Intelligent base station antenna system

Abstract

The invention relates to the technical field of wireless communication, in particular to an intelligent base station antenna system, which comprises a multi-source sensing fusion step: acquiring base station work parameters and a terminal measurement report in real time, analyzing and extracting signal features and performing time sequence alignment, and generating a multi-dimensional sensing data set; a space-time topology reconstruction step: performing multipath propagation environment inversion based on the data set, and constructing a three-dimensional telephone traffic thermodynamic diagram reflecting the user density in a three-dimensional space; interference potential energy analysis: simulating beam overflow based on the thermodynamic diagram, calculating overlapping coverage and generating an interference potential energy evaluation result; an intelligent decision optimization step: constructing a multi-objective optimization function, and carrying out global search to solve optimal antenna weight configuration containing amplitude and phase weight; a self-adaptive reconstruction execution step: converting the weight into a beam forming instruction, driving an antenna to adjust a radiation pattern and monitoring performance in real time; according to the invention, a traditional rigid mode is broken, automatic adjustment of the coverage envelope along with traffic flow is realized, and the overall capacity of the network is improved.

Description

Technical Field

[0001] This invention relates to the field of wireless communication technology, specifically to an intelligent base station antenna system. Background Technology

[0002] Mobile communication network optimization is a core aspect of ensuring user experience and network performance. Among them, the reasonable setting of base station antenna parameters is crucial for improving coverage quality. Traditional base station antenna optimization schemes usually rely on static drive test data or planning simulations to build the network coverage foundation by presetting fixed azimuth angles, downtilt angles, and beam weights. This method can quickly achieve regional signal coverage in the early stages of network construction.

[0003] With the evolution of mobile communication technology, although technologies such as electronic downtilt and beamforming have emerged to enhance coverage flexibility, existing antenna parameter configuration modes are still mainly based on static assumptions of a two-dimensional plane. This cannot adapt to the dynamic traffic distribution changes brought about by traffic tidal effects, making it difficult to accurately match beam coverage with actual traffic hotspots in time and space. At the same time, the lack of real-time perception and avoidance mechanisms for co-channel interference from neighboring cells in three-dimensional space easily leads to pilot pollution or ineffective coverage. It is difficult to balance improving the signal quality of the cell and suppressing interference from neighboring cells in complex propagation environments, thus limiting further optimization of the overall network capacity and energy efficiency. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides an intelligent base station antenna system. Specifically, the technical solution of this invention includes:

[0005] The multi-source sensing fusion module is used to collect the current operating parameter data of the base station antenna and the measurement report data from the terminal in real time. It parses and cleans the measurement report data, extracts the signal strength, signal-to-noise ratio, angle of arrival and time advance data, and aligns the current operating parameter data with the extracted data in time sequence to generate a multi-dimensional sensing dataset.

[0006] The spatiotemporal topology reconstruction module is used to invert the multipath propagation environment in virtual space based on multidimensional sensing datasets and combined with geographic information system data, and to construct a three-dimensional traffic heat map that reflects the distribution of user density in three-dimensional space.

[0007] The interference potential energy analysis module is used to simulate the signal overflow situation of the current beam pattern at the edge of the neighboring cell based on the three-dimensional traffic heat map, calculate the overlap coverage, and generate interference potential energy assessment results accordingly.

[0008] The intelligent decision optimization module is used to construct a multi-objective optimization function with the goal of maximizing the signal strength of the hotspot area of ​​the cell and minimizing the interference potential energy of the neighboring cell. It performs a full-domain search on the three-dimensional traffic heat map and calculates the optimal antenna weight configuration.

[0009] The adaptive reconfiguration execution module is used to convert the optimal antenna weight configuration into beamforming commands or electrically adjustable downtilt commands, drive the reconfigurable antenna actuator to adjust the radiation pattern, and monitor the adjusted network performance indicators in real time to form a closed-loop control.

[0010] Preferred methods for generating multidimensional sensing datasets include:

[0011] Obtain scheduling information for each transmission time interval recorded by the base station to determine the effective service period;

[0012] Sampling points that fall within the valid business hours are selected from the measurement report data;

[0013] Based on a preset statistical deviation threshold or interquartile range rule, outlier removal is performed on the selected sampling points to obtain valid sampling points.

