Uplink-based beam weight determination method and system

Abstract

The application relates to an uplink-based beam weight determination method, which comprises the following steps: obtaining user equipment measurement data of a target cell and a target cell adjacent area, the user equipment measurement data comprising uplink path loss, time advance and spatial orientation information of a user; constructing an equivalent uplink distance model according to the uplink path loss, the time advance and the spatial orientation information; determining target beam weights of base station antennas based on the equivalent uplink distance model and a gradient-aware particle swarm optimization algorithm, and issuing the target beam weights to a base station antenna array. Through the application, the problem that poor uplink quality is caused by inaccurate beam weights is solved, optimization is directly carried out based on uplink original data, and deviation of uplink performance caused by optimization of downlink data is avoided. The equivalent uplink distance model is constructed, and single optimization dimension and incapability of coping with complex uplink requirements are avoided.

Description

Technical Field

[0001] This application relates to the field of mobile communication technology, and in particular to a method and system for determining beam weights based on the uplink. Background Technology

[0002] With the development of mobile communication technology, network traffic, especially uplink traffic (such as ultra-high-definition video streaming, large file uploads, and IoT terminal data collection), has grown rapidly. The uplink has become a key bottleneck for system capacity and user experience, and automatic antenna mode control technology is a crucial means to optimize network performance.

[0003] Most existing technologies are optimized based on downlink measurement reports (such as RSRP and SINR), which implicitly assume that uplink and downlink channels are reciprocal. However, in real-world networks, due to user equipment (UE) transmit power limitations and uplink-specific interference scenarios, the characteristics of uplink and downlink channels are not entirely consistent. This means that optimization results based on downlink measurements may fail in the uplink, making it impossible to accurately address issues such as weak uplink coverage and high interference. Furthermore, existing solutions mainly rely on signal strength information, lacking in-depth utilization of user spatial location and channel fading characteristics, resulting in a single optimization dimension and difficulty in achieving refined beam management. Summary of the Invention

[0004] This application provides a method, system, electronic device, and storage medium for determining uplink beam weights, in order to at least solve the problem of poor uplink quality caused by inaccurate beam weights in related technologies.

[0005] In a first aspect, embodiments of this application provide a method for determining beam weights based on the uplink, the method comprising:

[0006] Acquire user equipment measurement data of the target cell and neighboring cells of the target cell. The user equipment measurement data includes uplink path loss, timing advance, and user spatial location information.

[0007] Based on the uplink path loss, the timing advance, and the spatial orientation information, an equivalent uplink distance model is constructed.

[0008] With the goal of optimizing the statistical value of the equivalent uplink distance, the target beam weight of the base station antenna is determined based on the equivalent uplink distance model and the gradient-aware particle swarm optimization algorithm, and the target beam weight is sent down to the base station antenna array.

[0009] In some embodiments, constructing an equivalent uplink distance model based on the uplink path loss, the timing advance, and the spatial orientation information includes:

[0010] The uplink path loss is normalized to obtain the normalized path loss.

[0011] The time advance is normalized to obtain the normalized time advance.

[0012] Based on the spatial orientation information and the main lobe direction of the current antenna beam, the beam spatial mismatch is calculated.

[0013] The normalized path loss, the normalized timing advance, and the beam spatial mismatch are weighted and fused to generate the equivalent uplink distance model.

[0014] In some embodiments, the spatial orientation information includes horizontal angle of arrival and vertical angle of arrival; the equivalent uplink distance model includes:

[0015]

[0016] Among them, D eq For the equivalent uplink distance, PL norm For the normalized uplink path loss, TA norm As the time lead, G(θ) h ,θ v ) is the beam space mismatch function, θ h θ is the horizontal angle of arrival. v α is the vertical angle of arrival, and β, γ are weighting coefficients.

[0017] In some embodiments, the objective of optimizing the statistics of the equivalent uplink distance includes:

[0018] The objective is to minimize the average equivalent uplink distance for all users; or

[0019] The goal is to maximize the equivalent uplink distance for edge users at a preset percentile.

[0020] In some embodiments, the method further includes:

[0021] The performance indicators of the base station antenna array after operation based on the target beam weight are monitored, including the uplink signal-to-interference-noise ratio and throughput.

[0022] The performance index is evaluated based on a preset threshold. The evaluation result determines whether the performance corresponding to the target beam weight is better than the historical beam weight. If so, the target beam weight is continued to be used; otherwise, it reverts to the historical beam weight.

[0023] In some embodiments, determining the target beam weights of the base station antenna based on the equivalent uplink distance model and gradient-aware particle swarm optimization algorithm includes:

[0024] By incorporating the gradient-aware term into the velocity update of the standard particle swarm optimization algorithm, a velocity update model is obtained.

