, each received branch signal is first subjected to IQ-path orthogonalization, dispersion compensation, constant modulus algorithm, frequency offset recovery, and phase offset recovery digital signal processing in sequence to obtain constellation point data as the first in the global fusion enhanced multi-branch network. Road input data ,in .

5. The method according to claim 1, characterized in that, The first N branches of the globally fused enhanced multi-branch network described in S5 each include a sliding window module and a cascaded sub-network. The window length of the sliding window module is M, and the sub-network is used to extract the independent features from the corresponding received branch signals within the sliding window range. The (N+1)th branch includes a cascaded two-dimensional convolution module and a flattening module. The kernel size of the two-dimensional convolution module is... The flattening module is used to convert the output of the two-dimensional convolution module into a flattening module. The dimensions are flattened to M; the independent features and the joint features are concatenated via a Concat layer and then input into the hidden layer and the output layer.

6. The method according to claim 1, characterized in that, When the globally fused enhanced multi-branch network described in S5 identifies and outputs the optimal beam combination from the candidate beam combination state set, the network determines the optimal beam combination based on the channel quality, turbulence intensity, received power, bit error rate, or Q factor of each candidate synthetic aperture unit; wherein the turbulence intensity is characterized by the scintillation index of the received branch signal, and the scintillation index is defined as: in Indicates the first The flicker index of the branch signal received by the path. Indicates the first The received light intensity of the branch signal in the path. This represents the variance of the received light intensity. and These represent the time mean of the received light intensity and the time mean of the square of the received light intensity, respectively. The higher the scintillation index, the stronger the power fluctuation of the corresponding receiving branch signal caused by atmospheric turbulence.

7. The method according to any one of claims 1 to 6, characterized in that, The number of receiving branch signals corresponding to the optimal beam combination 1 to Any integer value in the range allows the optimal beam combination to include two, three, or more receiving branch signals, or it may exclude receiving branch signals that are severely affected by atmospheric turbulence; when At that time, the candidate beam combination state set contains a total of The candidate combination states, and the number of received branch signals corresponding to the optimal beam combination. Taken from 1, 2, 3 or 4; the method is applied to single-link transmission scenarios from a single satellite to a single ground station, multi-link fusion transmission scenarios from multiple satellites to a single ground station, or multi-link networking transmission scenarios from multiple satellites to multiple ground stations, in order to suppress light intensity flicker, power fluctuations and communication interruptions of the satellite-to-ground laser communication link caused by atmospheric turbulence, foggy weather or link fading.

8. A turbulence-resistant space-to-ground laser communication system based on wavelength-polarization reconfigurable synthetic aperture receiver, characterized in that, include: The transmitting module is configured to modulate satellite-to-ground downlink data signals onto optical carriers of at least two different wavelengths and to perform polarization orthogonal modulation on the optical carriers to form an incoherent multi-beam transmission signal with wavelength and polarization dimensions. The beam pointing module is configured to control the pointing of the incoherent multi-beam transmission signal output by the transmitting module, so that the incoherent multi-beam transmission signal is transmitted downlink along the satellite-to-ground laser communication link. The multi-aperture receiving module is configured to receive the incoherent multi-beam transmitted signal after transmission through atmospheric turbulence, and output the optical signal corresponding to each receiving aperture. The wavelength-polarization separation module is configured to sequentially perform polarization separation and wavelength separation on the optical signal output by the multi-aperture receiving module to obtain N receiving branch signals. Each receiving branch signal is characterized by wavelength identifier, polarization identifier, aperture identifier and path identifier and serves as a candidate synthetic aperture unit. A reconfigurable synthetic aperture receiver module is configured to construct a candidate beam combination state set based on the N candidate synthetic aperture units. The candidate beam combination state set contains all combinations of selecting m beams from the N candidate synthetic aperture units for incoherent merging. The total number of candidate beam combination states is The received branch signals corresponding to the optimal beam combination output by the global fusion enhanced multi-branch network module are noncoherently combined to recover the satellite-to-ground downlink data signal. The globally fused enhanced multi-branch network module includes N+1 branches. The first N branches receive the corresponding received branch signals and extract independent features of each path through sub-networks. The N+1th branch receives the weighted fused data of the N received branch signals. The joint features between the N received branch signals are extracted using a two-dimensional convolutional kernel. The independent features and the joint features are then concatenated and passed through a hidden layer and an output layer. The optimal beam combination under the current turbulent conditions is identified and output from the candidate beam combination state set and sent to the reconfigurable synthetic aperture receiving module. Indicates the first The path receives branch signals. Indicates the first The weighting coefficients corresponding to the branch signals received by the path. .

9. The system according to claim 8, characterized in that, The transmitting module includes a wavelength demultiplexer and a polarization controller. The wavelength demultiplexer is used to separate at least two incoherent wavelengths, and the polarization controller is used to achieve orthogonal polarization output of X and Y polarization for each optical carrier. The beam pointing module includes a voice coil motor-driven fast steering mirror array, a micromirror array antenna, or an optical scanner. The multi-aperture receiving module includes a dual-aperture receiving antenna, a multi-aperture receiving array, or a distributed receiving array composed of multiple ground stations. The wavelength-polarization separation module includes a half-wave plate, a polarization beam splitter, or a polarization separation device, as well as a wavelength selection switch, an optical filter, or a wavelength demultiplexer. The first N branches of the globally fused enhanced multi-branch network module each include a sliding window module and a sub-network, and the N+1th branch includes a convolution kernel with a size of [missing information]. The two-dimensional convolutional module and the output of the two-dimensional convolutional module are... The flattening is an M-dimensional flattening module. The independent features and the joint features are concatenated by a Concat layer and then input into the hidden layer and the output layer.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method according to any one of claims 1 to 7.

Images