Inter-beam interference suppression method, electronic device, storage medium and program product

Abstract

The application provides a beam interference suppression method, an electronic device, a storage medium and a program product, and relates to the technical field of communication. The method uses prior information available in an NTN system to predict the inter-user interference strength, uses a pre-calculated interference strength representation to gather strong interference users in the same user subset for joint equalization, so that the beam interference is effectively suppressed within the subset, and different user subsets are processed independently, solving the defect that the independent beam processing scheme cannot suppress the beam interference, and improving the signal detection performance of the beam edge users. Furthermore, through subset division, the matrix dimension of joint equalization can be reduced from the total number of users to the size of each subset, the total calculation complexity is reduced, the spaceborne processor can complete real-time processing within the channel coherence time, and the bottleneck that the full-volume joint equalization scheme is not feasible in engineering is overcome.

Description

Technical Field

[0001] This application relates to the field of communication technology, and more specifically, to a method for suppressing inter-beam interference, an electronic device, a storage medium, and a program product. Background Technology

[0002] In non-terrestrial network (NTN) satellite communication systems, multi-beam antenna technology is widely used to achieve wide-area coverage and capacity enhancement. Satellites project multiple cellular coverage cells onto the ground by forming multiple narrow beams, with each beam corresponding to a ground cell, providing access services to user terminals. To maximize system throughput with limited spectrum resources, a full-frequency reuse strategy is typically used between adjacent beams, meaning all beams in the system use the exact same time-frequency resource block for data transmission. This strategy can maximize spectrum utilization efficiency, but it also introduces a serious technical problem: inter-beam interference. Specifically, in the uplink transmission direction, due to the non-ideal characteristics of the antenna pattern, uplink signals transmitted by user terminals within a beam's coverage area are not only received by the main lobe of their serving beam but also leak into adjacent receiving beams through sidelobes, causing co-channel interference to the uplink signals of other users in adjacent beams. As the user density and the number of beams increase, inter-beam interference gradually becomes the main bottleneck limiting the system's uplink capacity.

[0003] To address the aforementioned issues, two main interference suppression schemes exist in existing technologies. Scheme 1 involves joint equalization for all users, modeling the uplink signals of all active users within all beams as a unified MIMO (Multiple-Input Multiple-Output) system, and eliminating spatial interference between all users in one step by inverting the global channel matrix. Theoretically, this scheme can completely suppress inter-beam interference, but its computational complexity is as high as O(K³) (K being the total number of users), far exceeding the real-time processing capabilities of onboard processors in NTN large-coverage multi-user scenarios. Furthermore, it requires extremely high accuracy in real-time channel estimation, making it difficult to implement in engineering under long propagation delays. Scheme 2 involves independent beam processing, treating each beam as an independent link and using only the channel information of that beam for detection, completely ignoring inter-beam interference. This scheme has low computational complexity and is suitable for onboard platforms, but when users are located at the beam edge, the interference power of adjacent beams can be comparable to the useful signal power, resulting in severe degradation of detection performance and limiting the spectral efficiency gain of full-frequency reuse.

[0004] Therefore, in the existing technology, the full-scale joint equalization scheme has strong interference suppression capability but is too complex to be feasible, while the independent beam processing scheme has low complexity but is seriously insufficient in interference suppression capability. Neither of them has achieved an effective balance between interference suppression effect and on-board computational complexity. Summary of the Invention

[0005] The purpose of this application is to provide an inter-beam interference suppression method, electronic device, storage medium, and program product to improve the problem that existing suppression schemes cannot achieve an effective balance between interference suppression effect and onboard computational complexity.

[0006] In a first aspect, embodiments of this application provide an inter-beam interference suppression method applied to a spaceborne processor, the method comprising: Obtain prior information corresponding to the target satellite, including the location information of each user terminal, satellite ephemeris information, beam antenna gain pattern, and the transmit power and scheduling bandwidth of each user terminal; Determine the interference intensity characterization quantity between any two user terminals based on the prior information; The user terminals are divided into multiple user subsets according to the interference intensity characterization quantity, wherein user terminals whose interference intensity characterization quantity meets the strong interference condition are divided into the same user subset. Joint equalization processing is performed on the uplink signals of each user terminal within each user subset.

[0007] In the above implementation process, the prior information available in the NTN system is used to predict the interference intensity between users, replacing the traditional interference acquisition method that relies on real-time channel estimation, thus avoiding the problem of unreliable channel feedback information under long propagation delays. Based on the pre-calculated interference intensity characterization, users with strong interference are clustered into the same user subset for joint equalization, which effectively suppresses inter-beam interference within the subset. At the same time, different user subsets are processed independently, fundamentally solving the defect that independent beam processing schemes cannot suppress inter-beam interference, and significantly improving the signal detection performance of users at the beam edge. Furthermore, by partitioning the subsets, the matrix dimension of joint equalization can be reduced from the global number of users to the size of each subset, reducing the overall computational complexity and enabling the onboard processor to complete real-time processing within the channel coherence time, overcoming the bottleneck that the full joint equalization scheme is not feasible in engineering. Therefore, this scheme achieves an effective balance between onboard computing power constraints and interference suppression performance.

[0008] Secondly, embodiments of this application provide an inter-beam interference suppression device applied to a spaceborne processor, the device comprising: The information acquisition module is used to acquire prior information corresponding to the target satellite. The prior information includes the location information of each user terminal, satellite ephemeris information, beam antenna gain pattern, and the transmit power and scheduling bandwidth of each user terminal. An interference determination module is used to determine the interference intensity characterization quantity between any two user terminals based on the prior information. The subset partitioning module is used to divide the user terminal into multiple user subsets according to the interference intensity characterization quantity, wherein user terminals whose interference intensity characterization quantity satisfies the strong interference condition are divided into the same user subset. The equalization processing module is used to perform joint equalization processing on the uplink signals of each user terminal within each user subset.

[0009] Thirdly, embodiments of this application provide an electronic device, including a processor and a memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the steps of the method provided in the first aspect above are performed.

[0010] Fourthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the steps of the method provided in the first aspect above.

[0011] Fifthly, embodiments of this application provide a computer program product, including computer program instructions, which, when read and executed by a processor, perform the steps of the method provided in the first aspect above.

[0012] Other features and advantages of this application will be set forth in the following description and will be apparent in part from the description or may be learned by practicing embodiments of this application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings. Attached Figure Description

[0013] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 A flowchart illustrating an inter-beam interference suppression method provided in this application embodiment; Figure 2 A structural block diagram of an inter-beam interference suppression device provided in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of an electronic device for performing an inter-beam interference suppression method, provided as an embodiment of this application. Detailed Implementation

[0015] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0016] It should be noted that the terms "system" and "network" in the embodiments of this invention can be used interchangeably. "Multiple" refers to two or more; therefore, in the embodiments of this invention, "multiple" can also be understood as "at least two". "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / ", unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship.

[0017] It should also be noted that all actions involving the acquisition of signals, information, or data in this application are carried out in compliance with the relevant data protection laws and policies of the country where the application is located, and with the authorization granted by the owner of the relevant device.

[0018] This application provides an inter-beam interference suppression method. This method utilizes prior information available in the NTN system to predict the inter-user interference intensity, replacing the traditional interference acquisition method that relies on real-time channel estimation, thus avoiding the problem of unreliable channel feedback information under long propagation delays. Based on pre-calculated interference intensity characteristics, strong interfering users are clustered into the same user subset for joint equalization, effectively suppressing inter-beam interference within the subset. Simultaneously, different user subsets are processed independently, fundamentally solving the deficiency of independent beam processing schemes in suppressing inter-beam interference, significantly improving the signal detection performance of users at the beam edge. Furthermore, through subset partitioning, the matrix dimension of joint equalization can be reduced from the global user count K to the size of each subset. K l The total computational complexity is from O(K³) Reduced to Σ O(K l ³) This enables the onboard processor to complete real-time processing within the channel coherence time, overcoming the bottleneck of the full joint equalization scheme being impractical in engineering. Therefore, this scheme achieves an effective balance between onboard computing power constraints and interference suppression performance.

[0019] Please refer to Figure 1 , Figure 1 A flowchart of an inter-beam interference suppression method provided in this application embodiment, the method being applied to a spaceborne processor, includes the following steps: Step S110: Obtain prior information corresponding to the target satellite.

[0020] The prior information includes the location information of each user terminal, satellite ephemeris information, beam antenna gain pattern, and the transmit power and scheduling bandwidth of each user terminal. This prior information does not need to be obtained through real-time channel estimation or feedback; it can be obtained in advance through system configuration or broadcast or reporting mechanisms. Therefore, this scheme can pre-calculate the interference intensity based on the prior information. In this scheme, the user terminal refers to the active user terminal served by the target satellite. An active user terminal can be understood as a user terminal that has been allocated uplink transmission resources by the system scheduler within the current scheduling cycle or transmission time interval and is currently or will soon be in a data transmission state. The target satellite can be any satellite.