[0025] Based on the speed update model and the equivalent uplink distance model, the target beam weights of the base station antenna are determined.

[0026] In some embodiments, the speed update model includes:

[0027]

[0028] Among them, V i (t+1) is the velocity vector of particle i at time t+1, ω is the inertial weight, and V i (t) is the velocity vector of particle i at time t, c1 is the individual learning factor, r1 is the first random number uniformly distributed in the range [0, 1], and P best_i X is the historical best position found by particle i itself. i (t) is the position vector of particle i at time t, c2 is the social learning factor, r2 is the second random number uniformly distributed in the range [0, 1], and P best It is the globally optimal position found by the entire population, c3 is the gradient acceleration constant, and G is the global historical optimal position found by the entire population. i (t) is the gradient-aware term.

[0029] Secondly, embodiments of this application provide an uplink-based beam weight determination system, the system comprising:

[0030] The data acquisition module is used to acquire user equipment measurement data of the target cell and neighboring cells of the target cell. The user equipment measurement data includes uplink path loss, timing advance, and user spatial location information.

[0031] The model building module is used to construct an equivalent uplink distance model based on the uplink path loss, the time lead, and the spatial orientation information.

[0032] The weight determination module is used to determine the target beam weight of the base station antenna based on the equivalent uplink distance model and the gradient-aware particle swarm optimization algorithm, with the goal of optimizing the statistical value of the equivalent uplink distance, and then send the target beam weight to the base station antenna array.

[0033] Thirdly, embodiments of this application provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the uplink-based beam weight determination method as described in the first aspect above.

[0034] Fourthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the uplink-based beam weight determination method as described in the first aspect above.

[0035] Compared to related technologies, the uplink-based beam weight determination method provided in this application directly optimizes based on native uplink data, avoiding the deviation caused by using downlink data to optimize uplink performance. This ensures that subsequent optimization measures accurately match uplink service requirements, solving the problem of poor uplink quality due to inaccurate beam weights. Constructing an equivalent uplink distance model simultaneously balances uplink signal quality, transmission delay, and beam matching, avoiding the inability to address complex uplink demands due to a single optimization dimension. Employing a gradient-aware particle swarm optimization algorithm avoids getting stuck in local optima in the early stages and ensures rapid convergence to the optimal solution in later stages, improving optimization efficiency and accuracy. Attached Figure Description

[0036] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0037] Figure 1 This is a flowchart of a beam weight determination method based on the uplink according to an embodiment of this application;

[0038] Figure 2 This is a simulation diagram of the optimization effect of the GS-PSO algorithm according to an embodiment of this application;

[0039] Figure 3 This is a flowchart of a weight optimization process according to an embodiment of this application;

[0040] Figure 4 This is a structural block diagram of an uplink-based beam weight determination system according to an embodiment of this application;

[0041] Figure 5 This is a schematic diagram of the internal structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0042] To make the objectives, technical solutions, and advantages of this application clearer, the application is described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application.

[0043] Obviously, the accompanying drawings described below are merely some examples or embodiments of this application. Those skilled in the art can apply this application to other similar scenarios based on these drawings without any inventive effort. Furthermore, it is understood that although the efforts made in this development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, any changes to design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as insufficient disclosure of the content of this application.

[0044] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that is mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0045] Unless otherwise defined, the technical or scientific terms used in this application shall have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms “a,” “an,” “an,” “the,” and similar words used in this application do not indicate quantity limitation and may indicate singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units not listed, or may include other steps or units inherent to these processes, methods, products, or devices. The terms “connected,” “linked,” “coupled,” and similar words used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. “Multiple” used in this application refers to two or more. “And / or” describes the relationship between related objects, indicating that three relationships may exist; for example, “A and / or B” can represent: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following objects are in an "or" relationship. The terms "first," "second," and "third" used in this application are merely to distinguish similar objects and do not represent a specific ordering of the objects.

[0046] This embodiment provides a method for determining beam weights based on the uplink. Figure 1 This is a flowchart of the uplink-based beam weight determination method according to an embodiment of this application, as shown below. Figure 1As shown, the process includes the following steps:

[0047] Step S101: Obtain user equipment measurement data of the target cell and neighboring cells of the target cell. The user equipment measurement data includes uplink path loss, timing advance, and user spatial location information.

[0048] Real-time acquisition and fusion of multi-dimensional user equipment (UE) measurement data from the serving cell (target cell) and neighboring cells. Serving cell data includes reference received power (RSRP), angle of arrival (DOA, including horizontal and vertical), uplink path loss, and timing advance (TA); neighboring cell data includes RSRP, horizontal angle of arrival (HDOA), and vertical angle of arrival (VDOA).

