Beamforming method and apparatus for massive MIMO systems

Abstract

The application provides a beamforming method and device for a large-scale MIMO system, comprising: selecting an optimal antenna subset from a plurality of antennas of the large-scale MIMO system by using a zebra optimization algorithm; wherein parameters of the zebra optimization algorithm are adjusted in real time based on fuzzy logic heuristic guidance; performing channel state information estimation by using a compressed sensing algorithm to generate channel state information of the optimal antenna subset; calculating an optimal beamforming vector based on the channel state information; and controlling phases and amplitudes of an antenna array corresponding to the optimal antenna subset based on the optimal beamforming vector. By using the zebra optimization algorithm with fuzzy logic heuristic guidance to select antennas, dynamic adjustment of antenna selection can be achieved, and adaptive optimization of beamforming can be achieved, so that efficient performance and resource utilization efficiency can be maintained in different communication environments.

Description

Technical Field

[0001] This invention relates to the field of wireless technology, and in particular to a beamforming method and apparatus for a large-scale MIMO system. Background Technology

[0002] In massive MIMO (Multiple-Input Multiple-Output) systems, antenna selection and beamforming are key technologies for improving communication performance. Antenna selection aims to select a subset of available antennas to reduce system hardware overhead and computational complexity while maintaining system performance as much as possible. Beamforming, on the other hand, enhances signal strength in the target direction by adjusting the phase and amplitude of the antenna array, while suppressing interference in other directions, thereby improving the system's signal transmission quality and capacity.

[0003] Existing massive MIMO systems perform well in stable or slowly changing communication environments, optimizing signal transmission and system capacity through mature antenna selection and beamforming techniques. However, existing systems face challenges in highly dynamic and variable communication environments. Due to a lack of real-time adaptability and flexibility, existing beamforming schemes may exhibit unstable performance when dealing with rapidly changing channel conditions and mobile user scenarios, affecting overall communication performance and user experience.

[0004] Therefore, how to cope with the dynamically changing communication environment and provide better performance in complex communication environments is an urgent problem to be solved. Summary of the Invention

[0005] This invention provides a beamforming method and apparatus for a large-scale MIMO system to address the shortcomings of existing technologies in terms of performance instability when dealing with rapidly changing channel conditions and mobile user scenarios, thereby achieving dynamic adjustment of antenna selection and adaptive optimization of beamforming.

[0006] This invention provides a beamforming method for a large-scale MIMO system, comprising: The Zebra optimization algorithm is used to select the optimal subset of antennas from multiple antennas in a large-scale MIMO system; wherein the parameters of the Zebra optimization algorithm are adjusted in real time based on fuzzy logic heuristics. Channel state information is estimated using a compressed sensing algorithm to generate the channel state information of the optimal antenna subset; Based on the channel state information, the optimal beamforming vector is calculated; Based on the optimal beamforming vector, the phase and amplitude of the antenna array corresponding to the optimal antenna subset are controlled.

[0007] According to the present invention, a beamforming method for a massive MIMO system is provided, wherein the zebra optimization algorithm is used to select an optimal subset of antennas from multiple antennas in the massive MIMO system, including: The Zebra Optimization Algorithm is used to randomly select a preset number of antennas from multiple antennas in a large-scale MIMO system, generate multiple solutions to form an initial population, and iteratively optimize the system step by step; where each solution represents a subset of antennas. In each iteration, the solution with the largest fitness value in the current population is determined as the current optimal solution; Based on the current optimal solution and random perturbation, other solutions in the current population are updated, and the preset number of solutions with the largest fitness values ​​are selected to form a new population. Stop iterating when the algorithm converges and output the optimal subset of antennas.

[0008] According to the beamforming method for a large-scale MIMO system provided by the present invention, determining the solution with the largest fitness value in the current population as the current optimal solution includes: For each solution in the current population, calculate the total signal gain based on the gain of each antenna; Calculate the total energy consumption based on the energy consumption of each antenna; Computational complexity based on the number of antennas; The fitness value of each solution is obtained by weighting the total signal gain, the total energy consumption, and the complexity. The solution with the highest fitness value in the current population is determined as the current optimal solution.

[0009] According to the beamforming method for a large-scale MIMO system provided by the present invention, the step of updating other solutions in the current population based on the current optimal solution and random perturbation includes: Calculate the random disturbance based on the disturbance amplitude control coefficient and the random vector; The other solutions in the current population are updated based on the search range adjustment coefficient, the current optimal solution, the perturbation amplitude adjustment coefficient, and the random perturbation.

[0010] According to the beamforming method for a large-scale MIMO system provided by the present invention, the parameters of the zebra optimization algorithm are adjusted in real time in the following manner: During each iteration, the diversity and convergence rate of the current population are calculated; Based on the diversity and convergence speed of the current population, fuzzy reasoning is performed to obtain the membership degree of the search range adjustment coefficient and the membership degree of the perturbation amplitude adjustment coefficient; the search range adjustment coefficient is used to control the strength of the current solution towards the optimal solution, and the perturbation amplitude adjustment coefficient is used to control the intensity of the introduced random perturbation. The membership degrees of the search range adjustment coefficient and the disturbance amplitude adjustment coefficient are defuzzified to obtain the membership quantification values ​​of the search range adjustment coefficient and the disturbance amplitude adjustment coefficient. The parameters of the Zebra optimization algorithm are adjusted in real time based on the membership quantification values ​​of the search range adjustment coefficient and the perturbation amplitude adjustment coefficient.

[0011] According to a beamforming method for a large-scale MIMO system provided by the present invention, the calculation of the diversity and convergence rate of the current population includes: The diversity of the current population is calculated based on the absolute value of the difference between each solution in the current population and the mean of all solutions in each dimension, the number of solutions in the current population, and the preset number. The convergence rate of the current population is calculated based on the optimal fitness value in the current iteration and the optimal fitness value in the previous iteration.

[0012] According to the beamforming method for a large-scale MIMO system provided by the present invention, the step of performing fuzzy inference based on the diversity and convergence speed of the current population to obtain the membership degree of the search range adjustment coefficient and the membership degree of the perturbation amplitude adjustment coefficient includes: Using triangular membership functions, the diversity of the current population is fuzzed based on the highest, lowest, and medium diversity to obtain the membership degree of the search range adjustment coefficient of the current population. Using triangular membership functions, the convergence speed of the current population is fuzzified based on the highest, lowest, and medium convergence speeds to obtain the membership degree of the perturbation amplitude adjustment coefficient of the current population.

[0013] According to the beamforming method for a large-scale MIMO system provided by the present invention, the step of estimating channel state information using a compressed sensing algorithm to generate channel state information for the optimal antenna subset includes: The channel matrix corresponding to the optimal antenna subset is converted into a sparse vector using a sparse basis; Based on the number of non-zero elements in the sparse vector and the total length of the sparse vector, the sparsity of the channel matrix under the sparse basis is calculated. If the channel matrix is ​​determined to meet the compressed sensing conditions based on the sparsity, the channel matrix is ​​sampled using a measurement matrix to obtain a measurement vector; The original channel matrix is ​​reconstructed from the measurement data using a compressed sensing algorithm, thereby obtaining the channel state information of the optimal antenna subset.

[0014] According to the beamforming method for a large-scale MIMO system provided by the present invention, the step of calculating the optimal beamforming vector based on the channel state information includes: Determine the objective function for maximizing signal gain and interference suppression; When the objective function is a convex function, a convex optimization algorithm is used to solve the objective function to obtain the optimal beamforming vector; When the objective function is non-convex, the gradient descent method is used to solve the objective function to obtain the optimal beamforming vector.