[0021] Prior information can be synchronously updated by the onboard scheduler before the start of each scheduling cycle, ensuring that its timeliness is better than the channel coherence time, thereby guaranteeing the effectiveness of the interference intensity characterization.

[0022] Specifically, the location information of each user terminal can be acquired in real time by the Global Navigation Satellite System (GNSS) receiver equipped on the user terminal itself, generating three-dimensional positioning data containing longitude, latitude, and altitude. This data is then reported to the satellite's onboard processor via the uplink control channel at a preset period. After receiving the data, the satellite's onboard processor converts the longitude, latitude, and altitude coordinates into a three-dimensional position vector in the Earth-Centered Earth-Fixed (ECEF) coordinate system, so as to facilitate subsequent spatial geometric calculations with the satellite position.

[0023] The satellite ephemeris information is directly read from the satellite platform by the onboard processor of the target satellite. The satellite platform continuously updates the satellite's three-dimensional position vector and velocity vector in the ECEF coordinate system, and provides attitude angle data (including yaw angle, pitch angle and roll angle) of the satellite body coordinate system relative to the ECEF coordinate system. This information ensures the accuracy of subsequent beam pointing calculations.

[0024] A beam antenna gain pattern is a directional function describing the antenna's ability to radiate or receive signals in different directions in space. It is typically represented as a three-dimensional surface plot or a two-dimensional polar coordinate plot, with the horizontal axis representing the azimuth angle and the vertical axis representing the antenna gain value (in dBi or linear ratio). The main lobe of the beam has the highest gain, while the side lobes show a significant decrease in gain. The beam antenna gain pattern can be pre-programmed into a lookup table (LUT) in the onboard storage unit. This LUT has been calibrated through anechoic chamber testing before satellite launch and stores the gain values ​​of each receiving beam under different combinations of elevation and azimuth angles in a two-dimensional discrete array. The onboard processor can retrieve and query this data whenever needed.

[0025] The transmit power and scheduling bandwidth of each user terminal are determined by the satellite-side scheduler when allocating resources in each scheduling cycle. The transmit power is configured through the uplink power control algorithm, and the scheduling bandwidth is dynamically allocated to each user according to the user's service requirements and channel quality. The above power and bandwidth parameters are known output information of the satellite scheduler, and the onboard processor can directly obtain them from the scheduler's resource allocation records.

[0026] Step S120: Determine the interference intensity characterization quantity between any two user terminals based on prior information.

[0027] Among them, the interference intensity characterization measure can be used to quantify the degree of mutual interference between any two user terminals.

[0028] In some implementations, the interference intensity between any two user terminals can be characterized by the channel spatial correlation coefficient. For example, the onboard processor first calculates the azimuth angle (including elevation and azimuth) of each user terminal relative to the target satellite based on the location information and satellite ephemeris information. Then, it uses the azimuth angle of each user terminal to query the beam antenna gain pattern and obtains the antenna gain vector from each user terminal to its serving beam and to adjacent beams. This antenna gain vector can be regarded as the spatial feature vector of the user terminal in the multi-beam space. Based on this, the onboard processor calculates the correlation coefficient between the spatial feature vectors of any two user terminals, such as using cosine similarity or Pearson correlation coefficient, and uses this correlation coefficient as the interference intensity characterization between them. The spatial correlation coefficient typically ranges from 0 to 1. The closer the correlation coefficient is to 1, the lower the spatial distinguishability of the two user terminals and the stronger the inter-beam interference between them; the closer the correlation coefficient is to 0, the higher the spatial isolation of the two user terminals and the weaker the interference. In this case, the larger the interference intensity characterization, the greater the interference intensity it represents.

[0029] In other embodiments, the interference intensity can be represented by the signal-to-interference ratio (SIR). In this case, the larger the interference intensity is, the smaller the interference intensity. The specific calculation method of SIR will be explained in detail in subsequent embodiments, and will not be elaborated here.

[0030] Step S130: Divide the user terminal into multiple user subsets according to the interference intensity characterization quantity.

[0031] Among them, user terminals whose interference intensity, as represented by the interference intensity characterization quantity, meets the strong interference condition are divided into the same user subset.

[0032] Strong interference conditions are used to determine whether the interference intensity between two user terminals reaches the threshold condition requiring joint equalization processing. For example, user terminals whose interference intensity, as represented by the interference intensity characterization quantity, is greater than or equal to a preset threshold are classified into the same user subset.

[0033] A preset threshold is used to measure the strength of inter-beam interference between any two user terminals. The value of this threshold determines the granularity of the user subset division. For example, the lower the threshold value, the more user terminals are judged as having strong interference, and the stronger the interference suppression capability of joint equalization, but the computational complexity also increases accordingly; conversely, the higher the threshold value, the greater the interference. The quantization of the preset threshold varies depending on the quantization method of the interference strength representation. For details, please refer to the relevant descriptions in subsequent embodiments. For example, when the interference strength representation is the channel spatial correlation coefficient, the preset threshold is also in the dimension of the coefficient; when the interference strength representation is the SIR, the preset threshold is also in the dimension of the SIR. It should be noted that the specific mathematical expression of the strong interference condition depends on the choice of the interference strength representation: when the representation is positively correlated with the interference strength (such as the correlation coefficient), the strong interference condition is that the interference strength representation ≥ the preset threshold; when the representation is negatively correlated with the interference strength (such as the SIR), the strong interference condition is that the interference strength representation ≤ the preset threshold.

[0034] In actual engineering deployments, this preset threshold can be dynamically configured based on the real-time computing power constraints of the onboard processor and the system performance requirements. Alternatively, a fixed value can be set according to actual needs.

[0035] A user subset refers to a user group obtained by the onboard processor after dividing all user terminals into groups based on the interference intensity between them. Multiple user subsets may be included. Within any two user terminals in one subset, the interference intensity represented by the interference intensity characterization quantity satisfies the strong interference condition. However, the interference intensity between any two user terminals belonging to different user subsets does not satisfy the strong interference condition, indicating that they are at a weak interference level, and independent processing will not significantly affect the detection performance.

[0036] In practical implementation, after calculating the interference strength representation for every two user terminals, the interference strength represented by the interference strength representation is compared with the strong interference condition. For example, if the interference strength representation is the channel spatial correlation coefficient, then the channel spatial correlation coefficient is used to represent the interference strength; that is, the larger the coefficient, the greater the interference strength. The strong interference condition can be that the channel spatial correlation coefficient is greater than or equal to a preset threshold, for example, the preset threshold can be set to 0.7 as needed. If the channel spatial correlation coefficient is greater than or equal to 0.7, then the user terminals are assigned to the same user subset. Finally, if the channel spatial correlation coefficient between any two user terminals in this user subset is greater than or equal to 0.7, then it meets the strong interference condition. The remaining user terminals can be assigned to other user subsets, which may contain user terminal pairs with a channel spatial correlation coefficient less than 0.7.

[0037] Step S140: Perform joint equalization processing on the uplink signals of each user terminal within each user subset.

[0038] Joint equalization refers to an equalization method that treats the uplink signals of multiple user terminals as a whole MIMO system for joint detection. In this scheme, joint equalization is performed within a subset of users, and spatial interference between users within the subset is eliminated by inverting the channel submatrix.

[0039] In some implementations, during joint equalization processing, for each user subset, the channel gain from each user terminal in the user subset to each associated beam is extracted to construct the channel sub-matrix of the user subset; based on the channel sub-matrix, zero-forcing equalization or minimum mean square error equalization is used to calculate the equalization matrix; the equalization matrix is ​​applied to the received signal vector of the user subset to obtain the estimated value of the uplink signal of each user terminal in the user subset.

[0040] For example, for each user subset obtained by partitioning (l=1,2,...,L), where L represents the total number of user subsets, and the following processing is performed: The onboard processor first needs to construct the channel sub-matrix for this subset. Specifically, the user subset... Including K l For each user terminal, the onboard processor determines the beam set associated with that subset based on the service beam information of each user terminal within that subset. (including | (each beam), then extract this K l One user terminal here | The channel gains of each beam are arranged in a manner that prioritizes "beams as users" to form a... | | OK K l Channel submatrix of columns , For | |× A dimensional matrix. Each element in this channel submatrix. This represents the channel gain between the k-th user terminal and the m-th receiving beam. This channel gain can be directly constructed from the geometric parameters (antenna gain, path loss, propagation phase) in the aforementioned prior information, without relying on pilot estimation.

[0041] In obtaining the channel submatrix Then, the onboard processor determines the number of user terminals within that subset. Number of associated beams | | Based on the relationship, select an appropriate equilibrium algorithm to calculate the equilibrium matrix. W l .when ≤ | | When the number of users does not exceed the number of associated beams and the channel submatrix is ​​of full rank, zero-forcing (ZF) equalization can be used, and the equalization matrix is ​​arranged according to... calculate, express The conjugate transpose of the matrix completely eliminates interference between users within a subset but enhances noise; alternatively, minimum mean square error (MMSE) equalization can be used, with the equalization matrix following the formula... Calculation, where Let I be the noise power and I be the identity matrix. express The conjugate transpose of the method achieves a trade-off between interference cancellation and noise enhancement to minimize mean square error.