[0014] The coordinate system is transformed between the angle of arrival data corresponding to the effective sampling point and the normal direction of the base station antenna to obtain the absolute spatial angle;

[0015] The absolute spatial angle, signal strength, and signal-to-noise ratio are mapped to the current engineering parameter data according to the timestamp index, thus completing the generation of the multidimensional sensing dataset.

[0016] Preferred methods for constructing three-dimensional call traffic heatmaps include:

[0017] The coverage area is divided into three-dimensional grid cells of preset size;

[0018] Based on the time lead data, the radial distance between the user and the base station is estimated, and combined with the absolute spatial angle, the spatial projection position of each effective sampling point in the three-dimensional grid cell is determined.

[0019] Density clustering analysis is performed on the effective sampling points within each three-dimensional grid cell to calculate the traffic density value of each three-dimensional grid cell.

[0020] Kriging interpolation or Gaussian process regression is used to smoothly fill sparse regions, generating a continuous three-dimensional traffic heatmap.

[0021] Preferably, the method for generating the disturbance potential assessment results includes:

[0022] Obtain the location information of neighboring base stations and the geographical boundary information of the neighboring cell edges;

[0023] Mark the interference-sensitive areas that belong to the coverage area of ​​the neighboring cell in the 3D traffic heat map;

[0024] The power spectral density distribution of the current beam configuration in the interference-sensitive region was simulated using a ray tracing model.

[0025] Calculate the cumulative received power in the interference-sensitive area and compare it with the preset interference tolerance threshold;

[0026] If the accumulated received power is greater than the interference tolerance threshold, a high interference potential energy marker is generated, and the interference overflow of the excess portion is calculated.

[0027] If the cumulative received power is less than or equal to the interference tolerance threshold, a low interference potential energy marker is generated.

[0028] High or low disturbance potential energy markers are used as the results of disturbance potential energy assessment.

[0029] Preferably, the method for calculating the optimal antenna weight configuration includes:

[0030] Initialize a random population of antenna weights, each containing amplitude weights and phase weights;

[0031] Substitute the antenna weights of each group into the preset antenna pattern synthesis formula to generate the corresponding virtual beamform.

[0032] Calculate the target coverage gain of the virtual beamform in the three-dimensional traffic heatmap, and calculate the fitness function value that is positively correlated with the target coverage gain and negatively correlated with the interference spillover amount according to the preset weighting logic.

[0033] Based on the fitness function value, the antenna weight population is selected, crossovered and mutated to generate a new generation of antenna weight population;

[0034] Repeat the iteration until the fitness function value converges or the preset number of iterations is reached, and select the set of antenna weights with the highest fitness function value as the optimal antenna weight configuration.

[0035] Preferably, the method for driving the reconfigurable antenna actuator to adjust the radiation pattern includes:

[0036] Analyze the optimal antenna weight configuration to separate the phase adjustment parameter and the amplitude adjustment parameter;

[0037] If the interference potential energy assessment result is a high interference potential energy marker, then based on the phase adjustment parameter, a radiation null point is generated in the physical direction of the corresponding interference sensitive area to form a null trap beam.

[0038] If the displacement of the spatial geometric centroid of the call hotspot area in the three-dimensional call heat map exceeds the preset threshold within the preset time window, it is determined to be a tidal distribution change. Based on the amplitude adjustment parameter, the beam main lobe width is dynamically compressed or widened to match the geometric envelope of the call hotspot area.

[0039] The generated null beam configuration or main lobe width configuration is sent to the phase shifter and attenuator network to complete the adjustment of the radiation pattern.

[0040] Preferred methods for real-time monitoring and adjustment of network performance indicators to form closed-loop control include:

[0041] Within the preset observation window after adjusting the radiation pattern, re-acquire measurement report data;

[0042] Calculate the adjusted average signal-to-noise ratio for the entire network and the edge user rate;

[0043] Calculate the performance gain difference before and after the adjustment;

[0044] If the performance gain difference is less than the preset minimum positive return threshold, it is determined that there is a deviation in the current environment model, and a model correction instruction is triggered.