[0049] Because the uplink and downlink path losses are consistent, the uplink path loss calculation formula is as follows:

[0050]

[0051] Among them, UL path loss For uplink path loss, P Tx_BS P is the reference signal power transmitted by the base station. Rx_UE The power transmitted to the terminal (RSRP).

[0052] Step S102: Construct an equivalent uplink distance model based on uplink path loss, time lead, and spatial orientation information.

[0053] Based on multidimensional data, an innovative equivalent uplink distance (D) is constructed. eq The model integrates information such as uplink path loss, lead time, and user spatial location into a unified optimization metric. This algorithm abandons the traditional optimization approach that targets a single metric (such as RSRP) and instead adopts a new, comprehensive "equivalent uplink distance" model to drive the AI ​​algorithm, achieving more intelligent uplink optimization.

[0054] In some embodiments, step S102 specifically includes:

[0055] Step S1021: Normalize the uplink path loss to obtain the normalized path loss.

[0056] Normalized path loss is used to eliminate absolute differences in path loss across different base stations and environments, mapping the path loss of all users to a relatively comparable scale. Normalized path loss PL norm The calculation formula is:

[0057]

[0058] Among them, PL max and PLmin The system presets the maximum and minimum path loss values. For example, PL min The expected optimal loss can be set (e.g., 80dB), PL max It can be set to the maximum loss that the system can tolerate (e.g., 140dB).

[0059] Step S1022: Normalize the time advance to obtain the normalized time advance.

[0060] Normalized time lead is used to convert time lead (TA) into a dimensionless indicator reflecting relative distance. norm The calculation formula is:

[0061]

[0062] Among them, TA min Typically 0 or a very small value, representing users near the base station. TA max This is the TA value corresponding to the cell coverage radius. Users exceeding this value may be located at the cell edge.

[0063] Step S1023: Calculate the beam spatial mismatch based on the spatial orientation information and the main lobe direction of the current antenna beam.

[0064] Beam spatial mismatch is primarily used to quantify the degree of deviation between the user's current location and the main lobe direction of the antenna beam. Beam spatial mismatch G(θ) h ,θ v The formula for calculating ) is:

[0065]

[0066] Among them, ф h ф v These are the horizontal and vertical azimuth angles of the main lobe of the current antenna beam, respectively, determined by the weights the algorithm is currently trying. (BW) h BW v The horizontal and vertical widths of the antenna beam can be fixed parameters or controlled by weights.

[0067] Calculate the angular difference between the user's angle and the beam center, and normalize it using the beamwidth. The greater the beam mismatch, the further the user is from the beam center, and the worse the channel conditions.

[0068] Step S1024: The normalized path loss, normalized timing advance, and beam space mismatch are weighted and fused to generate an equivalent uplink distance model.

[0069] In some embodiments, spatial orientation information includes horizontal angle of arrival and vertical angle of arrival; the equivalent uplink distance model includes:

[0070]

[0071] Among them, D eq Equivalent uplink distance, a dimensionless relative value. The larger the value, the higher the uplink transmission difficulty for the user and the greater the required base station receiving gain.

[0072] PL norm This is the normalized uplink path loss, in dB. Uplink path loss is the power attenuation of the signal during its transmission from the UE to the base station, and it is the most important factor determining uplink quality.

[0073] TAnorm is the time advance, reflecting the propagation delay of radio waves from the UE to the base station, and is positively correlated with physical distance. The larger the value, the greater the time delay. eq The larger the value.

[0074] G(θ h ,θ v ) is the beam space mismatch function, θ h θ is the horizontal angle of arrival. v α represents the vertical angle of arrival. α, β, and γ are weighting coefficients that determine whether the algorithm prioritizes path loss distance or beam alignment.

[0075] This model function calculates the degree of match between the current antenna beam pattern and the user's actual location. If the user is located exactly in the center gain region of the main lobe of the beam, the function returns a small value (e.g., 0); if the user is at the edge or side lobe of the beam, the return value is large. This simulates a scenario where "the user is close, but not aligned with the beam, making communication difficult."

[0076] The coefficients α, β, and γ need to be trained with a large amount of historical data or dynamically adjusted according to network strategies. They are used to balance the weights of path loss, latency, and beam mismatch on the overall "equivalent distance". Their values ​​can be obtained through training with historical data (e.g., using machine learning regression analysis to analyze the impact of each factor on uplink throughput) or dynamically set by network strategies (e.g., in dense urban areas, beam alignment is more important, so the weight of γ is increased; in wide coverage scenarios, path loss and distance are the main issues, so the weights of α and β are increased).