[0042] When using ZF equalization When the matrix is ​​singular (i.e., the number of users exceeds the number of associated beams), MMSE equalization needs to be used to ensure matrix invertibility, or a regularized ZF method should be employed. This is done when calculating the equalization matrix. W l The process involves matrix inversion operations. The onboard processor can use Cholesky decomposition or Singular Value Decomposition (SVD) for numerical solutions to improve the numerical stability and computational efficiency of the operation.

[0043] Obtain the equilibrium matrixW l Subsequently, the onboard processor extracts the received signal vector corresponding to the subset from the frequency domain received data of each associated beam. y l ( y l For | | A 3D column vector, where each element corresponds to the received signal of a beam on the target time-frequency resource), and then an equalization matrix is ​​formed. With the received signal vector y l Multiplication, that is ,get dimensional estimated signal vector Each element corresponds to an estimated value of the uplink signal transmitted by a user terminal. At this point, signal detection for each user terminal within this subset is complete.

[0044] The onboard processor independently executes the above processing flow for each of the L subsets obtained in step S130. The balanced operation of each subset does not affect each other. It can be executed serially or concurrently on different parallel pipelines of the onboard multi-core processor or field-programmable gate array (FPGA) to further shorten the processing latency.

[0045] It should be noted that the channel submatrix The construction can be performed without relying on real-time pilot channel estimation. In one implementation, the elements of the channel submatrix can be constructed according to... Calculation, where The antenna gain of user terminal k in the direction of beam m is obtained by querying the beam antenna gain pattern from the azimuth angle. Let k be the free space path loss for user terminal k. Let λ be the Euclidean distance from the target satellite to the user terminal k, and λ be the signal wavelength. This geometry construction method utilizes the predictable satellite orbits and known user locations in NTN scenarios, avoiding the feedback delay problem caused by pilot estimation in traditional terrestrial cellular networks.

[0046] The above method, by constructing a channel submatrix within a user subset and performing ZF or MMSE equalization, enables the precise elimination of spatial interference between strongly interfering users within the subset through matrix inversion. This effectively suppresses the impact of inter-beam interference on signal detection and overcomes the performance degradation problem caused by treating strong interference as noise in independent beam processing. Furthermore, since joint equalization is performed independently only within each subset, different subsets do not interfere with each other, reducing the dimension of matrix inversion from the global user count K to the number of users within each subset. K lThe computational load of a single balancing operation is significantly reduced, and the balancing operations of each subset can be executed in parallel, further shortening the overall processing latency. This enables the onboard processor to complete real-time detection under limited computing resources and strict latency constraints.

[0047] In the above implementation process, the prior information available in the NTN system is used to predict the interference intensity between users, replacing the traditional interference acquisition method that relies on real-time channel estimation, thus avoiding the problem of unreliable channel feedback information under long propagation delays. Based on the pre-calculated interference intensity characterization, users with strong interference are clustered into the same user subset for joint equalization, which effectively suppresses inter-beam interference within the subset. At the same time, different user subsets are processed independently, fundamentally solving the defect that independent beam processing schemes cannot suppress inter-beam interference, and significantly improving the signal detection performance of users at the beam edge. Furthermore, by partitioning the subsets, the matrix dimension of joint equalization can be reduced from the global number of users to the size of each subset, reducing the overall computational complexity and enabling the onboard processor to complete real-time processing within the channel coherence time, overcoming the bottleneck that the full joint equalization scheme is not feasible in engineering. Therefore, this scheme achieves an effective balance between onboard computing power constraints and interference suppression performance.

[0048] Based on the above embodiments, in the method of determining the interference intensity characterization between any two user terminals according to prior information, it is also possible to calculate the distance from each user terminal to the target satellite and the azimuth angle of each user terminal relative to the target satellite based on the location information of each user terminal and the satellite ephemeris information; calculate the free space path loss of each user terminal based on the distance and the system carrier frequency; obtain the first antenna gain of each user terminal on its serving beam and the second antenna gain on the adjacent beam based on the azimuth angle and the beam antenna gain pattern; calculate the useful signal received power spectral density of each user terminal and the interference signal received power spectral density of each user terminal to other user terminals based on the free space path loss, first antenna gain, second antenna gain, transmit power and scheduling bandwidth of each user terminal; and determine the interference intensity characterization between any two user terminals based on the useful signal received power spectral density and the interference signal received power spectral density.

[0049] The onboard processor first calculates the distance from each user terminal to the target satellite and the orientation angle of each user terminal relative to the target satellite, based on the location information of each user terminal and the satellite ephemeris information. Specifically, for any user terminal k, the onboard processor calculates its position vector in the ECEF coordinate system. Position vector of target satellite s Subtraction yields the relative vector , This represents the x-axis coordinate of user terminal k in the ECEF coordinate system. This represents the y-coordinate of user terminal k in the ECEF coordinate system. This represents the z-axis coordinate of user terminal k in the ECEF coordinate system; This represents the x-axis coordinate of the target satellite s in the ECEF coordinate system. This represents the y-axis coordinate of the target satellite s in the ECEF coordinate system. This represents the z-axis coordinate of the target satellite s in the ECEF coordinate system. express and The deviation between them express and The deviation between them express and The deviation between them. Then, the Euclidean norm of this relative vector is calculated, which gives the distance from user terminal k to target satellite s. Simultaneously, based on the coordinate components of this relative vector, the orientation angle of the user terminal k relative to the satellite's local coordinate system, including the elevation angle, is calculated using inverse trigonometric functions. and azimuth .

[0050] After obtaining the distance from each user terminal to the satellite Subsequently, the onboard processor, based on this distance and the system carrier frequency... f c Calculate the free space path loss for each user terminal. The calculation formula is: , where c is the speed of light. Free-space path loss reflects the energy attenuation of a signal during propagation in ideal free space due to increased distance and beam spread. Its value is determined by both distance and carrier frequency; the greater the distance or the higher the frequency, the greater the path loss.

[0051] After obtaining the azimuth angle of each user terminal relative to the satellite, the onboard processor uses that azimuth angle ( , Using the index , query the pre-stored beam antenna gain pattern to obtain the first antenna gain of each user terminal on its serving beam. and the second antenna gain on adjacent beams The first antenna gain refers to the gain from the user terminal k to its serving beam. The antenna gain in the direction of beam direction is typically located within the main lobe of the beam and is relatively high. The second antenna gain refers to the gain from the same user terminal k to a certain adjacent beam. The antenna gain value in the direction of the beam is usually located at the edge of the main lobe or in the side lobe region, and the value is relatively low.

[0052] In some implementations, the beam antenna gain pattern is stored in the onboard storage unit in the form of a look-up table (LUT). The look-up table stores the antenna gain of each receiving beam in the corresponding direction, indexed by the azimuth angle. In the above method of obtaining the first antenna gain of each user terminal on its serving beam and the second antenna gain on the adjacent beam, the first antenna gain of each user terminal on its serving beam and the second antenna gain on the adjacent beam can be found in the look-up table according to the azimuth angle. If the azimuth angle does not fall on a discrete grid point of the look-up table, the antenna gain in that direction is calculated using a bilinear interpolation algorithm to obtain the first antenna gain of each user terminal on its serving beam and the second antenna gain on the adjacent beam.

[0053] The lookup table (LUT) is a data structure that stores discrete data points in tabular form, enabling fast lookups without real-time computation. In this scheme, the LUT stores the antenna gain values ​​of each received beam in various spatial directions, with rows and columns corresponding to discrete sampling points for elevation and azimuth angles, respectively, and each entry representing an antenna gain value.

[0054] Discrete grid points are a finite number of discrete azimuth angle sampling points in a lookup table. Due to limited storage space, LUTs cannot record antenna gain values ​​in all continuous directions in space. Instead, they can only discretize the space in certain steps (e.g., 1° elevation angle × 1° azimuth angle), and these sampling points are called grid points.

[0055] The onboard processor obtains the direction angles (elevation angle θ and azimuth angle) of each user terminal relative to the satellite. Then, the pre-stored beam antenna gain pattern (LUT) is retrieved using that azimuth angle as an index. This lookup table was calibrated via ground testing before satellite launch and is stored in a two-dimensional table format. The row index corresponds to the discrete sampling points of the elevation angle θ, and the column index corresponds to the azimuth angle. The discrete sampling points are represented by an entry for the antenna gain value in the corresponding direction (usually stored in dBi or as a linear ratio). For each receiving beam m of the target satellite, a corresponding lookup table exists, recording the gain response of that beam in various spatial directions. Therefore, the onboard processor needs to determine the gain response based on the azimuth angle of the user terminal k. The service beams at user terminal k are respectively m k The first antenna gain is retrieved from the corresponding lookup table. and in each adjacent beam m j The second antenna gain was retrieved from the corresponding lookup table. .