[0045] In response to the model correction command, the multipath fading factor in the spatiotemporal topology reconstruction module is adjusted, and the disturbance potential energy analysis and intelligent decision optimization process is re-executed.

[0046] Preferably, the system further includes an energy efficiency management module, used for:

[0047] Based on the 3D call heat map, the total call demand at the current moment is calculated.

[0048] Compare the total call volume demand with the preset energy-saving activation threshold;

[0049] If the total traffic demand is less than the energy-saving start threshold, a channel shutdown command is generated to lock some antenna array channels and recalculate the optimal antenna weight configuration of the remaining channels to maintain the basic coverage waveform.

[0050] If the total call demand is greater than or equal to the energy-saving activation threshold, then the full-channel operation mode will be maintained.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] 1. This system effectively filters out noise interference during non-service transmission periods and abnormal data caused by instantaneous shadow fading by multi-source sensing fusion and spatiotemporal topology reconstruction, combined with effective service period filtering and statistical deviation rules based on scheduling information, ensuring the purity of input data. During coordinate transformation, it corrects the fundamental deviation of introducing electronic downtilt angle errors into the rotation matrix, and uses only mechanical downtilt angle to construct an absolute spatial coordinate system, eliminating systematic positioning errors. Furthermore, by utilizing time lead and angle of arrival data, combined with Gaussian kernel density estimation and Kriging interpolation algorithms, it generates continuous high-precision three-dimensional traffic heatmaps without satellite positioning assistance, thereby accurately perceiving the three-dimensional traffic distribution inside buildings and high-rise areas, solving the technical problem that traditional two-dimensional planar analysis cannot cope with three-dimensional coverage of high-rise buildings.

[0053] 2. This system, through interference potential energy analysis and intelligent decision optimization, introduces a modified ray tracing model and linearized interference spillover calculation logic, avoiding the confusion of physical meaning caused by direct accumulation of logarithmic units, and accurately quantifying the cumulative energy projection of the beam onto the edge of the neighboring cell. On this basis, a multi-objective fitness function containing positive coverage gain and negative interference penalty is constructed, and a heuristic search algorithm is used to find the optimal value in a large parameter space. A nonlinear penalty mechanism forces the algorithm to avoid high interference areas. This process comprehensively considers amplitude and phase weights, and can solve the optimal antenna weight configuration that maximizes the signal strength of the hotspot in the cell and minimizes the risk of pilot pollution in the neighboring cell without increasing hardware costs, thus achieving a dynamic balance between network capacity and interference control.

[0054] 3. This system achieves flexible antenna radiation pattern shaping and deep energy saving through adaptive reconfiguration execution and energy efficiency management; it generates radiation nulls in interference-sensitive directions using phase perturbation method to accurately eliminate co-channel interference to neighboring cells; it also dynamically adjusts the beam main lobe width using amplitude taper technology based on spatial centroid displacement monitoring of traffic hotspots, allowing it to automatically adapt to traffic envelope changes caused by tidal effects, much like a fluid; furthermore, it uses a sparse array masking mechanism to lock some channels during low-load periods and re-optimizes the weights of the remaining active channels, significantly reducing base station operating energy consumption while ensuring that the basic coverage waveform does not exhibit black holes, breaking the rigid mode of traditional antennas operating on all channels at all times.

[0055] 4. This system, through closed-loop control and model self-correction mechanisms, endows the system with self-diagnosis and environmental adaptability. The system monitors the adjusted average signal-to-noise ratio and edge user rate in real time, and calculates the dimensionless comprehensive performance gain using a standardized method based on historical statistical standard deviation, eliminating evaluation bias caused by different physical dimensions. When the actual performance gain does not reach the expected threshold, the feedback control process is automatically triggered, and the multipath fading factor in the environmental model is dynamically adjusted using a discretized proportional-integral-differential algorithm. This forces the optimization algorithm to find a beam solution again under conditions that are closer to the real physical environment or more stringent constraints, thereby avoiding strategy failure caused by changes in building reflection characteristics or model parameter distortion, and ensuring the long-term robustness of network performance in dynamic environments. Attached Figure Description