[0015] The present invention also provides a beamforming device for a massive MIMO system, comprising: The antenna subset selection module is used to select the optimal antenna subset from multiple antennas in a large-scale MIMO system using the Zebra optimization algorithm; wherein the parameters of the Zebra optimization algorithm are adjusted in real time based on fuzzy logic heuristic guidance. The channel state information estimation module is used to estimate the channel state information using a compressed sensing algorithm to generate the channel state information of the optimal antenna subset. The beamforming optimization module is used to calculate the optimal beamforming vector based on the channel state information. The control module is used to control the phase and amplitude of the antenna array corresponding to the optimal antenna subset based on the optimal beamforming vector.

[0016] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a beamforming method for any of the aforementioned large-scale MIMO systems.

[0017] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the beamforming method for any of the above-described large-scale MIMO systems.

[0018] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements a beamforming method for any of the large-scale MIMO systems described above.

[0019] The beamforming method and apparatus for massive MIMO systems provided by this invention employ a zebra optimization algorithm guided by fuzzy logic heuristics to select an optimal subset of antennas from multiple antennas in a massive MIMO system. A compressed sensing algorithm is then used to estimate channel state information (CSO) to generate CSO for the optimal antenna subset. Based on the CSO, an optimal beamforming vector is calculated, and the phase and amplitude of the antenna array corresponding to the optimal antenna subset are controlled based on this vector. By using a zebra optimization algorithm guided by fuzzy logic heuristics for antenna selection, this invention enables dynamic adjustment of antenna selection, thereby achieving adaptive optimization of beamforming. This maintains high performance and resource utilization efficiency under different communication environments. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is a schematic flowchart of the beamforming method for a large-scale MIMO system provided in an embodiment of the present invention.

[0022] Figure 2 This is one of the structural schematic diagrams of the beamforming device for a large-scale MIMO system provided in the embodiments of the present invention.

[0023] Figure 3 This is the second schematic diagram of the beamforming device of the large-scale MIMO system provided in the embodiments of the present invention.

[0024] Figure 4 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0026] In the description of embodiments of the present invention, the terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Those skilled in the art will understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0027] This invention employs a zebra optimization algorithm guided by fuzzy logic for antenna selection, enabling dynamic adjustment of antenna selection and adaptive optimization of beamforming. This allows for maintaining high performance and resource utilization efficiency in different communication environments.

[0028] Figure 1 This is a schematic flowchart of the beamforming method for a large-scale MIMO system provided in an embodiment of the present invention. (Refer to...) Figure 1 This invention provides a beamforming method for a large-scale MIMO system, which may specifically include the following steps: Step 101: The Zebra Optimization Algorithm is used to select the optimal subset of antennas from multiple antennas in the large-scale MIMO system; wherein the parameters of the Zebra Optimization Algorithm are adjusted in real time based on fuzzy logic heuristic guidance.

[0029] It should be noted that the beamforming method for the large-scale MIMO system provided in this embodiment of the invention can be executed by an electronic device, a component in the electronic device, an integrated circuit, or a chip. The electronic device can be a mobile electronic device or a non-mobile electronic device. For example, a mobile electronic device can be a mobile phone, tablet computer, laptop computer, PDA, wearable device, ultra-mobile personal computer (UMPC), netbook, or personal digital assistant (PDA), etc., while a non-mobile electronic device can be a server, network attached storage (NAS), personal computer (PC), television (TV), ATM, or self-service machine, etc. This embodiment of the invention does not specifically limit the specific implementation of these methods.

[0030] In this embodiment of the invention, an optimal subset of antennas can be selected from a large number of antennas in a massive MIMO system, thereby reducing computational complexity while ensuring system performance.

[0031] In some embodiments, multiple antenna subsets can be generated as an initial population to ensure broad coverage of the search space. Then, based on a fuzzy rule base and fuzzy inference mechanism, the search strategy of the Zebra Optimization Algorithm can be dynamically adjusted to enhance the algorithm's global search capability and adaptability. The objective function value for each antenna subset can be calculated by comprehensively considering signal gain, energy consumption, and computational complexity, thereby evaluating the merits of each subset. The Zebra Optimization Algorithm can be executed to progressively optimize and select the final optimal antenna subset.

[0032] Compared to existing fixed heuristic methods for antenna selection and beamforming, which cannot adaptively adjust to real-time environmental changes, especially in dynamic wireless environments (such as user movement and rapid changes in channel conditions), the optimization effect is greatly limited. This invention introduces the Zebra Optimization Algorithm (ZOA) combined with fuzzy logic heuristic guidance to achieve dynamic adjustment and adaptive optimization. It can dynamically select antenna subsets and adjust beamforming parameters according to real-time monitored environmental changes.

[0033] Compared to existing zero-interference or minimum mean square error algorithms for adjusting beam weights, which are difficult to effectively suppress interference when multipath interference is severe, this invention combines a dynamic adjustment mechanism guided by fuzzy logic heuristics. This mechanism can dynamically optimize beamforming based on real-time CSI information, enhance signal directivity, and reduce the impact of multipath interference.

[0034] Step 102: Channel state information is estimated using a compressed sensing algorithm to generate the channel state information of the optimal antenna subset.

[0035] In this embodiment of the invention, compressed sensing technology can be used to accurately estimate the channel state information (CSI) of the determined optimal antenna subset, generating high-precision CSI data as the basis for beamforming.

[0036] In some embodiments, the sparsity of the channel matrix can be analyzed, and a suitable sparse basis can be selected. Low-dimensional sampling of the channel matrix reduces the amount of sampled data while maintaining the integrity of key channel information. High-dimensional channel state information can be reconstructed from the generated low-dimensional data. Finally, parameters during compressed sensing sampling and channel reconstruction can be adjusted in real time to cope with dynamic changes in the channel environment.

[0037] Compared to traditional CSI estimation methods (such as linear estimation), which may fail to provide sufficiently accurate channel information in low signal-to-noise ratio (SNR) environments (such as long-distance communication and underground infrastructure communication), thus affecting beamforming and antenna selection, this invention utilizes improved compressed sensing technology and sparsity analysis to effectively extract and reconstruct CSI even under low SNR conditions, ensuring the accuracy of channel information.

[0038] Step 103: Calculate the optimal beamforming vector based on the channel state information.

[0039] In this embodiment of the invention, precise CSI data can be used to optimize the antenna beamforming vector to maximize signal gain and suppress interference, thereby ensuring the communication quality and overall performance of the system.

[0040] In some embodiments, an objective function can be designed based on the reconstructed CSI data, taking into account factors such as signal gain, interference suppression, and energy efficiency. An optimization algorithm is then used to calculate the optimal beamforming vector, determining the antenna's transmission direction and power allocation. Based on the calculated optimal beamforming vector, the corresponding antenna is activated to achieve optimal signal transmission.

[0041] Step 104: Based on the optimal beamforming vector, control the phase and amplitude of the antenna array corresponding to the optimal antenna subset.

[0042] In this embodiment of the invention, after calculating the optimal beamforming vector, antenna activation and signal transmission can be performed. That is, according to the calculated beamforming vector, the corresponding antenna is activated, and signal transmission is performed according to the optimal vector. A subset of antennas in the optimal solution can be activated, other unselected antennas can be turned off, and the selected antenna subset can be recorded and used as a reference for subsequent channel measurements, thereby providing a basis for subsequent CSI estimation and beamforming.

[0043] In some embodiments, the optimal beamforming vector can be used. Each antenna in the optimal antenna subset is configured with the appropriate phase and amplitude to form the desired beam pattern. Phased array technology can be used to precisely control the phase and amplitude of the antennas, so that the signal forms the strongest beam in a specified direction.

[0044] In some embodiments, signal transmission and real-time adjustments can be performed to ensure optimal system performance during actual transmission and to make adjustments based on real-time feedback. Key performance indicators during transmission, such as signal strength, interference level, and signal-to-noise ratio, can be monitored in real time, and the beamforming vector or other transmission parameters can be adjusted based on the monitoring results.

[0045] When signal quality degradation or increased interference is detected, the beamforming vector can be adjusted through a dynamic feedback mechanism. This feedback mechanism may include adjusting phase and amplitude, recalculating the beamforming vector, or, in extreme cases, reselecting an antenna subset and re-estimating the CSI.