[0056] During the query process, if the user terminal's azimuth angle falls exactly on a discrete grid point in the lookup table, the onboard processor directly reads the gain value stored at that grid point as the query result. However, since user terminals are continuously distributed within the beam coverage area, their azimuth angles are usually not exactly equal to the discrete sampled values ​​in the lookup table. In this case, the azimuth angle will fall between four adjacent grid points in the lookup table. Taking the pitch angle dimension as an example, suppose the user terminal's pitch angle θ is located at a discrete sampled point... and Between, azimuth angle Located at discrete sampling points and Between these points, the onboard processor reads the gain values ​​corresponding to the four grid points: , , , Then, the antenna gain value in that direction is calculated using a bilinear interpolation algorithm.

[0057] The basic idea of ​​bilinear interpolation is to perform linear interpolation in two dimensions. The specific calculation process is as follows: First, perform two linear interpolations in the pitch angle dimension to obtain the azimuth angle. and Interpolation results for the target pitch angle θ in the direction:

[0058] Then, linear interpolation is performed along the azimuth dimension to obtain the target direction angle. Gain value at:

[0059] The interpolation order of the two dimensions can be interchanged; that is, interpolation can be performed first in the azimuth dimension and then in the elevation dimension, with the same final result. Through the above bilinear interpolation calculation, the onboard processor can obtain the first antenna gain of the user terminal on its serving beam. and in adjacent beams m j The second line gain Furthermore, this gain value is a smooth estimate over continuous angles, avoiding quantization errors caused by discretization grids.

[0060] It should be noted that the antenna gain values ​​stored in the lookup table can be in either linear ratio or dB value form. If stored in dB value form, bilinear interpolation should be performed in the dB domain; if stored in linear ratio form, interpolation should be performed in the linear domain. In practical engineering, due to the significant gain difference between the main lobe and side lobe regions of the antenna gain pattern, dB domain interpolation usually yields smoother interpolation results. Therefore, it is preferable to use a combination of dB value storage and dB domain linear interpolation, converting the interpolated values ​​to linear ratios for subsequent power spectral density calculations.

[0061] Furthermore, the azimuth angle exhibits periodicity at 0° and 360°, and the azimuth angle of the target approaches the boundary (e.g., When approaching 0° or 360°, the onboard processor needs to perform orbital processing at the boundaries. For example, if =358° =360°, and the target =359.5°, then interpolation can be performed normally using the formula above; if the target If the azimuth angle is 0.5°, then interpolation needs to be performed using two grid points: 358° and 360° (equivalent to 0°), or using two grid points: 0° and 2°, depending on the boundary definition of the lookup table. Before interpolation, the onboard processor should normalize the azimuth angle to ensure it falls within the valid range defined by the lookup table (usually 0°~360°), thus ensuring the correctness of the interpolation boundary conditions.

[0062] By storing the beam antenna gain pattern in the form of a lookup table (LUT) in the onboard storage unit, the onboard processor does not need to calculate complex antenna pattern functions in real time. It can quickly obtain the antenna gain values ​​of each user on the serving beam and adjacent beams simply by looking up the table, significantly reducing the online computational overhead and storage resource consumption of the onboard processor. Furthermore, considering the actual situation where the user's azimuth angle may not fall on the discrete grid points of the LUT, a bilinear interpolation algorithm is used to perform two-dimensional linear interpolation using the gain values ​​of the nearest neighbor grid points around the target point. This provides a smooth and accurate antenna gain estimate over continuous angles, avoiding the quantization error caused by directly taking the nearest grid point, thus ensuring the accuracy of subsequent interference intensity characterization calculations.

[0063] After obtaining the free-space path loss, first antenna gain, second antenna gain, and transmit power and scheduling bandwidth from the prior information for each user terminal, the onboard processor calculates the useful signal received power spectral density for each user terminal and the interference signal received power spectral density of each user terminal to other user terminals. Specifically, for user terminal i, its uplink signal is affected by the satellite's serving beam. m i During reception, the received power spectral density of the useful signal is:

[0064] in, This represents the transmit power of user terminal i. This represents the first antenna gain of user terminal i. This represents the scheduling bandwidth of user terminal i. This represents the free-space path loss of user terminal i. The received power spectral density of the useful signal. This reflects the received strength of the useful signal from user i per unit frequency. For the interference generated by user terminal j on user terminal i, its received power spectral density is:

[0065] in, This represents the transmit power of user terminal j. This represents the second antenna gain of user terminal i. This represents the scheduling bandwidth of user terminal j. This represents the free-space path loss of user terminal j. The received power spectral density of the interference signal. This reflects the interference signal of user terminal j in the serving beam of user terminal i. m i The received signal strength within a unit frequency range.

[0066] Finally, the onboard processor determines the interference intensity characterization quantity between any two user terminals based on the aforementioned received power spectral density of useful signals and received power spectral density of interference signals.

[0067] In the above implementation process, by utilizing prior information such as satellite ephemeris, user location, antenna pattern, transmit power and scheduling bandwidth, the power spectral density of the useful signal and interference signal of each user terminal is pre-calculated directly on the on-board processor, and the interference intensity characterization quantity between any pair of user terminals is determined accordingly. This eliminates the traditional interference acquisition method that relies on real-time pilot channel estimation and avoids the problem of outdated and unreliable channel state information feedback under long propagation delay in NTN scenarios.

[0068] Based on the above embodiments, in the implementation method of determining the interference intensity characterization quantity between any two user terminals according to the useful signal received power spectral density and the interference signal received power spectral density, the ratio of the useful signal received power spectral density of user terminal i to the interference signal received power spectral density of user terminal i to user terminal j can be calculated to obtain the first interference intensity characterization quantity of user terminal i to user terminal j, and the ratio of the useful signal received power spectral density of user terminal j to the interference signal received power spectral density of user terminal j to user terminal i can be calculated to obtain the second interference intensity characterization quantity of user terminal j to user terminal i. Then, based on the first interference intensity characterization quantity and the second interference intensity characterization quantity, the interference intensity characterization quantity between user terminal i and user terminal j can be determined.

[0069] For example, for any user terminal i and user terminal j, the interference between them is mutual. The SIR values ​​in the two directions can be calculated separately. The SIR value can characterize the ratio of useful signal power to interference signal power. In this scheme, the larger the SIR, the weaker the interference between users, and the smaller the SIR, the stronger the interference between users.

[0070] The first disturbance intensity characterization quantity is:

[0071] in, This represents the first antenna gain of user terminal j. This represents the gain of the second antenna of user terminal j.

[0072] The second interference intensity characterization quantity is:

[0073] After obtaining the first interference intensity characterization value and the second interference intensity characterization value, in some embodiments, the smaller value between the first interference intensity characterization value and the second interference intensity characterization value can be determined as the interference intensity characterization value between user terminal i and user terminal j, that is... .

[0074] The significance of using a smaller value is that the direction with a lower SIR represents the situation where the interference between user terminal pairs is most severe. Using this value as the basis for subsequent user subset division can ensure that strongly interfering user terminal pairs are identified and classified into the same user subset for joint processing in any direction.

[0075] The onboard processor repeats the above calculation process for all active user terminals, ultimately constructing the complete SIR matrix S, where... S(i,j)=SIR(i,j) diagonal elements S(i,i)Let the matrix be positive infinity (indicating that the user terminal itself has no interference). The rows and columns of this matrix correspond to all user terminals in the system, and each element represents two user terminals. (i,j) The intensity of interference between them.

[0076] In an alternative implementation, after constructing the SIR matrix, post-processing can be performed to optimize the partitioning criteria. Specifically, the onboard processor can apply a preset correction factor to the original SIR matrix, such as uniformly subtracting a bias (or multiplying by a scaling factor less than 1) from each SIR value, thereby tending to form a larger user subset and improving the redundancy of interference suppression; conversely, it can also uniformly add a bias (or multiply by a scaling factor greater than 1), making the threshold relatively lower, thereby tending to form a smaller user subset and reducing computational complexity. This post-processing method allows the onboard processor to adapt to dynamically changing system load and computing power constraints by adjusting the SIR matrix itself without changing the threshold settings.

[0077] In the above implementation process, by calculating the SIR values ​​in the two directions respectively and determining the interference intensity characterization quantity between user terminal pairs accordingly, it is possible to accurately capture the bidirectional interference asymmetry caused by factors such as differences in transmit power, differences in scheduling bandwidth, and spatial asymmetry of antenna gain, and avoid the misjudgment of interference intensity that may be caused by traditional single-direction estimation methods.