[0077] Step S103: With the goal of optimizing the statistical value of the equivalent uplink distance, the target beam weight of the base station antenna is determined based on the equivalent uplink distance model and the gradient-aware particle swarm optimization algorithm, and the target beam weight is sent to the base station antenna array.

[0078] The optimization goal of this embodiment is not D for a single user. eq Instead, it is all users D eq The statistical value.

[0079] In some embodiments, optimizing the statistics of the equivalent uplink distance includes:

[0080] The objective is to minimize the average equivalent uplink distance for all users.

[0081] The formula for the first objective function, Objective1 (to improve overall performance), is as follows:

[0082]

[0083] Minimizing the average equivalent uplink distance for all users means that the algorithm aims to universally improve the uplink channel quality for all users across the network.

[0084] The goal is to maximize the equivalent uplink distance for edge users at a preset percentile.

[0085] The formula for the second objective function, Objective2 (improving fairness), is as follows:

[0086]

[0087] Maximize the D of users at the preset percentile (e.g., the 5th percentile). eq This means that the algorithm is making every effort to improve the performance of the 5% of users with the worst experience, aiming to "provide a safety net" and enhance network fairness.

[0088] In some embodiments, step S103, which determines the target beam weights of the base station antenna based on the equivalent uplink distance model and gradient-aware particle swarm optimization algorithm, includes:

[0089] Step S1031: Integrate the gradient-aware term into the velocity update of the standard particle swarm optimization algorithm to obtain the velocity update model.

[0090] Step S1032: Determine the target beam weights of the base station antenna based on the speed update model and the equivalent uplink distance model.

[0091] Gradient-aware PSO (GA-PSO) is deeply integrated with the "equivalent uplink distance model". By utilizing the gradient information provided by the equivalent uplink distance model, the search direction of the PSO algorithm is intelligently guided, thereby achieving more efficient and accurate optimization. Figure 2 This is a simulation diagram of the optimization effect of a GS-PSO algorithm according to an embodiment of this application.

[0092] The gradient-aware term formula in the Gradient-aware PSO (GA-PSO) algorithm is as follows:

[0093]

[0094] Among them, G i (t) is the gradient sensing direction vector of the i-th particle at time t, and its direction is expected to be the direction that will cause the objective function to decrease the fastest.

[0095] η is the gradient learning rate, used to control the step size of the gradient direction.

[0096] ▽F(X i (t) is the gradient of the objective function at the particle's current position. It is a vector pointing in the direction in which the objective function value increases the fastest.

[0097] F(X) is the objective function. i (t)+ΔX)-F(X i (t)): By calling D twice eq The model calculates the change in the objective function before and after weight fine-tuning to approximate the gradient. When the optimization objective is the average equivalent uplink distance of all users, F(X) = Mean(D) eq ).

[0098] ΔX: A tiny random perturbation vector (with the same dimension as the antenna weights).

[0099] This formula uses numerical difference to estimate in real time how the current antenna weights should be fine-tuned to most effectively reduce the global average equivalent distance. -▽F represents taking the opposite direction of the gradient (i.e., the direction of the fastest descent of the function).

[0100] By incorporating a gradient-aware term into the speed update of standard PSO, a new intelligent guidance mechanism is formed, resulting in the following speed update model:

[0101]

[0102] Here, Vi(t+1) is the velocity vector of particle i at time t+1, representing the direction and magnitude of the antenna weight parameters that need to be adjusted in the next iteration. It is a vector with the same dimension as the number of antenna weights.

[0103] ω is the inertia weight, which controls the degree to which particles inherit their current velocity, used to balance global exploration and local development. A large ω value indicates that particles have a strong ability to explore new areas, while a small ω value indicates that particles tend to search more precisely within the current area.

[0104] V i (t) is the velocity vector of particle i at time t.

[0105] c1 is the individual learning factor, which adjusts the step size of a particle's learning towards its own historical best position. It represents the degree of trust the algorithm has in the individual's experience.

[0106] r1 is the first random number uniformly distributed in the range [0, 1]. Introducing randomness prevents the algorithm from getting trapped in local optima too early, meaning that even with historical experience, the exploration still has a certain degree of uncertainty.

[0107] P best_i It is the historical best position found by particle i itself, a vector representing the position of particle i in all past iterations that makes the global average equivalent distance Mean(D) the best position. eq The set of antenna weights with the smallest weights.

[0108] X i (t) is the position vector of particle i at time t. It is a vector that directly represents the current set of antenna weight configurations (such as phase and amplitude) being evaluated.

[0109] c2 is the social learning factor, which adjusts the step size of a particle's learning towards the group's historical best position. It represents the degree of trust the algorithm has in the collective intelligence of the group.

[0110] r2 is a second random number that is uniformly distributed in the range [0, 1] and is independent of r1.