[0046] This invention employs a zebra optimization algorithm guided by fuzzy logic to select an optimal subset of antennas from multiple antennas in a large-scale MIMO system. A compressed sensing algorithm is used to estimate channel state information (CSO) for this optimal subset, generating CSO data for that subset. Based on this CSO data, an optimal beamforming vector is calculated, and the phase and amplitude of the antenna array corresponding to the optimal subset are then controlled using this vector. By using this fuzzy logic-guided zebra optimization algorithm for antenna selection, this invention enables dynamic adjustment of antenna selection, thereby achieving adaptive optimization of beamforming. This maintains high performance and resource utilization efficiency under various communication environments.

[0047] In an optional embodiment, the step of employing the Zebra Optimization Algorithm to select the optimal subset of antennas from multiple antennas in a massive MIMO system may specifically include: Step S11: Using the Zebra optimization algorithm, a preset number of antennas are randomly selected from multiple antennas in the large-scale MIMO system to generate multiple solutions to form an initial population, and then iteratively optimized step by step; wherein, each solution is used to represent a subset of antennas; Step S12: In each iteration, the solution with the largest fitness value in the current population is determined as the current optimal solution; Step S13: Based on the current optimal solution and random perturbation, update the other solutions in the current population, and select the preset number of solutions with the largest fitness values ​​to form a new population; Step S14: Stop iterating when the algorithm converges and output the optimal antenna subset.

[0048] In this embodiment of the invention, the problem space can be defined first, with a total of N... t A subset is formed by selecting d antennas from a given set. The problem space can be defined as a set. Each solution X i It is a set containing d elements, which can represent which antennas were selected.

[0049] In some embodiments, N may be assumed t =128, and d=16.

[0050] In this embodiment of the invention, N solutions can be generated, each solution representing a subset of antennas. Randomly selected from N... t Select d antennas without repetition from the d antennas to generate an initial solution X. iRepeat the above steps N times to generate the initial population.

[0051] In this embodiment of the invention, the fitness of each solution can be calculated to measure its performance in terms of signal gain, energy consumption, and computational complexity. Based on the fitness of each solution, the current optimal solution in the current population is determined, providing a reference for the next update. The optimal solution X can be used as a reference. best The introduced randomness is used to update other solutions in the population.

[0052] In some embodiments, after each iteration, the N solutions with the highest fitness values ​​can be selected to form a new population; solutions with fitness values ​​below a threshold are excluded and replaced with randomly generated solutions to maintain population diversity, thereby ensuring that the solutions in the population always maintain a high fitness through population selection.

[0053] In some embodiments, it can be determined whether the algorithm has reached a convergence state to stop iteration, and the final optimal subset of antennas can be output when the algorithm converges for subsequent channel state information (CSI) estimation. When the convergence condition is met, the solution X with the highest fitness value in the current population can be extracted. best As the optimal subset of antennas.

[0054] One of the convergence conditions is reaching the maximum number of iterations T. max For example, it could be set to 100 iterations. A second convergence condition could be that the fitness value of the optimal solution in the population changes by less than a set threshold over multiple consecutive iterations. ,For example =0.001. If any condition is met, the algorithm converges and outputs the optimal solution X. best For example, the final subset of antennas could be: X best ={3,7,12,21,25,31,44,50,63,70,88,91,94,105,110,127}.

[0055] In an optional embodiment, determining the solution with the highest fitness value in the current population as the current optimal solution may specifically include: Step S121: For each solution in the current population, calculate the total signal gain based on the gain of each antenna; Step S122: Calculate the total energy consumption based on the energy consumption of each antenna; Step S123: Calculate the complexity based on the number of antennas; Step S124: Perform a weighted calculation on the total signal gain, the total energy consumption, and the complexity to obtain the fitness value of each solution; Step S125: Determine the solution with the largest fitness value in the current population as the current optimal solution.

[0056] In some embodiments, for each solution X i The fitness value f(X) can be calculated. i The specific calculation formula is as follows: (1); in, Is the solution X i fitness value, Is the solution X i signal gain, Is the solution X i Energy consumption, Is the solution X i The complexity, , and These are the weighting coefficients for signal gain, energy consumption, and complexity, respectively.

[0057] In some embodiments, the formula for calculating signal gain can be: (2); in, Is the solution X i signal gain, G(A ij ) It is the Ath ij The gain of each antenna can be obtained through channel state information (CSI) or pre-measurement.

[0058] In some embodiments, the formula for calculating energy consumption can be: (3); in, Is the solution X i Energy consumption, P(A) ij ) is the Ath ij Energy consumption of each antenna.

[0059] In some embodiments, the formula for calculating complexity can be: (4); in, Is the solution X i The computational complexity is given by , where d is the number of antennas. For large-scale MIMO systems, the computational complexity primarily depends on the size of the antenna subset.

[0060] In this embodiment of the invention, the fitness values ​​f(X) of all solutions can be compared in the initial population. i Find the solution X with the largest fitness value. best Record this solution as the current optimal solution, and continue to use it to guide the updating of other solutions.

[0061] In an optional embodiment, updating other solutions in the current population based on the current optimal solution and the random perturbation may specifically include: Step S131: Calculate the random disturbance based on the disturbance amplitude control coefficient and the random vector; Step S132: Update other solutions in the current population based on the search range adjustment coefficient, the current optimal solution, the perturbation amplitude adjustment coefficient, and the random perturbation.

[0062] In some embodiments, for each solution X i The update formula can be: X i (t+1)=X i (t)+α·(X best -X i (t))+β·Random Perturbation(5); Among them, X i (t) is the current solution X i X i (t+1) is the updated solution X. i X best It is the current optimal solution. RandomVector is a random vector uniformly distributed in the range [0, 1]. α is the search range adjustment coefficient, and β is the perturbation amplitude adjustment coefficient.

[0063] The formula for calculating random disturbances can be: Random Perturbation=γ·(Random Vector-0.5)(6); Where γ is the control coefficient for the disturbance amplitude, and Random Vector is a random vector uniformly distributed in the range [0, 1].

[0064] In this embodiment of the invention, a new solution can be generated by updating the formula, and its fitness value can be evaluated. The position X of each solution can be updated. i (t+1) and calculate its new fitness value f(X). i (t+1)). If f(X) i (t+1))>f(X best If X ), then the optimal solution is updated to X. i (t+1).

[0065] In an optional embodiment, the parameters of the Zebra optimization algorithm can be adjusted in real time through the following steps: Step S21: In each iteration, calculate the diversity and convergence rate of the current population; Step S22: Based on the diversity and convergence speed of the current population, perform fuzzy inference to obtain the membership degree of the search range adjustment coefficient and the membership degree of the perturbation amplitude adjustment coefficient; the search range adjustment coefficient is used to control the strength of the current solution towards the optimal solution, and the perturbation amplitude adjustment coefficient is used to control the intensity of the introduced random perturbation. Step S23: Defuzzify the membership degree of the search range adjustment coefficient and the membership degree of the disturbance amplitude adjustment coefficient to obtain the membership quantification value of the search range adjustment coefficient and the membership quantification value of the disturbance amplitude adjustment coefficient. Step S24: Based on the membership quantification values ​​of the search range adjustment coefficient and the perturbation amplitude adjustment coefficient, the parameters of the Zebra optimization algorithm are adjusted in real time.

[0066] In this embodiment of the invention, fuzzy logic heuristics can be used to adjust the parameters of the Zebra optimization algorithm in real time, improving its adaptability and search capabilities, especially when dealing with complex optimization problems.

[0067] In some embodiments, fuzzy input variables, fuzzy output variables, and fuzzy rules can be determined to achieve the design of a fuzzy rule base.

[0068] Fuzzy input variables can include population diversity (D) and convergence rate (C). Population diversity measures the diversity of solutions within the population, reflecting whether the algorithm might get stuck in local optima. Convergence rate measures the rate at which the optimal solution converges, reflecting how quickly the algorithm approaches the global optimum.