[0078] However, in actual deployments, depending on system performance requirements and application scenarios, those skilled in the art can also use other aggregation strategies to determine paired SIRs, such as: In an alternative implementation, the larger of the two directional SIRs can be used as the pairwise SIR between user pairs, i.e. This method means that the direction with weaker interference (higher SIR) is used as the measure of interference intensity for the user terminal pair. In actual communication, the mutual interference between two user terminals is not always equal. When the SIR in one direction is already sufficiently high, for example, if user terminal j has a lower transmit power or its signal to beam m... i The sidelobe gain is extremely small. Even if user terminal i causes strong interference to user terminal j, user terminal j may not cause the same level of interference to user terminal i. In this case, using a larger value as the pairwise SIR can avoid "collateral damage" to the user terminal pair due to extremely weak interference in one direction. This allows the user terminal pairs with originally asymmetrical interference to be kept in different subsets and processed independently, thereby further reducing computational complexity.

[0079] In another alternative implementation, the arithmetic mean of the two-way SIR can be used as the pairwise SIR between user terminal pairs, i.e. This approach comprehensively considers interference from both directions, avoiding the extremes of being overly conservative when taking a smaller value or overly aggressive when taking a larger value; it represents a compromise strategy. When the SIR difference between the two directions is significant, the average value can smooth out this difference, resulting in a paired SIR that falls between the two. The advantage of this approach is a more balanced assessment of interference, but the disadvantage is that the average value may mask strong interference in one direction. That is, when the SIR in one direction is extremely low while the SIR in the other direction is extremely high, the average value may still be high, potentially leading to users with strong interference being incorrectly assigned to different subsets. To overcome this drawback, a weighted average approach can be used, for example, assigning different weights to the SIR in the two directions based on the transmit power or scheduling bandwidth of each user terminal. Alternatively, the geometric mean of the SIRs in the two directions can be used as the paired SIR.

[0080] In an alternative implementation, the aggregation strategy can be dynamically selected based on the system's current operating state. Specifically, when the system is in an interference-limited region (such as when inter-beam interference is the main factor limiting performance), the onboard processor adopts a conservative strategy that takes the smaller value to ensure that all user terminals with strong interference are identified and grouped into the same user subset for joint processing. When the system is in a noise-limited region (i.e., where the interference impact is relatively small), the onboard processor switches to an aggressive strategy that takes the larger value to reduce the subset size and computational complexity. When the system is in a transitional region between the two, an average value strategy is used for balance. This dynamic switching strategy allows the onboard processor to flexibly adjust the interference representation method according to the real-time link quality, achieving an optimal balance between performance and complexity.

[0081] In another alternative implementation, instead of aggregating the SIR values ​​in both directions into a single scalar value, the values ​​in both directions can be treated as a two-dimensional vector. The subsequent user subset partitioning steps are directly input. During partitioning, the onboard processor sets thresholds for both directions. Only when the SIR (Sequential Intrusion Rate) in both directions is less than or equal to the preset threshold (i.e., the strong interference condition is met) are two users assigned to the same subset. If the SIR in only one direction is less than or equal to the threshold, the two users are allowed to be in different subsets. This method preserves all information about bidirectional interference and avoids information loss during aggregation. However, its cost is that the subset partitioning algorithm needs to adapt to two-dimensional criteria, increasing the complexity of the partitioning logic.

[0082] In another alternative implementation, the sum of two-way SIRs can be used as an alternative measure of paired SIRs, i.e. Alternatively, its reciprocal form (i.e., the sum of interferences) can be used as a measure of interference intensity. The physical meaning of this method is that the total interference energy between two user terminals is contributed by interference from both directions; the greater the total interference, the more necessary it is to group both terminals into the same user subset for joint processing. The key difference between this method and taking the smaller value is that taking the smaller value only considers the worst one from both directions, while taking the sum value considers the contributions from both directions simultaneously.

[0083] In an alternative implementation, the values ​​processed by nonlinear mapping can be used as paired SIRs. Specifically, the onboard processor first converts the SIRs in both directions to dB values, then takes the difference between them, and dynamically determines the aggregation method based on the magnitude of the difference: if the difference between the two SIRs is less than a preset threshold (e.g., 3 dB), it indicates that they are comparable, and the arithmetic mean is taken; if the difference is greater than the threshold, the smaller value is taken. In this method, when the SIRs in the two directions differ significantly, the smaller direction dominates the interference characteristics of the user terminal pair and should be given priority; when the difference is not significant, the average value can more comprehensively reflect the overall interference level. This dynamic aggregation method has better scene adaptability due to its adaptive adjustment strategy based on the symmetry of bidirectional interference.

[0084] It should be noted that all the above-mentioned aggregation methods fall within the scope of this application. In engineering practice, the onboard processor can flexibly choose any one of the above methods or a combination thereof to determine the pairwise SIR between user terminal pairs based on factors such as the statistical characteristics of interference distribution, onboard computing power margin, and user service quality requirements. Regardless of the aggregation strategy adopted, as long as its core idea is based on a comprehensive evaluation of the original SIR values ​​in two directions, and the evaluation result is used as the basis for subsequent user subset partitioning, it does not depart from the protection scope of this application.

[0085] Understandably, if the above-mentioned SIR is used to represent the interference intensity, then when performing user subset partitioning, the interference intensity represented by SIR satisfies the strong interference condition if SIR is less than or equal to a preset threshold. This is because the interference intensity represented by SIR is negatively correlated with SIR. Therefore, if SIR is less than or equal to the preset threshold, it can be considered that its interference intensity is relatively large and satisfies the strong interference condition. At this time, the corresponding user terminal pairs can be partitioned into the same user subset.

[0086] Based on the above embodiments, in the above implementation method for determining the interference intensity characterization quantity between any two user terminals, the interference intensity characterization quantity does not necessarily need to be characterized by SIR, but can be simplified to an approximate measure based on antenna gain difference. Specifically, the onboard processor can query the antenna gain pattern based on the azimuth angle of each user terminal to obtain the antenna gain of each user in its serving beam and in each adjacent beam. Then, it directly calculates the difference (or ratio) between the antenna gain of the same user terminal in the serving beam direction and the antenna gain in the adjacent beam direction, which is used as a measure of the potential interference intensity caused by the user terminal to adjacent beams. In the NTN scenario, free space path loss mainly depends on the distance from the user terminal to the satellite, while the difference in distance from different user terminals to the satellite within the same beam coverage area is relatively small. Therefore, the spatial selectivity of antenna gain is the dominant factor determining the inter-beam interference intensity. This simplified method does not require obtaining the transmit power and scheduling bandwidth information of each user, has a smaller computational load, and is suitable for scenarios where the computing power resources of the onboard processor are extremely limited.

[0087] In an alternative implementation, the interference intensity can also be estimated based on the geometric distance ratio. Specifically, the onboard processor calculates the distance from any two user terminals to the satellite, and then, combined with the serving beam pointing directions of the two user terminals, calculates the angular distance between the two users and the satellite (i.e., the angle between the two user direction vectors). This angular distance is used as the geometric representation of the interference intensity. The basic principle is: the closer the azimuth angles of the two user terminals are, the closer they are in space to the same direction, the lower the spatial isolation between the beams, and the stronger the mutual interference; conversely, the greater the difference in azimuth angles, the higher the spatial isolation, and the weaker the interference.

[0088] In another alternative implementation, the interference intensity can be characterized by the signal-to-interference-plus-noise ratio (SINR) instead of the signal-to-interference-plus-noise ratio (SIR), that is, an additional noise power term is introduced into the denominator of the above SIR calculation formula. Specifically, for user terminal i and user terminal j, the SINR calculation formula in the direction of user terminal i is:

[0089] in, Let be the receiver noise power spectral density. Then, following the same method as SIR, the smaller value in each direction is taken as the paired SINR. The advantage of this method is that when the system operates in a noise-limited area rather than an interference-limited area (e.g., in scenarios with low user density within the beam coverage area), SINR more accurately reflects the actual signal detection quality than SIR. Using it as a characterization of interference intensity can avoid unnecessary computational overhead caused by over-segmenting user subsets when noise is dominant.

[0090] It should be noted that the various alternative implementation methods mentioned above are not mutually exclusive. In actual deployment, the onboard processor can flexibly select and combine the above methods according to the completeness of the current prior information, the onboard computing power margin, the system working mode (interference-limited or noise-limited) and the user distribution characteristics, so as to achieve the optimal balance between the accuracy of the interference intensity representation and the computational efficiency.

[0091] Based on the above embodiments, a greedy algorithm can be used to implement the division of multiple user subsets. For example, a new subset is established using any unassigned user terminal as a seed. The remaining unassigned user terminals are traversed. If the interference intensity represented by the interference intensity representation quantity between the terminal and all existing members in the current subset satisfies the strong interference condition, then the user terminal is added to the current subset. The above process is repeated until all user terminals are assigned.