[0056] The present invention will be further explained below with reference to the accompanying drawings and embodiments:

[0057] Figure 1 This is a structural diagram of the system of the present invention. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0059] Example 1:

[0060] Please see Figure 1 A smart base station antenna system, comprising:

[0061] The multi-source sensing fusion module is used to collect the current operating parameter data of the base station antenna and the measurement report data from the terminal in real time. It parses and cleans the measurement report data, extracts the signal strength, signal-to-noise ratio, angle of arrival and time advance data, and aligns the current operating parameter data with the extracted data in time sequence to generate a multi-dimensional sensing dataset.

[0062] The spatiotemporal topology reconstruction module is used to invert the multipath propagation environment in virtual space based on multidimensional sensing datasets and combined with geographic information system data, and to construct a three-dimensional traffic heat map that reflects the distribution of user density in three-dimensional space.

[0063] The interference potential energy analysis module is used to simulate the signal overflow situation of the current beam pattern at the edge of the neighboring cell based on the three-dimensional traffic heat map, calculate the overlap coverage, and generate interference potential energy assessment results accordingly.

[0064] The intelligent decision optimization module is used to construct a multi-objective optimization function with the goal of maximizing the signal strength of the hotspot area of ​​the cell and minimizing the interference potential energy of the neighboring cell. It performs a full-domain search on the three-dimensional traffic heat map and calculates the optimal antenna weight configuration.

[0065] The adaptive reconfiguration execution module is used to convert the optimal antenna weight configuration into beamforming commands or electrically adjustable downtilt commands, drive the reconfigurable antenna actuator to adjust the radiation pattern, and monitor the adjusted network performance indicators in real time to form a closed-loop control.

[0066] This embodiment provides an intelligent base station antenna system. This system aims to address the technical problems of existing base station antenna parameters, which are typically set based on static drive test data, making them unable to adapt to dynamic traffic changes caused by tidal effects, and the difficulty in real-time avoidance of co-channel interference from neighboring cells. The system includes a multi-source sensing fusion module, configured to collect current operating parameter data of the base station antenna and measurement report data from terminals in real time. It parses and cleans the measurement report data, extracting signal strength, signal-to-noise ratio, angle of arrival, and timing advance data. The current operating parameter data is then time-aligned with the extracted data to generate a multi-dimensional sensing dataset. This module forms the system's data foundation, connecting the baseband processing unit (BBU) and antenna control unit of the base station via a physical interface, and collecting operating parameter data in real time, including azimuth angle, mechanical downtilt angle, electronic downtilt angle, and current weights. The system also includes a spatiotemporal topology reconstruction module, configured to invert the multipath propagation environment in virtual space based on the multi-dimensional sensing dataset and combined with geographic information system data, constructing a three-dimensional traffic heatmap reflecting the user density distribution in three-dimensional space. This heatmap serves as a digital twin basis for decision-making, transforming discrete signal data into a visualized spatial distribution.

[0067] The system further includes an interference potential energy analysis module, configured to simulate signal spillover at the edge of neighboring cells based on a three-dimensional traffic heatmap, calculate overlap coverage, and generate interference potential energy assessment results accordingly. The concept of interference potential energy is introduced here, referring to the cumulative intensity of beam energy projected onto the non-service area, i.e., the edge of neighboring cells, to characterize the risk level of pilot pollution. The system also includes an intelligent decision optimization module, configured to construct a multi-objective optimization function to maximize the signal strength in the hotspot area of ​​the cell and minimize the interference potential energy of neighboring cells. It performs a full-domain search on the three-dimensional traffic heatmap to calculate the optimal antenna weight configuration. This module uses a heuristic search algorithm to calculate configuration parameters including amplitude and phase weights. The system includes an adaptive reconfiguration execution module, configured to convert the optimal antenna weight configuration into beamforming commands or electrically adjustable downtilt commands, driving the reconfigurable antenna actuator to adjust the radiation pattern and monitoring the adjusted network performance indicators in real time to form closed-loop control. This module drives the phase shifter and attenuator network through the antenna control interface, converting digital signals into physical actions.