[0069] The fuzzy output variables can include a search range adjustment coefficient and a perturbation amplitude adjustment coefficient. The search range adjustment coefficient α can be used to control the strength of the solution's tendency to approach the optimal solution during foraging behavior. The perturbation amplitude adjustment coefficient β can be used to control the intensity of the introduced random perturbation; increasing β can expand the search range and avoid local optima.

[0070] Examples of fuzzy rules can be found in Table 1: Table 1

[0071] Fuzzy rule one can be designed as follows: if D is low and C is high, then α increases and β decreases, focusing the search on the region near the current optimal solution.

[0072] Fuzzy rule one can be designed as follows: if D is high and C is low, then α decreases and β increases, expanding the search range and exploring new solution spaces.

[0073] In this embodiment of the invention, a fuzzy rule base can be used to derive the membership degrees of fuzzy output variables α and β based on the membership degrees of fuzzy input variables D and C. If D belongs to "medium" and C belongs to "low", then according to the rule base, α belongs to "medium" and β belongs to "high".

[0074] In this embodiment of the invention, real-time monitoring and feedback can be performed to ensure that the algorithm can adaptively adjust during operation and avoid getting trapped in local optima.

[0075] In some embodiments, the population diversity D and convergence speed C can be calculated in real time during each iteration, and α and β can be dynamically adjusted based on the fuzzy inference results, thereby achieving adaptive adjustment of the Zebra Optimization Algorithm during operation. For example, if a decrease in population diversity and an increase in convergence speed are detected, α is increased and β is decreased to concentrate the search.

[0076] This invention updates fuzzy inference rules using real-time monitored data, optimizes the search strategy, and ensures the algorithm can search effectively at different stages. A feedback mechanism continuously optimizes the algorithm parameters, making it adaptive in dynamic environments.

[0077] In one optional embodiment, calculating the diversity and convergence rate of the current population may specifically include: Step S211: Calculate the diversity of the current population based on the absolute value of the difference between each solution in the current population and the mean of all solutions in each dimension, the number of solutions in the current population, and the preset number. Step S212: Calculate the convergence rate of the current population based on the optimal fitness value in the current iteration and the optimal fitness value in the previous iteration.

[0078] In this embodiment of the invention, population diversity D can be used to measure the diversity of solutions in the population, reflecting whether the algorithm may get trapped in a local optimum. The formula for calculating population diversity is as follows: (7); in, Let represent the mean of all solutions along the j-th dimension. X represents the solution in the j-th dimension. i D represents population diversity, and N represents the number of solutions.

[0079] The convergence rate C can be used to measure the rate at which the optimal solution converges, reflecting how quickly the algorithm approaches the global optimum. The formula is as follows: (8); Where, f(X) best f(X) represents the optimal fitness value in the t-th iteration. best(t-1)) represents the optimal fitness value in the (t-1)th iteration, and C is the convergence rate.

[0080] In an optional embodiment, the step of performing fuzzy inference based on the diversity and convergence speed of the current population to obtain the membership degrees of the search range adjustment coefficient and the perturbation amplitude adjustment coefficient may specifically include: Step S221: Using the triangular membership function, the diversity of the current population is fuzzed based on the highest diversity, lowest diversity, and medium diversity to obtain the membership degree of the search range adjustment coefficient of the current population. Step S222: Using the triangular membership function, the convergence speed of the current population is fuzzified based on the highest convergence speed, lowest convergence speed, and medium convergence speed to obtain the membership degree of the perturbation amplitude adjustment coefficient of the current population.

[0081] In this embodiment of the invention, membership functions for fuzzy variables can be defined. For example, for population diversity D, the following triangular membership function can be defined (the membership function for convergence velocity C is similar to that for population diversity D, and will not be described again): (9); Among them, D 高 It has the highest population diversity, D 中 It is of medium population diversity, D 低 μ represents the minimum population diversity, D represents the current population diversity, and μ represents the minimum population diversity. D低 It is the membership degree that represents the lowest population diversity, μ D中 It represents a membership degree of moderate population diversity, μ D高 It represents the membership degree that corresponds to the highest level of population diversity.

[0082] Substituting population diversity D or convergence rate C into the above three formulas, we can calculate the membership values ​​for low, medium, and high membership degrees, respectively. The larger the membership value, the higher the membership degree (low, medium, high), thus transforming D and C into fuzzy sets, such as "low", "medium", and "high".

[0083] Membership functions can be used to transform D and C into fuzzy sets, such as "low", "medium", and "high".

[0084] In this embodiment of the invention, the centroid method can be used to convert the membership degrees of fuzzy output variables α and β into specific numerical values, which can then be used to adjust the parameters of the Zebra optimization algorithm. The defuzzification formula can be as follows: (10); (11); in, μ αi and μ βi It corresponds to the membership degree of the fuzzy set. α i and β i This corresponds to the output value.

[0085] In summary, the embodiments of this invention improve the specific application of the Zebra Optimization Algorithm in antenna selection for large-scale MIMO systems, including the algorithm's initialization steps, iterative optimization process, dynamic parameter adjustment mechanism guided by fuzzy logic, and how the algorithm achieves a balance between global and local optimization to avoid local optima. By introducing the Zebra Optimization Algorithm combined with fuzzy logic, global optimization of antenna subsets is achieved, avoiding local optima problems, improving selection accuracy, and reducing computational complexity.

[0086] In an optional embodiment, the step of using a compressed sensing algorithm to estimate channel state information and generate channel state information for the optimal antenna subset may specifically include: Step S31: Use a sparse basis to convert the channel matrix corresponding to the optimal antenna subset into a sparse vector; Step S32: Calculate the sparsity of the channel matrix under the sparse basis based on the number of non-zero elements in the sparse vector and the total length of the sparse vector. Step S33: If the channel matrix meets the compressed sensing conditions based on the sparsity, the channel matrix is ​​sampled using the measurement matrix to obtain the measurement vector; Step S34: Using a compressed sensing algorithm, the original channel matrix is ​​reconstructed from the measurement data to obtain the channel state information of the optimal antenna subset.

[0087] Sparsity analysis is the first step in channel state information estimation. This analysis ensures that the channel matrix has good sparsity under a certain basis, so that compressed sensing technology can be used for efficient sampling and reconstruction.

[0088] In this embodiment of the invention, a sparse basis suitable for the channel matrix can be selected to maximize its sparsity. Commonly used sparse bases include Discrete Cosine Transform (DCT) bases, Discrete Wavelet Transform (DWT) bases, and Fourier Transform bases. DCT bases are suitable for capturing the frequency domain characteristics of the channel, especially when only a few frequency components dominate the channel spectrum. DWT bases are suitable for time-frequency analysis and can effectively capture the multipath characteristics of the channel, especially when multipath effects are significant.

[0089] In some embodiments, the channel matrix H can be transformed to a selected sparse basis to obtain a sparse representation. θ : .in, It is the inverse transformation matrix of the sparse basis.

[0090] In some embodiments, the sparsity k of the channel matrix under a sparse basis can be calculated. This is achieved by calculating the number of non-zero elements k in the sparse representation vector θ: Where k represents a vector θ whose absolute value is greater than a certain threshold. The number of elements (e.g.) ).

[0091] In some embodiments, the sparsity ratio r can be calculated: Where n is the total length of the sparse vector θ, and k is the number of non-zero elements in the sparse representation vector θ.

[0092] In this embodiment of the invention, after the sparsity analysis confirms that the channel matrix has good sparsity, compressed sensing sampling is performed to reduce the amount of sampled data while maintaining high reconstruction accuracy, thereby realizing compressed sensing sampling.

[0093] In some embodiments, a measurement matrix Φ can be designed to sample the channel matrix, i.e., a suitable measurement matrix Φ, such as a Gaussian random matrix or a partial Fourier matrix, can be selected to map the high-dimensional channel matrix to a low-dimensional space. Where y is the sampled low-dimensional measurement vector, and Φ is the measurement matrix. It is the channel matrix. It is a sparse base. It is a sparse representation.