[0092] Greedy algorithms are a strategy that approximates the global solution by selecting the best local decision at each step. In this scheme, the greedy algorithm is manifested in the fact that when constructing a subset, if an unassigned user has a strong interfering relationship with all members in the current subset, it is immediately added to the current subset, regardless of the global impact of this decision on future subset partitioning. Although this strategy cannot guarantee a globally optimal subset partitioning (for example, it may generate a large number of subsets), it can obtain a feasible partitioning result that meets the interference suppression requirements with relatively low computational complexity.

[0093] This implementation uses a greedy algorithm to partition users into subsets. The basic idea is to construct each subset in an iterative manner, and in each decision step, try to gather the users with strong interference into the same subset.

[0094] The specific process of partitioning is as follows: First, the onboard processor initializes an unassigned user set, marking all K user terminals as "unassigned," at which point no subsets have been established. Then, the onboard processor arbitrarily selects a user terminal from the current unassigned user set as a seed user, creating a new subset with this seed user as its initial member. The seed user can be selected randomly, in order of user terminal number, or preferentially based on the geographical location of the user terminals, prioritizing those located at the beam edge. The selection strategy does not affect the correctness of the partitioning result, but different seed selection orders may affect the final number and size distribution of subsets.

[0095] After the seed users are selected, the onboard processor enters the member expansion phase of the current subset. Specifically, the onboard processor traverses every other user terminal in the currently unassigned user set. For each user terminal to be examined, it calculates the interference strength characterization quantity (such as the SIR value mentioned above) between the terminal and all existing members in the current subset, and then judges the relationship between these SIR values ​​and the preset threshold γ. The preset threshold γ is used as a benchmark value for determining the strength of interference, and the unit is dB or linear ratio. User terminal pairs with an SIR less than or equal to γ ​​(i.e., the interference strength characterized by the interference strength characterization quantity is greater than the preset threshold) are identified as having a strong interference relationship, should be assigned to the same subset, and marked as "assigned"; user terminal pairs with an SIR greater than γ are identified as having a weak interference relationship and are allowed to be distributed in different subsets.

[0096] If the SIR value between user terminal i and every existing member in the current subset is less than or equal to the preset threshold γ (i.e., SIR(i,j)<γ holds true for all members j in the current subset), it indicates that the user terminal has a strong interference relationship with all user terminals in the current subset, satisfying the strong interference condition. In this case, the user terminal is added to the current subset and removed from the unassigned user set. Conversely, if the SIR value between the user terminal and at least one existing member in the current subset is greater than the threshold γ (i.e., the interference between them is weak), the user terminal is not added to the current subset and is retained in the unassigned user set for processing by subsequent subsets.

[0097] It should be noted that the condition "SIR less than or equal to the preset threshold γ" in the above judgment condition corresponds to strong interference, that is, meeting the strong interference condition, because the lower the SIR value, the stronger the interference. The size of the threshold γ determines the granularity of the division. A higher γ value means a more lenient requirement for interference strength, and more user terminal pairs are considered to be strong interference relationships, resulting in a larger subset size and fewer subsets. A lower γ value means a stricter requirement for interference strength, and only user terminal pairs with extremely strong interference are classified into the same subset, resulting in a smaller subset size and more subsets.

[0098] The onboard processor continues to execute the above traversal and addition operations until there is no user terminal in the current unassigned user set that can satisfy the condition that it has a strong interference relationship with all members in the current subset. At this point, the current subset is closed and no new members are accepted.

[0099] Subsequently, the onboard processor selects a new seed user from the unassigned user set, constructs the next subset, and repeats the above membership expansion process. This overall process iterates until all user terminals have been assigned to a subset, at which point the algorithm terminates and outputs the partitioning results for all subsets. The final partitioning results satisfy the following two conditions: the union of all subsets equals the total set of user terminals, and there is no intersection between any two different user subsets, meaning that each user terminal is assigned to one and only one user subset.

[0100] From the perspective of algorithm characteristics, this greedy partitioning algorithm has the following advantages: First, the algorithm has low complexity, with each user terminal being traversed at most once, and the overall computational cost is on the order of O(K²), far less than the O(K³) of full joint balancing, making it suitable for real-time execution by onboard processors. Second, the algorithm can guarantee that all strongly interfering user pairs are partitioned into the same user subset, because when the SIR between two user terminals is less than or equal to the threshold, regardless of which one is selected as the initial member, when traversing another user terminal later, that user terminal will definitely meet the strong interference condition and be added to the same user subset. Finally, any two user terminals in different user subsets, since they did not meet the strong interference condition during the partitioning process, have weak interference between them, and independent processing will not significantly affect the detection performance.

[0101] In the above implementation, the greedy partitioning algorithm achieves effective clustering of interference relationships between user terminals with low computational cost. Each user terminal is traversed at most once, and the overall computational complexity is O(K²), which is far less than the exponential complexity required for the global optimal partitioning, fully adapting to the real-time requirements of the onboard processor. At the same time, by limiting the strong interference condition, it ensures that any newly added user terminal has a strong interference relationship with all members in the subset, thus ensuring that the interference between any two user terminals in the same subset is at a strong interference level, providing structural guarantee for the effectiveness of subsequent joint equilibrium. This partitioning strategy also ensures that there are no strong interference user pairs between different subsets, because if there are strong interference pairs across subsets, the latter will necessarily meet the addition condition and be included in the subset where the former is located during the partitioning process. Therefore, the interference between subsets is controlled at a weak interference level, and independent processing will not produce significant performance loss.

[0102] In the process of partitioning user subsets as described above, in order to further control the upper limit of the computational complexity of a single joint balancing and prevent the size of a subset from becoming too large due to dense interference distribution, thus exceeding the real-time computing power capacity of the onboard processor, this implementation introduces a maximum subset size constraint in the greedy partitioning process. If the size of the current subset reaches the set maximum subset size, then adding new user terminals to the current subset stops, and then a new subset is re-established using any of the remaining unassigned user terminals as seeds.

[0103] Specifically, when executing the greedy partitioning algorithm described above, the onboard processor, in addition to determining whether the user terminal to be added has a strong interference relationship with all existing members in the current subset, also needs to additionally check the size constraint of the current subset. Let the maximum subset size be... N max Its value can be pre-configured as a fixed value during the system design phase. In some implementations, it can also be determined based on the current available computing power of the onboard processor. For example, the maximum dimension allowed for a single matrix inversion can be derived from the current available computing power of the onboard processor, and this dimension can be used as... N max .

[0104] During the process of traversing unassigned user terminals, for each user terminal to be examined, the onboard processor first determines whether the user and all existing members in the current subset are subject to strong interference conditions (e.g., all SIR values ​​are less than or equal to a preset threshold γ). If this condition is met, it further determines whether the number of existing members in the current subset is less than... N max If the current subset size has not yet reached... N max If the user terminal is already large enough, then add it to the current subset; if the current subset size has already reached a certain threshold... N max Even if the user terminal has a strong interference relationship with all members in the current subset, it is not allowed to join the current subset. Instead, it is kept in the unassigned user set and will be taken over by the newly created subset.

[0105] The order of judgment for the two conditions mentioned above can be flexibly adjusted. In one implementation, the onboard processor first judges the interference relationship and then the size constraint, that is, it prioritizes ensuring that users with strong interference are clustered together, and only refuses to join when the subset capacity is full; in another implementation, the onboard processor can first judge the size constraint and then judge the interference relationship, that is, it prioritizes ensuring that the size of each subset does not exceed the upper limit, and then considers the interference relationship.

[0106] After the onboard processor has traversed all user terminals in the current unassigned user set, if no unassigned user terminal simultaneously satisfies both of the above conditions (i.e., there are no users with strong interference relationships with all members of the current subset and the current subset's capacity is not full), then the current subset is closed and no new members are accepted. Subsequently, the onboard processor uses any of the remaining unassigned user terminals as a seed to create a new subset and repeats the above member expansion process from the initial size. This process iterates until all user terminals are assigned, at which point the algorithm terminates.

[0107] Using the above method, the size of each subset obtained in the final partition does not exceed [a certain value]. N maxThis restricts the dimension of the joint equilibrium matrix within each subset to a certain value. N max Within this range, the risk of computing power overload caused by the abnormal expansion of a subset due to uneven distribution of interference is effectively avoided.

[0108] It should be noted that the maximum subset size N max The value set affects the overall performance of the algorithm. N max The larger the value, the more user terminals a single subset can accommodate, the greater the possibility that strongly interfering users will be completely clustered in the same subset, and the better the interference suppression effect, but the higher the computational complexity of a single joint equilibrium. N max The smaller the value, the lower the computational load of a single balancing operation. However, this may cause a group of user terminals that were originally strongly interfering with each other to be split into multiple subsets due to capacity limitations, thus introducing residual interference between subsets that has not been eliminated, sacrificing detection performance to some extent. Therefore, N max The selection of the system needs to be comprehensively optimized based on the real-time computing power of the onboard processor, system latency requirements, and interference distribution characteristics, so as to achieve the best balance between interference suppression performance and computational overhead.