[0068] This embodiment constructs a closed-loop control mechanism for perception computing reconstruction, breaking the rigid mode of traditional fixed antenna beam coverage of fixed areas; the system can automatically adjust the coverage envelope according to the flow of traffic hotspots like a fluid, which not only improves the signal-to-noise ratio of users at the edge of the cell, but also reduces interference to neighboring cells by suppressing interference potential energy, thereby significantly improving the overall network capacity without increasing hardware costs.

[0069] Example 2:

[0070] Methods for generating multidimensional sensing datasets include:

[0071] Obtain scheduling information for each transmission time interval recorded by the base station to determine the effective service period;

[0072] Sampling points that fall within the valid business hours are selected from the measurement report data;

[0073] Based on a preset statistical deviation threshold or interquartile range rule, outlier removal is performed on the selected sampling points to obtain valid sampling points.

[0074] The coordinate system is transformed between the angle of arrival data corresponding to the effective sampling point and the normal direction of the base station antenna to obtain the absolute spatial angle;

[0075] The absolute spatial angle, signal strength, and signal-to-noise ratio are mapped to the current engineering parameter data according to the timestamp index, thus completing the generation of the multidimensional sensing dataset.

[0076] This embodiment details the specific logic for generating a multidimensional sensing dataset. This process aims to ensure the purity and spatial consistency of the input data. The system obtains the scheduling information of each transmission time interval recorded by the base station to determine the effective service period. This period refers to the time when the base station has actual data transmission and the resource block (RB) occupancy rate exceeds the preset benchmark value. The system filters out the sampling points that fall into the effective service period from the measurement report data, and performs outlier removal processing on the filtered sampling points based on the preset statistical deviation threshold or interquartile range rule to obtain the effective sampling points.

[0077] The system performs statistical analysis on the selected set of sampling points and calculates the upper quartiles. and lower quartile The interquartile range was obtained. And set the upper and lower limits for truncation as follows: and Coefficients in the formula This is a standard constant for box plots in statistics, used to identify mild outliers. This value can be adjusted based on the sensitivity of the business scenario to outlier data. to Adjust between; whenever the signal strength or signal-to-noise ratio value is not within the range Outliers within the closed interval are removed to filter out dirty data caused by equipment failure or transient shadow fading. The system performs coordinate transformation between the arrival angle data corresponding to the valid sampling points and the normal direction of the base station antenna to obtain the absolute spatial angle. This step requires establishing an accurate geometric mapping, and the specific transformation uses an Euler rotation matrix: Let the azimuth angle of the antenna mechanical normal in the base station's engineering parameters be... The mechanical tilt angle is ;in, and The coordinate transformation is based on the current time value read in real time through the antenna control unit interface, rather than preset static parameters, to ensure that the coordinate transformation can reflect the real-time physical attitude of the antenna.

[0078] It should be noted here that although the electron downtilt angle The beam pointing can be changed electronically, but the relative angle of arrival in the measurement report... This is the reading relative to the physical plane of the antenna array calculated by the baseband processing unit. It means that the angular offset caused by beamforming weights has been removed during the calculation process. Therefore, when performing absolute spatial coordinate system transformation, the rotation matrix should only use the mechanical downtilt angle that determines the physical plane's attitude. It is strictly forbidden to include the electron downtilt angle in the rotation matrix, otherwise it will lead to a systematic deviation in the coordinate system definition;

[0079] Then the absolute space coordinate vector The calculation is as follows:

[0080]

[0081] in, , They are respectively around shaft and The rotation matrix of the axis, in this embodiment, specifically uses the ZY follow-the-clock coordinate system rotation. To ensure the uniqueness of the coordinate transformation, it is defined as follows:

[0082]

[0083] Final absolute spatial azimuth , here and Representing absolute spatial coordinate vectors respectively Y-axis and X-axis components in a Cartesian coordinate system; absolute spatial downtilt angle , here Represents absolute spatial coordinate vector The Z-axis component in the Cartesian coordinate system; the system maps the absolute spatial angle, signal strength and signal-to-noise ratio to the current working parameter data according to the timestamp index, and completes the generation of the multidimensional sensing dataset;