[0094] In some embodiments, a measurement matrix can be used to perform low-dimensional sampling of the channel to obtain sampled data. The signal is sampled using the measurement matrix Φ to generate a measurement vector y. The sampling rate is determined by the channel sparsity and reconstruction accuracy requirements; generally, a lower sampling rate results in lower computational complexity but may affect reconstruction accuracy.

[0095] In this embodiment of the invention, after sampling is completed, the original channel matrix can be reconstructed using compressed sensing technology.

[0096] In some embodiments, the sparse coefficient vector θ can be reconstructed from the sampled data by solving the following optimization problem. : ; Where θ1 is the sparsity constraint, y is the sampled low-dimensional measurement vector, and Φ is the measurement matrix. It is a sparse base. It is a sparse representation.

[0097] In some embodiments, the channel matrix H can be reconstructed using the reconstructed sparse coefficient vector: Channel matrix reconstruction can remap low-dimensional sparse representations back to high-dimensional space, reconstructing the original channel state information matrix.

[0098] In this embodiment of the invention, the reconstruction accuracy of channel state information can be ensured by adjusting the sampling density, sparse basis selection, and measurement matrix design, thereby achieving accuracy control. Specifically, the sampling density is dynamically adjusted to adapt to different channel environments. For example, in cases of strong multipath effects, the sampling density is increased to ensure reconstruction accuracy; sparse basis selection or optimization improves the sparsity of the channel matrix, thereby improving reconstruction quality.

[0099] In this embodiment of the invention, key parameters of compressed sensing can be dynamically adjusted under different channel conditions to balance computational complexity and reconstruction accuracy, thereby achieving fuzzy logic control. Specifically, a fuzzy logic system can be used, taking channel environment parameters (such as signal-to-noise ratio and multipath effect intensity) as input and outputting appropriate sampling density, sparse basis type, and measurement matrix configuration.

[0100] In this embodiment of the invention, post-processing techniques can be used to optimize the reconstructed CSI results, ensuring their stability and accuracy in practical applications, thereby achieving optimized reconstruction results. Specifically, denoising processing can be performed to reduce errors during the reconstruction process and improve the stability of channel estimation; multiple sampling and reconstructions can be performed, and the average value can be taken to reduce random errors.

[0101] In one optional embodiment, calculating the optimal beamforming vector based on the channel state information may specifically include: Step S41: Determine the objective function for maximizing signal gain and interference suppression; Step S42: If the objective function is a convex function, a convex optimization algorithm is used to solve the objective function to obtain the optimal beamforming vector. Step S43: When the objective function is a non-convex function, the gradient descent method is used to solve the objective function to obtain the optimal beamforming vector.

[0102] In this embodiment of the invention, after obtaining accurate CSI, the system enters the beamforming optimization stage, with the goal of achieving optimal signal transmission by adjusting the phase and amplitude of the antenna.

[0103] In some embodiments, an objective function is designed to maximize signal gain and suppress interference in order to optimize the antenna beamforming vector. The objective function formula can be: (12); Where w is the beamforming vector, R s and R iare the covariance matrices of the target signal and the interference signal, respectively, and λ is the regularization parameter used to balance the relationship between signal gain and interference suppression.

[0104] Among them, R s It can be used to calculate the covariance matrix of the target signal based on the reconstructed CSI matrix, ensuring that the beam is directed towards the target user. R i It can be used to calculate the covariance matrix of interference signals based on CSI data, with the aim of minimizing the influence of other interference sources on the target signal. λ can be adjusted according to system requirements to achieve optimal signal gain and interference suppression.

[0105] In this embodiment of the invention, an appropriate optimization algorithm can be used to solve the beamforming vector so that the objective function is maximized.

[0106] In some embodiments, if the objective function J(w) is a convex function, it can be solved directly using a convex optimization algorithm, such as quadratic programming. The optimal beamforming vector can be obtained by solving the following convex optimization problem: (13); (14); in, It is the optimal beamforming vector.

[0107] In some embodiments, for non-convex problems, gradient descent can be used for iterative solutions. The beamforming vector w is initialized, and then the solution is obtained based on the gradient of the objective function. renew: (15); Where η is the learning rate, used to control the update step size. (k) w is the beamforming vector for the current iteration k. (k+1) It is the updated beamforming vector.

[0108] In some embodiments, a learning rate decay strategy can be used, that is, gradually reducing η as the number of iterations increases, to ensure that the algorithm converges to a local or global optimum. A multi-initial-value strategy can be adopted, starting gradient descent from multiple different initial values ​​and taking the final optimal result to increase the probability of a global optimum, thereby ensuring convergence in the beamforming vector calculation process and avoiding getting trapped in local optima.

[0109] This invention combines a fuzzy logic-heuristic-guided zebra optimization algorithm with compressed sensing technology. It intelligently selects the optimal subset of antennas and uses compressed sensing to accurately estimate channel state information, then calculates the beamforming vector that maximizes signal gain based on this data. Simultaneously, the system possesses dynamic adjustment capabilities, automatically adjusting optimization strategies by monitoring communication performance in real time to ensure consistently high performance and resource utilization efficiency under different communication environments.

[0110] Compared to traditional antenna selection and beamforming methods, in high-density user environments (such as stadiums and concert halls), increased interference and intense resource competition may prevent the effective selection of antenna subsets and beamforming, leading to signal quality degradation. This invention addresses this issue by using compressed sensing technology and sparsity analysis to accurately estimate channel state information and combining this with optimization algorithms for antenna selection and beamforming, ensuring optimized resource allocation and interference suppression even in high-density user environments.

[0111] In summary, the embodiments of this invention improve the application of compressed sensing technology in CSI estimation, particularly through a comprehensive scheme combining sparsity analysis and dynamic adjustment strategies. Specifically, this includes methods for selecting the optimal sparse basis, steps for designing the measurement matrix, optimization strategies for the sampling process, and specific implementation methods to improve reconstruction accuracy when channel sparsity is insufficient. By optimizing compressed sensing technology through sparsity analysis and dynamic adjustment, CSI reconstruction accuracy is improved, thereby ensuring the accuracy of channel information even under conditions of insufficient channel sparsity.

[0112] In summary, this invention provides specific technical solutions for beamforming optimization, including the design of the objective function, the selection and solution process of the optimization algorithm, and a specific implementation method for beamforming in complex communication environments using reconstructed CSI data. By designing an optimization algorithm for beamforming, signal transmission quality is improved, signal gain is maximized, and interference is effectively suppressed.

[0113] In one alternative embodiment, a dynamic adjustment and feedback mechanism can be implemented through real-time monitoring and feedback and dynamic adjustment of strategies.

[0114] In some embodiments, system performance can be monitored in real time to ensure optimal performance under various communication environments. Specifically, communication performance indicators such as signal strength, latency, throughput, and bit error rate can be monitored. These data are collected in real time using dedicated hardware or software and transmitted to a central control module for analysis, thereby achieving real-time data acquisition. Based on the collected data, the antenna phase and amplitude can be adjusted in real time, or the beamforming vector can be recalculated. If performance degradation or environmental changes are detected, a feedback control mechanism will automatically trigger an adjustment process to ensure the system adapts to new environmental conditions, thus achieving feedback control.

[0115] In some embodiments, the parameters and strategies of each module can be dynamically adjusted during system operation based on environmental changes and performance requirements. Specifically, different adjustment strategies can be designed to cope with possible environmental changes (such as the Doppler effect caused by mobile users) and changes in performance requirements. Strategies may include: reselecting antenna subsets, adjusting CSI estimation parameters, recalculating beamforming vectors, etc. The system can automatically switch to a suitable adjustment strategy based on monitoring results, implementing a strategy switching mechanism. For example, when an increase in user movement speed is detected, the system switches to a strategy that considers the Doppler effect to adjust the beam direction. The system has adaptive learning capabilities, enabling it to optimize adjustment strategies based on historical data and the current environment, reducing adjustment time and computational overhead.