[0109] Specifically, in each scheduling cycle or when a new user terminal connects to the target satellite and triggers the execution of this method, the onboard processor first performs system state awareness and obtains information in the following three dimensions as... N max The selected input parameters are as follows. The first dimension is the real-time computing power status of the onboard processor. The onboard processor reads hardware operating parameters such as current processor load rate, temperature, power consumption, and available storage space through its operating system or real-time monitoring module, and comprehensively evaluates the currently available peak computing power. C ava (The unit can be defined as the number of floating-point operations or fixed-point operations that can be performed per second). The second dimension is the system latency constraint, where the onboard processor obtains the maximum baseband processing latency allowed in the current scheduling cycle. T max (Determined by the system frame structure, scheduling cycle length, and channel coherence time), this delay limits the upper limit of the total time from when the user signal is received by the beam to when joint equalization processing is completed. The third dimension is the user interference distribution characteristics. Based on the SIR matrix calculated in step S120, the onboard processor extracts the statistical characteristics of the SIR between user terminals, including the number of user terminal pairs with an SIR less than or equal to a preset threshold, the number of strong interference neighbors of each user terminal (i.e., the number of other user terminals with an SIR less than or equal to the threshold with a certain user terminal) and their distribution, thereby quantifying the interference density of the current user group.

[0110] After acquiring information from the above three dimensions, the onboard processor executes... N max The optimization calculation. Specifically, the onboard processor first determines the computational constraints from the perspective of computing power. N max upper limit N cap According to the computational complexity model of matrix inversion, the computational cost required to invert a matrix of dimension n is approximately [missing information]. (α and β are constant coefficients related to the specific algorithm implementation; for example, α≈2 / 3 when using Cholesky decomposition for inversion.) The onboard processor will... C ava Substituting this into the model, and reserving a certain amount of computing power margin (e.g., 30% for handling sudden tasks or scheduling fluctuations), we can deduce the maximum matrix dimension that can be processed under the current computing power conditions. N cap Even if The largest integer that is true N cap ,in T slot The available time slot length allocated to baseband processing.

[0111] Meanwhile, the onboard processor determines from the perspective of latency constraints N max upper limit N delay Channel coherence time This is a key physical layer parameter in the NTN system, dependent on the relative velocity between the satellite and the user and the carrier frequency, and calculated based on the Doppler shift. The total baseband processing delay includes equalization matrix calculation delay and signal detection delay, with matrix inversion delay being the dominant factor. The onboard processor constrains this delay. T max Substituting into the calculation model, we can deduce that the processing latency does not exceed T max Maximum matrix dimension N delay .final, N max The initial value is taken as the smaller of the two upper limits mentioned above: This is to ensure that the computing power resources of the onboard processor and the system latency constraints are not exceeded.

[0112] After determining the maximum dimension under the aforementioned computing power and latency constraints, the onboard processor further incorporates interference distribution characteristics... Fine-tuning is performed. The onboard processor statistically analyzes the spatial distribution of strong interfering user terminal pairs in the current SIR matrix, and calculates the average number of strong interfering neighbors d' and the maximum number of strong interfering neighbors d'.max (That is, at most how many other users can a single user have a strong interference relationship with them). If d' is significantly less than If the difference between the two exceeds a set value, it indicates that the interference relationship between users is relatively sparse. A smaller subset size can cover most of the strongly interfering user pairs. In this case, the onboard processor can be appropriately downgraded. N max To save computing power, such as ,in The adjustment factor is less than 1 (e.g., 0.7). N min To ensure the minimum subset size for basic interference suppression capability (typically 2-4). If d max Greater than This indicates that some users have strong interference relationships with multiple other users (i.e., forming dense clusters of strong interference). In this case, if we directly... N max Restricted to This could cause the interference cluster to be split into multiple subsets, introducing residual interference between subsets. To mitigate this problem, the onboard processor can be moderately upgraded. N max However, due to limitations in computing power and latency, the increase cannot exceed [the limit]. Maximum: ,in An upward adjustment factor greater than 1 (e.g., 1.2, but no upward adjustment is made if it does not exceed 1.0, or according to...). d max / The proportion is determined dynamically.

[0113] In a more refined implementation, the onboard processor can forgo coarse-grained adjustments based on statistical features and instead employ an optimization method based on SIR matrix connectivity analysis. Specifically, the onboard processor treats all users as an undirected graph G=(V,E), where vertices V correspond to all users, and edges E correspond to pairs of user terminals with an SIR less than or equal to a threshold γ (i.e., edges with strong interference relationships). The onboard processor performs connectivity analysis on this undirected graph G to calculate the size of the largest connected subgraph in G (i.e., the largest group of users directly or indirectly connected through strong interference relationships). N conn .like N conn ≤ Then the onboard processor can directly... N max Set as At this point, the largest strong interference cluster can be completely contained within a single subset; if N conn > This means that the size of the largest interference cluster exceeds the single-processing capacity of the onboard processor. In this case, a trade-off needs to be made between interference suppression and computational power constraints. The onboard processor will... N max Set as It accepts the fact that the largest interfering cluster is split, while controlling the performance loss introduced by the split within an acceptable range through independent processing between subsets in subsequent steps. The onboard processor can also... N conn and The ratio of the two values ​​triggers an alarm signal, prompting the system scheduler to appropriately reduce the number of users scheduled simultaneously in subsequent scheduling cycles, or to adjust the power control strategy to reduce the interference intensity between users.

[0114] Finally, the onboard processor outputs the optimized maximum subset size. N max This is then used as the scale constraint input for the greedy partitioning algorithm in step S130. If, during the current scheduling cycle, the onboard processor experiences hardware state changes (such as frequency reduction due to increased temperature) or system latency requirements changes (such as switching to a shorter scheduling cycle), the system will be affected. N max The value can be recalculated in each scheduling cycle, achieving dynamic adaptive adjustment. Meanwhile, the onboard processor will... N max The selection results are recorded in the onboard log for offline analysis and parameter optimization by the ground operation and maintenance system, such as for adjusting constant coefficients in the computing power model or optimizing the threshold γ. N max Inter-strategy linkage between them.

[0115] Please refer to the above method embodiments. Figure 2 , Figure 2 This is a structural block diagram of an inter-beam interference suppression device 200 provided in an embodiment of this application. The device 200 may be a module, program segment, or code on an electronic device. It should be understood that the device 200 corresponds to the above method embodiment and is capable of performing the various steps involved in the method embodiment. The specific functions of the device 200 can be found in the description above. To avoid repetition, detailed descriptions are appropriately omitted here.

[0116] Optionally, the device 200 includes: The information acquisition module 210 is used to acquire prior information corresponding to the target satellite. The prior information includes the location information of each user terminal, satellite ephemeris information, beam antenna gain pattern, and the transmit power and scheduling bandwidth of each user terminal. Interference determination module 220 is used to determine the interference intensity characterization quantity between any two user terminals based on the prior information; The subset partitioning module 230 is used to divide the user terminal into multiple user subsets according to the interference intensity characterization quantity, wherein user terminals whose interference intensity characterization quantity satisfies the strong interference condition are divided into the same user subset. The equalization processing module 240 is used to perform joint equalization processing on the uplink signals of each user terminal within each user subset.

[0117] Optionally, the interference determination module 220 is configured to: calculate the distance from each user terminal to the target satellite and the azimuth angle of each user terminal relative to the target satellite based on the location information of each user terminal and the satellite ephemeris information; calculate the free-space path loss of each user terminal based on the distance and the system carrier frequency; obtain the first antenna gain of each user terminal on its serving beam and the second antenna gain on the adjacent beam based on the azimuth angle and the beam antenna gain pattern; calculate the useful signal received power spectral density of each user terminal and the interference signal received power spectral density of each user terminal to other user terminals based on the free-space path loss of each user terminal, the first antenna gain, the second antenna gain, the transmit power and the scheduling bandwidth; and determine the interference intensity characterization quantity between any two user terminals based on the useful signal received power spectral density and the interference signal received power spectral density.

[0118] Optionally, the interference determination module 220 is used to calculate the ratio of the useful signal received power spectral density of user terminal i to the interference signal received power spectral density of user terminal i to user terminal j, to obtain a first interference intensity characterization quantity of user terminal i to user terminal j; calculate the ratio of the useful signal received power spectral density of user terminal j to the interference signal received power spectral density of user terminal j to user terminal i, to obtain a second interference intensity characterization quantity of user terminal j to user terminal i; and determine the interference intensity characterization quantity between user terminal i and user terminal j based on the first interference intensity characterization quantity and the second interference intensity characterization quantity.

[0119] Optionally, the interference determination module 220 is used to determine the smaller value between the first interference intensity characterization value and the second interference intensity characterization value as the interference intensity characterization value between user terminal i and user terminal j.