[0111] P best It is the globally optimal historical position found by the entire population, a vector representing the position of all particles in all past iterations that makes the globally average equivalent distance Mean(D) the best possible position. eq The set of antenna weights with the smallest values ​​is the target that the entire algorithm seeks.

[0112] c3 is the gradient acceleration constant, which adjusts the step size of the particle's search along the gradient direction. It determines the algorithm's feedback on the current instantaneous performance (D). eq The degree of trust in the model's calculation results.

[0113] G i (t) is the gradient-aware term, a vector, derived from D. eq The model calculations show that its direction is expected to reduce the global average equivalent distance Mean(D) most quickly. eq (direction).

[0114] Where F(X) = Mean(D) eq ).

[0115] The position update formula is:

[0116]

[0117] X i (t+1) is the position vector of particle i at time t+1. The new antenna weight configuration obtained after updating according to the new velocity will be used by D in the next iteration. eq Model evaluation.

[0118] X i (t) represents the position of particle i in the current generation; V i (t+1) represents the velocity of particle i in the next generation.

[0119] Objective function and D eq The model connection formula is:

[0120]

[0121] F(X) is the objective function, which is also the fitness function, and is the average of the equivalent uplink distance for all users.

[0122] D k eq This represents the equivalent uplink distance for the k-th user. (Based on model D) k eq =ɑ*PL k norm +β*TA k norm +γ*G(θ k h ,θ k v The calculations show that this comprehensively represents the user's uplink transmission difficulty.

[0123] The improved GA-PSO algorithm introduces D eq G calculated in real time by the model i The (t) term transforms the algorithm from a passive performance evaluator into an active, intelligent navigator that guides algorithm optimization. All parameters revolve around finding the minimum Mean(D) parameter. eq The target service is "antenna weights". Figure 3 This is a flowchart of a weight optimization process according to an embodiment of this application.

[0124] Through the above steps, optimization is performed directly based on raw uplink data, avoiding the bias that can occur when using downlink data to optimize uplink performance. This ensures that subsequent optimization measures are precisely matched with uplink service requirements, resolving the issue of poor uplink quality caused by inaccurate beam weights. Constructing an equivalent uplink distance model simultaneously balances uplink signal quality, transmission delay, and beam matching, avoiding optimization with a single dimension that cannot handle complex uplink demands. Employing a gradient-aware particle swarm optimization algorithm avoids getting stuck in local optima in the early stages while ensuring rapid convergence to the optimal solution in later stages, improving optimization efficiency and accuracy.

[0125] In some embodiments, the method further includes:

[0126] The performance indicators of the base station antenna array after operation based on the target beam weight are monitored. The performance indicators include the uplink signal-to-interference-noise ratio and throughput.

[0127] The performance indicators are evaluated based on preset threshold values. The evaluation results determine whether the performance of the target beam weight is better than that of the historical beam weight. If so, the target beam weight is used. If not, the historical beam weight is reverted to.

[0128] The target weighting scheme is distributed to the base station antenna array and takes effect. The system automatically monitors changes in key performance indicators such as uplink SINR and throughput, and evaluates the effect based on preset thresholds. Dynamic decisions are made based on the evaluation results: if performance improves, the new weights are solidified; if performance deteriorates, it automatically reverts to a historically stable version, forming an adaptive closed-loop optimization mechanism to ensure robust network operation.

[0129] The above method directly utilizes raw uplink data measured by base stations for optimization. It eliminates inaccurate indirect speculation, providing deterministic assurance, especially for critical uplink services such as live streaming and remote control. It proposes a comprehensive core metric: "equivalent uplink distance." This metric integrates information from multiple dimensions, including signal strength (path loss), transmission latency (TA), and user location (DOA), into an intelligent objective function. This allows the optimization process to simultaneously balance coverage, capacity, and latency, achieving a leap from "single-metric optimization" to "comprehensive user perception optimization." Combined with the PSO algorithm, the search efficiency is higher and the target is more precise. The fundamental shift in the optimization object is from "downlink coverage" to "uplink capacity and experience": directly utilizing uplink data to optimize the uplink, precisely improving the speed and stability of uplink services such as live streaming and IoT. The "equivalent uplink distance" model comprehensively assesses factors such as signal quality, distance, and angle, making optimization decisions more intelligent and accurate.

[0130] Example of solution effectiveness verification:

[0131] Select a cluster (3-7 stations) to cover high-user-density areas (commercial streets, convention centers, subway entrances / exits, etc.), focusing on improving the uplink experience and fairness for marginal users. Site selection: Choose macro or micro stations with a cell spacing of <300m and an antenna height of 20-30m, prioritizing coverage of traffic hotspots, and consider typical cell clusters (3-7 stations) as optimization units. Sampling period: Typical peak business hours (17:00–21:00) and off-peak hours (2:00–5:00), with 2 weeks of data collection for each period for training / validation.