[0116] In summary, the embodiments of the present invention provide a method for implementing dynamic adjustment and real-time feedback mechanisms in large-scale MIMO systems, including technical means for real-time performance monitoring, strategy design for parameter adjustment, and specific methods for ensuring that the system maintains optimal performance under different environmental conditions through adaptive optimization. Thus, through real-time monitoring and dynamic adjustment, the system can maintain optimal performance in complex communication environments and enhance its adaptability.

[0117] In one specific embodiment, the overall process of beamforming for the large-scale MIMO system provided by this invention can be as follows: Step 1: Initial Setup and Antenna Subset Selection: The initial setup begins by determining the total number of antennas and the size of the desired antenna subset. Next, the optimal antenna subset is selected using the Zebra Optimization Algorithm. This algorithm generates multiple initial antenna subsets and iteratively optimizes to find the subset that maximizes signal gain, minimizes energy consumption, and has low computational complexity. The entire process is guided by fuzzy logic heuristics to ensure an optimal balance between global and local searches.

[0118] Step 2, Sparsity Analysis: After determining the antenna subset, sparsity analysis is performed on its channel matrix. An appropriate sparse basis (such as a DCT basis or wavelet basis) is selected to convert the channel matrix into a sparse representation. Next, the sparsity of the channel matrix is ​​calculated to determine its suitability for compressed sensing technology. If the sparsity is insufficient, the sparse basis is adjusted or the channel matrix is ​​preprocessed to improve its sparsity, thereby ensuring the effectiveness of subsequent sampling and reconstruction.

[0119] Step 3, Compressed Sensing Sampling: Based on the sparsity analysis results, a measurement matrix is ​​designed and compressed sensing sampling is performed on the channel matrix. The measurement matrix compresses the channel matrix from a high-dimensional space to a low-dimensional space, generating measurement data. The sampling rate is adjusted according to the sparsity of the channel matrix and the desired reconstruction accuracy to maintain sufficient reconstruction accuracy while reducing the amount of data.

[0120] Step 4: Channel State Information Reconstruction: After sampling, the channel matrix is ​​reconstructed from the measurement data using compressed sensing algorithms to ensure that the reconstructed channel state information has high accuracy, providing reliable basic data for subsequent beamforming optimization.

[0121] Step 5, Beamforming Optimization: Using the reconstructed CSI, an objective function for beamforming is designed to maximize signal gain and suppress interference. Through optimization algorithms (such as convex optimization or gradient descent), the system calculates the optimal beamforming vector and adjusts the antenna's phase and amplitude based on this vector to achieve optimal signal transmission. The entire process ensures maximum signal transmission in the predetermined direction and minimizes interference.

[0122] Step 6: Dynamic Adjustment and Feedback: During system operation, a dynamic adjustment and feedback mechanism monitors communication performance in real time, including metrics such as signal strength, latency, and throughput. Based on the monitoring results, the system can dynamically adjust antenna selection, CSI estimation, and beamforming parameters to ensure optimal performance under different environments and requirements. Through adaptive adjustment strategies, it can flexibly respond to environmental changes and continuously optimize signal transmission quality.

[0123] Existing large-scale MIMO systems primarily employ fixed heuristic methods for antenna selection, typically optimizing the selection of the optimal antenna subset based on channel capacity or signal-to-interference plus noise ratio (SINR). While these methods offer some optimization in static or slowly varying channel environments, their performance optimization is significantly limited in dynamically changing communication environments. Regarding channel estimation, most of these systems rely on traditional linear estimation techniques, such as Minimum Mean Squared Error (MMSE). These techniques require prior knowledge of the environment and assume the channel remains stable over a certain period. Therefore, they may fail to accurately estimate the channel state under rapidly changing channel conditions, impacting the overall system's communication quality. Beamforming techniques mainly use mathematical optimization methods, such as Zero Forcing (ZF) and MMSE algorithms, to adjust beam weights, aiming to maximize the received quality of the target signal. However, these techniques often fix beamforming parameters and do not consider real-time changes in channel state. Especially in multipath and rapidly changing channel environments, their beamforming effect is limited, making it difficult to effectively suppress interference. Therefore, existing technologies have the following drawbacks: (1) Insufficient adaptability: Traditional systems usually rely on static or semi-static optimization methods, such as maximizing channel capacity or optimizing signal-to-interference-plus-noise ratio (SINR). These methods are difficult to adjust in time in scenarios where the channel state changes rapidly, resulting in performance degradation.

[0124] (2) Low channel estimation accuracy: Existing channel estimation techniques (such as MMSE) assume that the channel is stable. For multipath and rapidly changing channel environments, these estimation methods may not be able to accurately capture the real-time state of the channel, affecting beamforming and system performance.

[0125] (3) Fixed beamforming strategy: Beamforming strategy is often fixed and does not take into account the dynamic changes of the environment. This may lead to uneven signal coverage and interference problems in multi-user and high mobility environments.

[0126] (4) The system optimization is not flexible enough: Most of the optimization strategies of the existing systems are preset and periodic, lacking a mechanism for dynamic adjustment based on real-time data. This may lead to unsatisfactory resource allocation and response delay in the ever-changing communication network.

[0127] To address the aforementioned problems, this invention utilizes the Zebra Optimization Algorithm (ZOA) combined with fuzzy logic to achieve a more flexible and responsive antenna selection strategy, adapting to real-time changes in channel state, thereby realizing dynamic antenna selection. By combining compressed sensing technology for accurate estimation of Channel State Information (CSI), the estimation accuracy and adaptability of the system under non-sparse channel conditions are improved, thus achieving high-precision channel estimation. Dynamic beamforming optimization based on real-time CSI data is introduced to maximize signal gain and effectively suppress interference, especially in multipath and rapidly changing environments, achieving adaptive beamforming. Through real-time monitoring and data feedback mechanisms, system configuration and optimization strategies are dynamically adjusted to improve overall communication performance and resource utilization efficiency, thereby achieving real-time system optimization. Specifically, this invention has the following advantages: 1. Enhanced adaptability in dynamic communication environments. This invention achieves global optimization selection of antenna subsets by introducing the Zebra Optimization Algorithm combined with fuzzy logic heuristics, and dynamically adjusts optimization parameters during the selection process. This technique enables the system to adaptively respond to dynamically changing communication environments, avoiding the limitations of fixed heuristic methods in dynamic environments, thus providing superior performance in complex communication environments.

[0128] 2. Improved CSI estimation accuracy under rapidly changing channel conditions. This invention employs an improved compressed sensing technique in CSI estimation, selecting the optimal sparse basis and optimizing the sampling strategy through detailed sparsity analysis and dynamic adjustment strategies. This enables the system to achieve high-precision CSI reconstruction even under conditions of insufficient channel sparsity or rapid channel changes, significantly improving the accuracy of channel state information and overcoming the dependence of traditional linear estimation techniques on channel stability and prior knowledge.

[0129] 3. Real-time optimization of beamforming effect. This invention applies an optimization algorithm during the beamforming process, performing real-time optimization by designing a specific objective function and combining it with CSI data. Unlike the fixed beamforming parameters in existing technologies, the technical means of this invention can dynamically adjust beam weights to maximize signal reception quality, especially in multipath and rapidly changing channel environments, effectively suppressing interference and improving beamforming effect.

[0130] 4. Enhanced System Robustness and Adaptability. This invention integrates dynamic adjustment and real-time feedback mechanisms. By monitoring communication performance in real time, the system can adaptively adjust antenna selection, CSI estimation, and beamforming parameters according to environmental changes. This mechanism enhances the system's robustness and adaptability, enabling it to maintain optimal performance in various complex communication environments without relying on fixed parameters or prior knowledge of the environment.