[0120] Optionally, the beam antenna gain pattern is stored in the onboard storage unit in the form of a lookup table (LUT). The lookup table is indexed by the directional angle and stores the antenna gain of each receiving beam in the corresponding direction. The interference determination module 220 is used to look up the first antenna gain of each user terminal on its serving beam and the second antenna gain on the adjacent beam in the lookup table according to the directional angle. If the directional angle does not fall on a discrete grid point of the lookup table, the antenna gain in that direction is calculated using a bilinear interpolation algorithm to obtain the first antenna gain of each user terminal on its serving beam and the second antenna gain on the adjacent beam.

[0121] Optionally, the subset partitioning module 230 is used to establish a new subset with any unassigned user terminal as a seed; traverse the remaining unassigned user terminals, and if the interference intensity represented by the interference intensity representation quantity between the user terminal and all existing members in the current subset satisfies the strong interference condition, then add the user terminal to the current subset; repeat the above process until all user terminals are assigned.

[0122] Optionally, the subset partitioning module 230 is used to stop adding new user terminals to the current subset when the size of the current subset reaches the set maximum subset size; and to re-establish a new subset using any of the remaining unallocated user terminals as seeds.

[0123] Optionally, the maximum subset size is determined based on the currently available computing power of the onboard processor.

[0124] Optionally, the equalization processing module 240 is used to extract the channel gain from each user terminal to each associated beam for each user subset, and construct the channel sub-matrix of the user subset; based on the channel sub-matrix, calculate the equalization matrix using zero-forcing equalization or minimum mean square error equalization; and apply the equalization matrix to the received signal vector of the user subset to obtain the estimated value of the uplink signal of each user terminal in the user subset.

[0125] It should be noted that those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the device described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0126] Please refer to Figure 3 , Figure 3This application provides a schematic diagram of an electronic device for performing an inter-beam interference suppression method. The electronic device may include: at least one processor 310, such as a CPU; at least one communication interface 320; at least one memory 330; and at least one communication bus 340. The communication bus 340 is used to establish communication between these components. In this embodiment, the communication interface 320 is used for signaling or data communication with other node devices. The memory 330 may be a high-speed RAM or non-volatile memory, such as at least one disk storage device. Optionally, the memory 330 may also be at least one storage device located remotely from the processor. The memory 330 stores computer-readable instructions, which, when executed by the processor 310, enable the electronic device to perform the aforementioned method process.

[0127] Understandable. Figure 3 The structure shown is for illustrative purposes only; the electronic device may also include components that are more advanced than those shown. Figure 3 The more or fewer components shown, or having the same Figure 3 The different configurations shown. Figure 3 The components shown can be implemented using hardware, software, or a combination thereof.

[0128] This application provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it performs the method process executed by the electronic device in the above method embodiments.

[0129] This embodiment discloses a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions, and when the program instructions are executed by a computer, the computer can perform the methods provided in the above-described method embodiments.

[0130] In summary, this application provides an inter-beam interference suppression method, electronic device, storage medium, and program product. This method utilizes prior information available in the NTN system to predict the inter-user interference intensity, replacing the traditional interference acquisition method that relies on real-time channel estimation, thus avoiding the problem of unreliable channel feedback information under long propagation delays. Based on pre-calculated interference intensity characteristics, strong interfering users are clustered into the same user subset for joint equalization, effectively suppressing inter-beam interference within the subset. Simultaneously, different user subsets are processed independently, fundamentally solving the deficiency of independent beam processing schemes in suppressing inter-beam interference and significantly improving the signal detection performance of users at the beam edge. Furthermore, through subset partitioning, the matrix dimension of joint equalization can be reduced from the global number of users to the scale of each subset, reducing the overall computational complexity and enabling the onboard processor to complete real-time processing within the channel coherence time, overcoming the bottleneck of the full joint equalization scheme being impractical in engineering. Therefore, this solution achieves an effective balance between onboard computing power constraints and interference suppression performance.

[0131] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0132] Furthermore, the units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0133] Furthermore, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0134] In this document, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, without necessarily requiring or implying any such actual relationship or order between these entities or operations.

[0135] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for suppressing inter-beam interference, characterized in that, Applied to a spaceborne processor, the method includes: Obtain prior information corresponding to the target satellite, including the location information of each user terminal, satellite ephemeris information, beam antenna gain pattern, and the transmit power and scheduling bandwidth of each user terminal; Determine the interference intensity characterization quantity between any two user terminals based on the prior information; The user terminals are divided into multiple user subsets according to the interference intensity characterization quantity, wherein user terminals whose interference intensity characterization quantity meets the strong interference condition are divided into the same user subset. Joint equalization processing is performed on the uplink signals of each user terminal within each user subset.

2. The method according to claim 1, characterized in that, The step of determining the interference intensity characterization quantity between any two user terminals based on the prior information includes: Based on the location information of each user terminal and the satellite ephemeris information, calculate the distance from each user terminal to the target satellite and the azimuth angle of each user terminal relative to the target satellite; Based on the distance and system carrier frequency, calculate the free space path loss of each user terminal; Based on the azimuth angle and the beam antenna gain pattern, the first antenna gain of each user terminal on its serving beam and the second antenna gain on the adjacent beam are obtained. Based on the free space path loss of each user terminal, the first antenna gain, the second antenna gain, the transmit power and the scheduling bandwidth, calculate the useful signal received power spectral density of each user terminal and the interference signal received power spectral density of each user terminal to other user terminals. Based on the received power spectral density of the useful signal and the received power spectral density of the interference signal, an interference intensity characterization quantity between any two user terminals is determined.

3. The method according to claim 2, characterized in that, The step of determining the interference intensity characterization quantity between any two user terminals based on the received power spectral density of the useful signal and the received power spectral density of the interference signal includes: Calculate the ratio of the received power spectral density of the useful signal of user terminal i to the received power spectral density of the interference signal of user terminal i to user terminal j, and obtain the first interference intensity characterization quantity of user terminal i to user terminal j. The ratio of the received power spectral density of the useful signal of user terminal j to the received power spectral density of the interference signal of user terminal j to user terminal i is calculated to obtain the second interference intensity characterization quantity of user terminal j to user terminal i. Based on the first interference intensity characterization value and the second interference intensity characterization value, the interference intensity characterization value between user terminal i and user terminal j is determined.

4. The method according to claim 3, characterized in that, The step of determining the interference intensity characterization between user terminal i and user terminal j based on the first interference intensity characterization and the second interference intensity characterization includes: The smaller value between the first interference intensity characterization value and the second interference intensity characterization value is determined as the interference intensity characterization value between user terminal i and user terminal j.

5. The method according to claim 2, characterized in that, The beam antenna gain pattern is stored in the onboard storage unit in the form of a lookup table (LUT). The lookup table is indexed by the azimuth angle and stores the antenna gain of each received beam in the corresponding direction. The step of obtaining the first antenna gain of each user terminal on its serving beam and the second antenna gain on adjacent beams based on the azimuth angle and the beam antenna gain pattern includes: According to the azimuth angle, the lookup table is used to find the first antenna gain of each user terminal on its serving beam and the second antenna gain on the adjacent beam; If the azimuth angle does not fall on a discrete grid point of the lookup table, a bilinear interpolation algorithm is used to calculate the antenna gain in that direction to obtain the first antenna gain of each user terminal on its serving beam and the second antenna gain on the adjacent beam.

6. The method according to claim 1, characterized in that, The step of dividing the user terminal into multiple user subsets based on the interference intensity characterization includes: Create a new subset using any unassigned user terminal as the seed; Traverse the remaining unassigned user terminals. If the interference intensity represented by the interference intensity representation quantity between the user terminal and all existing members in the current subset satisfies the strong interference condition, then add the user terminal to the current subset. Repeat the above process until all user terminals have been allocated.

7. The method according to claim 6, characterized in that, The step of dividing the user terminal into multiple user subsets based on the interference intensity characterization quantity further includes: If the size of the current subset reaches the set maximum subset size, stop adding new user terminals to the current subset; A new subset is rebuilt using any of the remaining unassigned user terminals as the seed.

8. The method according to claim 7, characterized in that, The maximum subset size is determined based on the currently available computing power of the onboard processor.

9. The method according to claim 1, characterized in that, The joint equalization processing of the uplink signals of each user terminal within each user subset includes: For each user subset, extract the channel gain from each user terminal to each associated beam within that user subset, and construct the channel sub-matrix for that user subset; Based on the channel submatrix, the equalization matrix is ​​calculated using zero-forcing equalization or minimum mean square error equalization. The equalization matrix is ​​applied to the received signal vector of the user subset to obtain an estimate of the uplink signal of each user terminal within the user subset.

10. An electronic device, characterized in that, It includes a processor and a memory, the memory storing computer-readable instructions that, when executed by the processor, perform the method as described in any one of claims 1-9.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it performs the method as described in any one of claims 1-9.

12. A computer program product, characterized in that, It includes computer program instructions, which, when read and executed by a processor, perform the method as described in any one of claims 1-9.

Images