[0132] Equivalent uplink distance weight (emphasizing beam alignment): ɑ(PL) norm ) = 0.30; β(TA) norm =0.20; γ(G)=0.50.

[0133] Normalized boundary: PL min =80dB; PL max =140dB; TA min =0; TA max =128dB (TA units are quantized according to the system TS).

[0134] Beam parameters: BW h =10°, BW v =8° (narrow beam, which facilitates precise pointing control).

[0135] GA-PSO parameters (for initial simulation / offline training): Population size (particles) = 40; Number of iterations = 120; Initial inertia ω init =0.9 → linear decay to ω final =0.4;

[0136] c1 = 1.4, c2 = 1.4 (individual / social learning factor).

[0137] c3 (gradient acceleration) = 0.25;

[0138] η (gradient learning rate) = 0.05;

[0139] Initial perturbation amplitude ΔX = 0.01 (within the weight normalization range);

[0140] Weight vector X dimension: phase / amplitude of each antenna element or main lobe direction ф of each beam. h ф v Controllable quantities (defined by equipment capacity).

[0141] Issuance and Implementation: Dynamic issuance frequency T downThe frequency of data transmission and monitoring will be every 5 to 10 minutes (this can be extended in non-high-fluctuation scenarios). The initial pilot program will be set to a short cycle of once every 10 minutes.

[0142] Offline simulation: A training set is built using 7 days of MR data, and GA-PSO is run offline to find the initial weight set (average of multiple random seeds).

[0143] Small-scale field testing (A / B): Select two neighboring cell clusters for A / B comparison. A pilot application of GA-PSO weights (real-time distribution); B maintains historically stable weights.

[0144] Comparison period: at least 72 hours of continuous comparison (including peak / low peak).

[0145] Data acquisition frequency: Uplink SINR, uplink throughput, uplink retransmission rate, RSRP, number of users and other KPIs are sampled every 5 minutes.

[0146] Evaluation and rollback: If the scaling-up criteria are met, gradually scale up to a larger cluster and eventually promote it across the entire network.

[0147] Production strategy: During the promotion phase, GA-PSO will be run in a hybrid offline + online mode (frequent training during the day, and solidification of optimal weights or slow updates every half hour at night).

[0148] In this deployment, 104 sites and a total of 310 cells were selected for uplink quality optimization. The weights of 140 cells were adjusted, resulting in an optimization rate of 45.16%.

[0149] Table 1 is a comparative evaluation table of key KPI indicators before and after deployment according to an embodiment of this application. As shown in Table 1, before and after optimization: the basic KPIs remained stable before and after functional deployment; the CQI good rate increased from 96.51% to 96.99%, an increase of 0.49%; the uplink sensing rate increased from 11.6Mbps to 14.7Mbps, an increase of 3.1Mbps; and the downlink sensing rate increased from 185.84Mbps to 192.23Mbps, an increase of 6.39Mbps.

[0150] Table 1

[0151]

[0152] Table 2 is a comparison table of uplink quality effect evaluation according to an embodiment of this application. As shown in Table 2, the average uplink SINR is improved by 1.6dB, and the proportion of uplink SINR greater than 6 is improved by 1.1%.

[0153] Table 2

[0154]

[0155] Automatic rollback strategy: If the closed-loop evaluation module monitors the indicators in real time and finds that the indicators do not meet the standards, it will automatically restore the historical stable weights and mark the strategy as unavailable.

[0156] Solidification: Weighting schemes that have passed verification and have no anomalies within 7×24 hours in a steady state are written into the whitelist and solidified as the default scheme.

[0157] Manual intervention: If the rollback occurs more than 3 times consecutively, the raw MR data is manually analyzed. The distribution is adjusted by modifying parameters such as α, β, γ, ΔX, and η.

[0158] It should be noted that the steps shown in the above process or in the flowchart of the accompanying figures can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0159] This embodiment also provides an uplink-based beam weight determination system, which is used to implement the above embodiments and preferred embodiments, and will not be repeated hereafter. As used below, the terms "module," "unit," "subunit," etc., can refer to a combination of software and / or hardware that performs a predetermined function. Although the apparatus described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0160] Figure 4 This is a structural block diagram of the uplink-based beam weight determination system according to an embodiment of this application, such as... Figure 4 As shown, the system includes:

[0161] The data acquisition module 41 is used to collect user equipment measurement data of the target cell and neighboring cells of the target cell. The user equipment measurement data includes uplink path loss, timing advance, and user spatial location information.