[0131] The beamforming device for a large-scale MIMO system provided by the present invention is described below. The beamforming device for a large-scale MIMO system described below can be referred to in correspondence with the beamforming method for a large-scale MIMO system described above.

[0132] Figure 2 This is one of the structural schematic diagrams of the beamforming device for a large-scale MIMO system provided in an embodiment of the present invention. (Refer to...) Figure 2 This invention provides a beamforming device for a large-scale MIMO system, which may specifically include the following modules: Antenna subset selection module 210 is used to select the optimal antenna subset from multiple antennas in a large-scale MIMO system using the Zebra optimization algorithm; wherein, the parameters of the Zebra optimization algorithm are adjusted in real time based on fuzzy logic heuristic guidance. The channel state information estimation module 220 is used to estimate the channel state information using a compressed sensing algorithm to generate the channel state information of the optimal antenna subset; The beamforming optimization module 230 is used to calculate the optimal beamforming vector based on the channel state information. The control module 240 is used to control the phase and amplitude of the antenna array corresponding to the optimal antenna subset based on the optimal beamforming vector.

[0133] This invention employs a zebra optimization algorithm guided by fuzzy logic to select an optimal subset of antennas from multiple antennas in a large-scale MIMO system. It then uses compressed sensing to estimate channel state information (CSO) for this optimal subset, generating CSO data for that subset. Based on this CSO data, it calculates the optimal beamforming vector, thereby controlling the phase and amplitude of the antenna array corresponding to the optimal subset. By using this fuzzy logic-guided zebra optimization algorithm for antenna selection, this invention enables dynamic adjustment of antenna selection, leading to adaptive optimization of beamforming. This maintains high performance and resource utilization efficiency under various communication environments.

[0134] Figure 3 This is the second schematic diagram of the beamforming device for a large-scale MIMO system provided in this embodiment of the invention. (Refer to...) Figure 3 This invention provides another beamforming device for a large-scale MIMO system, which may specifically include the following modules: Antenna subset selection module 310 is used to select an optimal subset of antennas from a large number of antennas in a massive MIMO system, aiming to reduce computational complexity while ensuring system performance.

[0135] The antenna subset selection module may include an initial population generation unit, a fuzzy logic control unit, an objective function calculation unit, and an optimization execution unit. The initial population generation unit generates multiple antenna subsets as the initial population, ensuring broad coverage of the search space. The fuzzy logic control unit dynamically adjusts the Zebra Optimization algorithm's search strategy based on a fuzzy rule base and fuzzy inference mechanism, enhancing the algorithm's global search capability and adaptability. The objective function calculation unit calculates the objective function value for each antenna subset by considering signal gain, energy consumption, and computational complexity to evaluate its performance. The optimization execution unit executes the Zebra Optimization algorithm, progressively optimizing and selecting the final optimal antenna subset.

[0136] The channel state information estimation module 320 is used to accurately estimate the channel state information (CSI) of the selected antenna subset using compressed sensing technology, and generate high-precision CSI data as the basis for beamforming.

[0137] The channel state information estimation module may include a sparsity analysis unit, a compressed sensing sampling unit, a signal reconstruction unit, and a precision control unit. The sparsity analysis unit analyzes the sparsity of the channel matrix and selects a suitable sparse basis. The compressed sensing sampling unit performs low-dimensional sampling of the channel matrix, reducing the amount of sampled data while maintaining the integrity of key channel information. The signal reconstruction unit reconstructs high-dimensional channel state information from the low-dimensional data generated by the compressed sensing sampling unit. The precision control unit adjusts parameters in real time during the compressed sensing sampling and channel reconstruction processes to cope with dynamic changes in the channel environment.

[0138] The beamforming optimization module 330 is used to optimize the antenna beamforming vector using accurate CSI data to maximize signal gain and suppress interference, thereby ensuring the communication quality and overall performance of the system.

[0139] The beamforming optimization module may include an objective function design unit, a beamforming vector calculation unit, and an antenna activation control unit. The objective function design unit designs the objective function based on the reconstructed CSI data, comprehensively considering factors such as signal gain, interference suppression, and energy efficiency. The beamforming vector calculation unit uses optimization algorithms to calculate the optimal beamforming vector, determining the antenna's transmission direction and power allocation. The antenna activation control unit activates the corresponding antenna according to the calculated beamforming vector, achieving optimal signal transmission.

[0140] The dynamic adjustment and feedback mechanism module 340 is used to monitor the communication performance parameters of the system in real time, and dynamically adjust the parameters and strategies of each module based on this feedback information to ensure the optimal performance of the system under different communication environments.

[0141] The dynamic adjustment and feedback mechanism module may include a dynamic adjustment unit and a feedback control unit. The dynamic adjustment unit can be used to dynamically adjust relevant parameters and strategies at each stage of antenna subset selection, CSI estimation, and beamforming to adapt to environmental changes. The feedback control unit can be used to optimize loops by monitoring communication performance parameters (such as signal quality, delay, and throughput) in real time, ensuring the system's adaptive optimization capability.

[0142] Based on the above modules, the beamforming method for a large-scale MIMO system provided in this embodiment of the invention can be described as follows: First, the antenna subset selection module uses the Zebra Optimization Algorithm (ZOA) guided by fuzzy logic heuristics to select an optimal subset of antennas from the many antennas in the large-scale MIMO system, with the aim of reducing computational complexity while ensuring system performance.

[0143] Subsequently, this subset of antennas is passed to the CSI estimation module, which uses compressed sensing technology to sparsely sample and accurately reconstruct the channel state information of the selected antennas, generating high-precision CSI data, which is crucial for beamforming.

[0144] Next, the beamforming optimization module designs and optimizes the objective function based on CSI data, and calculates the beamforming vector that maximizes signal gain and suppresses interference, thereby ensuring the signal transmission quality and overall communication performance of the system.

[0145] Finally, the dynamic adjustment and feedback mechanism module monitors the system's communication performance parameters in real time, such as signal quality, latency, and throughput, and dynamically adjusts the parameters and strategies of each module based on this feedback information to adapt to the constantly changing communication environment. For example, when a deterioration in the communication environment is detected, the system may adaptively adjust by increasing the CSI sampling density or re-optimizing the antenna selection strategy.

[0146] Through the coordinated operation of these modules, this invention enables efficient antenna selection and beamforming optimization for large-scale MIMO systems in complex communication environments, thereby maximizing system performance and improving resource utilization efficiency.

[0147] Figure 4 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 4As shown, the electronic device may include a processor 410, a communications interface 420, a memory 430, and a communication bus 440, wherein the processor 410, communications interface 420, and memory 430 communicate with each other via the communication bus 440. The processor 410 can call logic instructions in the memory 430 to execute a beamforming method for a massive MIMO system. This method includes: using a zebra optimization algorithm to select an optimal subset of antennas from multiple antennas in the massive MIMO system; wherein the parameters of the zebra optimization algorithm are adjusted in real time based on fuzzy logic heuristics; using a compressed sensing algorithm to estimate channel state information and generate channel state information for the optimal antenna subset; calculating an optimal beamforming vector based on the channel state information; and controlling the phase and amplitude of the antenna array corresponding to the optimal antenna subset based on the optimal beamforming vector.

[0148] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0149] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the beamforming method for a massive MIMO system provided by the above methods. The method includes: using a zebra optimization algorithm to select an optimal subset of antennas from multiple antennas in a massive MIMO system; wherein the parameters of the zebra optimization algorithm are adjusted in real time based on fuzzy logic heuristic guidance; using a compressed sensing algorithm to estimate channel state information and generate channel state information for the optimal subset of antennas; calculating an optimal beamforming vector based on the channel state information; and controlling the phase and amplitude of the antenna array corresponding to the optimal subset of antennas based on the optimal beamforming vector.

[0150] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements a beamforming method for a massive MIMO system provided by the methods described above. This method includes: employing a zebra optimization algorithm to select an optimal subset of antennas from multiple antennas in a massive MIMO system; wherein the parameters of the zebra optimization algorithm are adjusted in real time based on fuzzy logic heuristics; employing a compressed sensing algorithm to estimate channel state information and generate channel state information for the optimal antenna subset; calculating an optimal beamforming vector based on the channel state information; and controlling the phase and amplitude of the antenna array corresponding to the optimal antenna subset based on the optimal beamforming vector.

[0151] The device embodiments described above are merely illustrative. 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 modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0152] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0153] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A beamforming method for a large-scale MIMO system, characterized in that, include: The Zebra optimization algorithm is used to select the optimal subset of antennas from multiple antennas in a large-scale MIMO system; wherein the parameters of the Zebra optimization algorithm are adjusted in real time based on fuzzy logic heuristics. Channel state information is estimated using a compressed sensing algorithm to generate the channel state information of the optimal antenna subset; Based on the channel state information, the optimal beamforming vector is calculated; Based on the optimal beamforming vector, the phase and amplitude of the antenna array corresponding to the optimal antenna subset are controlled; The step of employing the Zebra Optimization Algorithm to select the optimal subset of antennas from multiple antennas in a large-scale MIMO system includes: The Zebra Optimization Algorithm is used to randomly select a preset number of antennas from multiple antennas in a large-scale MIMO system, generate multiple solutions to form an initial population, and iteratively optimize the system step by step; where each solution represents a subset of antennas. In each iteration, the solution with the largest fitness value in the current population is determined as the current optimal solution; Based on the current optimal solution and random perturbation, other solutions in the current population are updated, and the preset number of solutions with the largest fitness values ​​are selected to form a new population. Stop iterating when the algorithm converges and output the optimal subset of antennas; The parameters of the Zebra optimization algorithm are adjusted in real time in the following way: During each iteration, the diversity and convergence rate of the current population are calculated; Based on the diversity and convergence speed of the current population, fuzzy reasoning is performed to obtain the membership degree of the search range adjustment coefficient and the membership degree of the perturbation amplitude adjustment coefficient; the search range adjustment coefficient is used to control the strength of the current solution towards the optimal solution, and the perturbation amplitude adjustment coefficient is used to control the intensity of the introduced random perturbation. The membership degrees of the search range adjustment coefficient and the perturbation amplitude adjustment coefficient are defuzzified to obtain the membership quantification values ​​of the search range adjustment coefficient and the perturbation amplitude adjustment coefficient. The parameters of the Zebra optimization algorithm are adjusted in real time based on the membership quantification values ​​of the search range adjustment coefficient and the perturbation amplitude adjustment coefficient.

2. The beamforming method for a large-scale MIMO system according to claim 1, characterized in that, The step of determining the solution with the largest fitness value in the current population as the current optimal solution includes: For each solution in the current population, calculate the total signal gain based on the gain of each antenna; Calculate the total energy consumption based on the energy consumption of each antenna; Computational complexity based on the number of antennas; The fitness value of each solution is obtained by weighting the total signal gain, the total energy consumption, and the complexity. The solution with the highest fitness value in the current population is determined as the current optimal solution.

3. The beamforming method for a large-scale MIMO system according to claim 1, characterized in that, The step of updating other solutions in the current population based on the current optimal solution and the random perturbation includes: Calculate the random disturbance based on the disturbance amplitude control coefficient and the random vector; The other solutions in the current population are updated based on the search range adjustment coefficient, the current optimal solution, the perturbation amplitude adjustment coefficient, and the random perturbation.

4. The beamforming method for a large-scale MIMO system according to claim 1, characterized in that, The calculation of the diversity and convergence rate of the current population includes: The diversity of the current population is calculated based on the absolute value of the difference between each solution in the current population and the mean of all solutions in each dimension, the number of solutions in the current population, and the preset number. The convergence rate of the current population is calculated based on the optimal fitness value in the current iteration and the optimal fitness value in the previous iteration.

5. The beamforming method for a large-scale MIMO system according to claim 1, characterized in that, The fuzzy inference based on the diversity and convergence speed of the current population, to obtain the membership degrees of the search range adjustment coefficient and the perturbation amplitude adjustment coefficient, includes: Using a triangular membership function, the diversity of the current population is fuzzed based on the highest, lowest, and medium diversity, to obtain the membership degree of the search range adjustment coefficient of the current population. Using triangular membership functions, the convergence speed of the current population is fuzzified based on the highest, lowest, and medium convergence speeds to obtain the membership degree of the perturbation amplitude adjustment coefficient of the current population.

6. The beamforming method for a large-scale MIMO system according to claim 1, characterized in that, The step of estimating channel state information using compressed sensing algorithm to generate channel state information for the optimal antenna subset includes: The channel matrix corresponding to the optimal antenna subset is converted into a sparse vector using a sparse basis; Based on the number of non-zero elements in the sparse vector and the total length of the sparse vector, the sparsity of the channel matrix under the sparse basis is calculated. If the channel matrix is ​​determined to meet the compressed sensing conditions based on the sparsity, the channel matrix is ​​sampled using a measurement matrix to obtain a measurement vector; The original channel matrix is ​​reconstructed from the measurement data using a compressed sensing algorithm, thereby obtaining the channel state information of the optimal antenna subset.

7. The beamforming method for a large-scale MIMO system according to claim 1, characterized in that, The step of calculating the optimal beamforming vector based on the channel state information includes: Determine the objective function for maximizing signal gain and interference suppression; When the objective function is a convex function, a convex optimization algorithm is used to solve the objective function to obtain the optimal beamforming vector; When the objective function is non-convex, the gradient descent method is used to solve the objective function to obtain the optimal beamforming vector.

8. A beamforming device for a massive MIMO system, characterized in that, include: The antenna subset selection module is used to select the optimal antenna subset from multiple antennas in a large-scale MIMO system using the Zebra optimization algorithm; wherein the parameters of the Zebra optimization algorithm are adjusted in real time based on fuzzy logic heuristic guidance. The channel state information estimation module is used to estimate the channel state information using a compressed sensing algorithm to generate the channel state information of the optimal antenna subset. The beamforming optimization module is used to calculate the optimal beamforming vector based on the channel state information. The control module is used to control the phase and amplitude of the antenna array corresponding to the optimal antenna subset based on the optimal beamforming vector. The step of employing the Zebra Optimization Algorithm to select the optimal subset of antennas from multiple antennas in a large-scale MIMO system includes: The Zebra Optimization Algorithm is used to randomly select a preset number of antennas from multiple antennas in a large-scale MIMO system, generate multiple solutions to form an initial population, and iteratively optimize the system step by step; where each solution represents a subset of antennas. In each iteration, the solution with the largest fitness value in the current population is determined as the current optimal solution; Based on the current optimal solution and random perturbation, other solutions in the current population are updated, and the preset number of solutions with the largest fitness values ​​are selected to form a new population. Stop iterating when the algorithm converges and output the optimal subset of antennas; The parameters of the Zebra optimization algorithm are adjusted in real time in the following way: During each iteration, the diversity and convergence rate of the current population are calculated; Based on the diversity and convergence speed of the current population, fuzzy reasoning is performed to obtain the membership degree of the search range adjustment coefficient and the membership degree of the perturbation amplitude adjustment coefficient; the search range adjustment coefficient is used to control the strength of the current solution towards the optimal solution, and the perturbation amplitude adjustment coefficient is used to control the intensity of the introduced random perturbation. The membership degrees of the search range adjustment coefficient and the perturbation amplitude adjustment coefficient are defuzzified to obtain the membership quantification values ​​of the search range adjustment coefficient and the perturbation amplitude adjustment coefficient. The parameters of the Zebra optimization algorithm are adjusted in real time based on the membership quantification values ​​of the search range adjustment coefficient and the perturbation amplitude adjustment coefficient.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the beamforming method for the massive MIMO system as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the beamforming method for the massive MIMO system as described in any one of claims 1 to 7.

11. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the beamforming method for the massive MIMO system as described in any one of claims 1 to 7.

Images