[0162] Model building module 42 is used to build an equivalent uplink distance model based on uplink path loss, time lead, and spatial orientation information.

[0163] The weight determination module 43 is used to determine the target beam weight of the base station antenna based on the equivalent uplink distance model and gradient-aware particle swarm optimization algorithm with the objective of optimizing the statistical value of the equivalent uplink distance, and then send the target beam weight to the base station antenna array.

[0164] In some embodiments, the model building module 42 includes:

[0165] The first preprocessing module is used to normalize the uplink path loss to obtain the normalized path loss.

[0166] The second preprocessing module is used to normalize the time advance to obtain the normalized time advance.

[0167] The third preprocessing module is used to calculate the beam spatial mismatch based on the spatial orientation information and the main lobe direction of the current antenna beam.

[0168] The model generation module is used to weight and fuse normalized path loss, normalized timing advance, and beam space mismatch to generate an equivalent uplink distance model.

[0169] In some embodiments, spatial orientation information includes horizontal angle of arrival and vertical angle of arrival; the equivalent uplink distance model includes:

[0170]

[0171] Among them, D eq For the equivalent uplink distance, PL norm For the normalized uplink path loss, TA norm As the time lead, G(θ) h ,θ v ) is the beam space mismatch function, θ h θ is the horizontal angle of arrival. v α is the vertical angle of arrival, and β, γ are weighting coefficients.

[0172] In some embodiments, the weight determination module 43 includes:

[0173] The first objective module aims to minimize the average equivalent uplink distance for all users.

[0174] The second objective module is used to maximize the equivalent uplink distance value of edge users at a preset percentile.

[0175] In some embodiments, the system further includes:

[0176] The performance acquisition module is used to monitor the performance indicators of the base station antenna array after it operates based on the target beam weights. The performance indicators include the uplink signal-to-interference-noise ratio and throughput.

[0177] The evaluation module is used to evaluate performance indicators based on preset indicator thresholds. Based on the evaluation results, it determines whether the performance corresponding to the target beam weight is better than the historical beam weight. If so, the target beam weight is continued to be used; otherwise, it reverts to the historical beam weight.

[0178] In some embodiments, the weight determination module 43 includes:

[0179] The model optimization module is used to incorporate gradient-aware terms into the velocity update of the standard particle swarm optimization algorithm to obtain a velocity update model.

[0180] The weight optimization module is used to determine the target beam weights of the base station antenna based on the speed update model and the equivalent uplink distance model.

[0181] In some embodiments, the speed update model includes:

[0182]

[0183] Among them, V i (t+1) is the velocity vector of particle i at time t+1, ω is the inertial weight, and V i (t) is the velocity vector of particle i at time t, c1 is the individual learning factor, r1 is the first random number uniformly distributed in the range [0, 1], and P best_i X is the historical best position found by particle i itself. i (t) is the position vector of particle i at time t, c2 is the social learning factor, r2 is the second random number uniformly distributed in the range [0, 1], and P best It is the globally optimal position found by the entire population, c3 is the gradient acceleration constant, and G is the global historical optimal position found by the entire population. i (t) is the gradient-aware term.

[0184] The system described above optimizes directly based on native uplink data, avoiding the biases that can occur when using downlink data to optimize uplink performance. This ensures that subsequent optimization measures are precisely matched to uplink service requirements, resolving the issue of poor uplink quality caused by inaccurate beam weights. Constructing an equivalent uplink distance model simultaneously balances uplink signal quality, transmission delay, and beam matching, preventing optimization from being limited to a single dimension and unable to handle complex uplink demands. Employing a gradient-aware particle swarm optimization algorithm avoids getting stuck in local optima in the early stages while ensuring rapid convergence to the optimal solution in later stages, improving optimization efficiency and accuracy.

[0185] It should be noted that the above modules can be functional modules or program modules, and can be implemented through software or hardware. For modules implemented through hardware, the above modules can reside in the same processor; or the above modules can be located in different processors in any combination.

[0186] This embodiment also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0187] Optionally, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor and the input / output device is connected to the processor.

[0188] Optionally, in this embodiment, the processor can be configured to perform the following steps via a computer program:

[0189] S1, acquire user equipment measurement data of the target cell and neighboring cells of the target cell. The user equipment measurement data includes uplink path loss, timing advance, and user spatial location information.

[0190] S2. Based on uplink path loss, time lead, and spatial orientation information, construct an equivalent uplink distance model.

[0191] S3 aims to optimize the statistical value of the equivalent uplink distance. Based on the equivalent uplink distance model and gradient-aware particle swarm optimization algorithm, it determines the target beam weight of the base station antenna and sends the target beam weight to the base station antenna array.

[0192] It should be noted that the specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementations, and will not be repeated here.

[0193] In one embodiment, Figure 5 This is a schematic diagram of the internal structure of an electronic device according to an embodiment of this application, such as... Figure 5 As shown, an electronic device is provided, which can be a server, and its internal structure diagram can be as follows. Figure 5 As shown, this electronic device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores data. The network interface communicates with external terminals via a network connection. When executed by the processor, the computer program implements an uplink-based beam weight determination method.

[0194] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the electronic device to which the present application is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0195] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0196] Those skilled in the art should understand that the technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments have been described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0197] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for determining beam weights based on the uplink, characterized in that, The method includes: Acquire user equipment measurement data of the target cell and neighboring cells of the target cell. The user equipment measurement data includes uplink path loss, timing advance, and user spatial location information. Based on the uplink path loss, the timing lead, and the spatial orientation information, an equivalent uplink distance model is constructed, including: The uplink path loss is normalized to obtain the normalized path loss. The time advance is normalized to obtain the normalized time advance. Based on the spatial orientation information and the main lobe direction of the current antenna beam, the beam spatial mismatch is calculated; The normalized path loss, the normalized time advance, and the beam space mismatch are weighted and fused to generate the equivalent uplink distance model. With the goal of optimizing the statistical value of the equivalent uplink distance, the target beam weight of the base station antenna is determined based on the equivalent uplink distance model and the gradient-aware particle swarm optimization algorithm, and the target beam weight is sent down to the base station antenna array.

2. The method according to claim 1, characterized in that, The spatial orientation information includes horizontal angle of arrival and vertical angle of arrival; the equivalent uplink distance model includes: in, D eq For equivalent uplink distance, PL norm This represents the normalized uplink path loss. TA norm For lead time, G( θ h , θ v ) is the beam space mismatch function. θ h The horizontal angle of arrival. θ v The angle of arrival is perpendicular. α , β , γ These are the weighting coefficients.

3. The method according to claim 1, characterized in that, The objective of optimizing the statistical value of the equivalent uplink distance includes: The objective is to minimize the average equivalent uplink distance for all users; or The goal is to maximize the equivalent uplink distance for edge users at a preset percentile.

4. The method according to claim 1, characterized in that, The method further includes: The performance indicators of the base station antenna array after operation based on the target beam weight are monitored, including the uplink signal-to-interference-noise ratio and throughput. The performance index is evaluated based on a preset threshold. The evaluation result determines whether the performance corresponding to the target beam weight is better than the historical beam weight. If so, the target beam weight is continued to be used; otherwise, it reverts to the historical beam weight.

5. The method according to claim 1, characterized in that, The determination of the target beam weights for the base station antenna based on the equivalent uplink distance model and gradient-aware particle swarm optimization algorithm includes: By incorporating the gradient-aware term into the velocity update of the standard particle swarm optimization algorithm, a velocity update model is obtained. Based on the speed update model and the equivalent uplink distance model, the target beam weights of the base station antenna are determined.

6. The method according to claim 5, characterized in that, The speed update model includes: Among them, V i (t+1) is the velocity vector of particle i at time t+1, ω is the inertial weight, and V i (t) is the velocity vector of particle i at time t, c1 is the individual learning factor, r1 is the first random number uniformly distributed in the range [0, 1], and P best_i X is the historical best position found by particle i itself. i (t) is the position vector of particle i at time t, c2 is the social learning factor, r2 is the second random number uniformly distributed in the range [0, 1], and G best It is the globally optimal position found by the entire population, c3 is the gradient acceleration constant, and G is the global historical optimal position found by the entire population. i (t) is the gradient-aware term.

7. A beam weight determination system based on uplink, characterized in that, The system includes: The data acquisition module is used to acquire user equipment measurement data of the target cell and neighboring cells of the target cell. The user equipment measurement data includes uplink path loss, timing advance, and user spatial location information. The model building module is used to construct an equivalent uplink distance model based on the uplink path loss, the timing lead, and the spatial orientation information, including: The uplink path loss is normalized to obtain the normalized path loss. The time advance is normalized to obtain the normalized time advance. Based on the spatial orientation information and the main lobe direction of the current antenna beam, the beam spatial mismatch is calculated; The normalized path loss, the normalized time advance, and the beam space mismatch are weighted and fused to generate the equivalent uplink distance model. The weight determination module is used to determine the target beam weight of the base station antenna based on the equivalent uplink distance model and the gradient-aware particle swarm optimization algorithm, with the goal of optimizing the statistical value of the equivalent uplink distance, and then send the target beam weight to the base station antenna array.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the uplink-based beam weight determination method as described in any one of claims 1 to 6.

9. A storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the uplink-based beam weight determination method as described in any one of claims 1 to 6.

Images