[0084] This embodiment effectively filters out noise interference during non-service transmission periods by filtering effective service periods based on scheduling information and removing IQR outliers. Combined with a coordinate system transformation and timestamp mapping mechanism based on the corrected rotation matrix, it corrects the fundamental deviation of introducing electronic downtilt errors into coordinate transformation, providing high-confidence and spatiotemporally aligned input data for subsequent accurate three-dimensional spatial inversion, and solving the engineering problem of difficult fusion of multi-source heterogeneous data.

[0085] Example 3:

[0086] Methods for constructing a 3D call traffic heatmap include:

[0087] The coverage area is divided into three-dimensional grid cells of preset size;

[0088] Based on the time lead data, the radial distance between the user and the base station is estimated, and combined with the absolute spatial angle, the spatial projection position of each effective sampling point in the three-dimensional grid cell is determined.

[0089] Density clustering analysis is performed on the effective sampling points within each three-dimensional grid cell to calculate the traffic density value of each three-dimensional grid cell.

[0090] Kriging interpolation or Gaussian process regression is used to smoothly fill sparse regions, generating a continuous three-dimensional traffic heatmap.

[0091] This embodiment details the specific logic for constructing a three-dimensional traffic heatmap. This process aims to reconstruct discrete sampling points into a continuous spatial field distribution. The system divides the coverage area into three-dimensional grid cells of preset size. Based on time advance data, the system estimates the radial distance between the user and the base station, and combines this with absolute spatial angles to determine the spatial projection position of each valid sampling point in the three-dimensional grid cell. The position estimation model is as follows:

[0092]

[0093] In the formula, the subscript Indicates the index of the valid sampling point. The range of values ​​is ,in, This represents the total number of valid sampling points selected within the transmission time interval. For the first The spatial coordinate vector of each sampling point; The known geographic coordinate vector of the base station antenna; The radial distance is estimated based on time lead, and the calculation formula is as follows: ,in, For time lead index value, As the basic unit of time, The speed of light; The direction vector function is specifically defined as the transformation from spherical coordinates to Cartesian coordinates: ; This is the absolute spatial azimuth. This is the absolute spatial tilt angle;

[0094] The system performs density clustering analysis on the effective sampling points within each 3D grid cell to calculate the traffic density value of each 3D grid cell; the specific density calculation uses a Gaussian kernel density estimation model.

[0095]

[0096] in, The number of points within the grid. As the center of the grid, Bandwidth is used to control the smoothness level; in this embodiment, bandwidth... The value is based on the grid side length. Confirm, set as Alternatively, Silverman's rule of thumb can be used to adaptively calculate based on the sample standard deviation to balance smoothing effects with detail preservation; The kernel function is Gaussian; in this embodiment, the standard normal distribution function is specifically used.

[0097]

[0098] The introduction of this function ensures that the influence of the sampling point on the surrounding grid decreases normally with distance; the system uses Kriging interpolation or Gaussian process regression to smoothly fill sparse regions, generating a continuous three-dimensional traffic heatmap; specifically, ordinary Kriging interpolation with the spherical model variogram is used here:

[0099]

[0100] in, For spatial distance, For variable range, Value of a nugget. The arch height, calculated as the sill value minus the nugget value, represents the structural variance; the parameters in the above variogram model... It was not set arbitrarily, but obtained through experimental variogram fitting: calculating different distance step sizes. Experimental variogram values ,in For distance equal to Number of point pairs and Positions and The observed variable value at the location is, in this embodiment, the traffic density value of the corresponding grid; the weighted least squares method is used to transform the theoretical variogram model. To approximate the experimental variogram point set, the specific fitting process utilizes the Levenberg-Marquardt algorithm to iteratively solve for the parameters. This continues until the sum of squared residuals is minimized; thus, the optimal parameter combination is calculated; after completing the variogram modeling, for any grid center to be interpolated... Its valuation The weight vector This is obtained by solving the following Kriging equations:

[0101]

[0102] in, for The semi-variogram matrix, whose elements , This is the semi-variation vector between the observed point and the unknown point, and its elements are... , These are Lagrange multipliers; this step uses the variogram of known sampling points to predict values ​​in unknown regions, thereby obtaining a continuous field distribution;

[0103] This embodiment cleverly utilizes the TA and AOA data commonly found in the existing network, and combines them with the Kriging interpolation algorithm to generate a high-precision three-dimensional user distribution map at low cost without the need for GPS assistance. This enables the base station to perceive the traffic distribution inside buildings and on high floors, solving the problem that traditional two-dimensional planar analysis cannot cope with three-dimensional coverage of high-rise buildings.

[0104] Example 4:

[0105] Methods for generating disturbance potential assessment results include:

[0106] Obtain the location information of neighboring base stations and the geographical boundary information of the neighboring cell edges;

[0107] Mark the interference-sensitive areas that belong to the coverage area of ​​the neighboring cell in the 3D traffic heat map;

[0108] The power spectral density distribution of the current beam configuration in the interference-sensitive region was simulated using a ray tracing model.

[0109] Calculate the cumulative received power in the interference-sensitive area and compare it with the preset interference tolerance threshold;

[0110] If the accumulated received power is greater than the interference tolerance threshold, a high interference potential energy marker is generated, and the interference overflow of the excess portion is calculated.

[0111] If the cumulative received power is less than or equal to the interference tolerance threshold, a low interference potential energy marker is generated.

[0112] High or low disturbance potential energy markers are used as the results of disturbance potential energy assessment.

[0113] This embodiment details the specific logic for generating interference potential assessment results. This process aims to quantify the potential harm of the beam to neighboring cells. According to an intelligent base station antenna system in Embodiment 3, the method for generating interference potential assessment results includes: acquiring the location information of neighboring base stations and the geographical boundary information of the neighboring cell edges; marking interference-sensitive areas belonging to the coverage area of ​​the neighboring cells in a three-dimensional traffic heatmap; these areas typically include locations where handover failures or high interference records have occurred historically; and simulating the power spectral density distribution of the current beamform in the interference-sensitive areas using a ray tracing model. This model is based on geometric optics principles, calculates the coherent superposition of the line-of-sight path and the reflection path, and introduces free-space path loss correction, with single-point received power... The calculation formula is:

[0114]

[0115] in, This represents the received power at sampling points within the interference-sensitive area, in units of: , Transmission power, unit: , Transmit antenna gain, unit: The formula internally represents the vector superposition of the direct and multipath signals in the complex domain, followed by modulo calculation of the synthesized amplitude; where... The direct beam amplitude, For the first The amplitude of the reflection path takes into account free space loss; For the first The reflectivity of the path, This refers to the phase lag caused by the path length difference; where, Based on the building surface material properties corresponding to the reflection points in the GIS database, such as glass curtain walls, concrete, and brick walls, the electromagnetic material parameters are obtained by querying a pre-set electromagnetic material parameter library.

[0116] It is defined as the phase shift caused by the difference in path length: ; in the formula The center carrier frequency of the base station system is given in Hz. The cumulative received power within the interference-sensitive area is calculated and compared to a preset interference tolerance threshold. If the cumulative received power exceeds the interference tolerance threshold, a high interference potential energy marker is generated, and the interference overflow of the excess portion is calculated. If the cumulative received power is less than or equal to the interference tolerance threshold, a low interference potential energy marker is generated. The high or low interference potential energy marker is used as the interference potential energy assessment result. The interference overflow calculation formula is modified to linear power accumulation to address the unclear physical meaning of logarithmic unit accumulation.

[0117]

[0118] in, This is the amount of interference overflow; This refers to the set of sampling points within the interference-sensitive area. This represents the cumulative received power at the m-th sampling point obtained through simulation calculation; This is the preset interference tolerance threshold. It should be noted that the formula first converts the power in dBm units to linear units mW, calculates the physical power difference exceeding the threshold, and then accumulates them to obtain the total interference power overflow value with clear physical meaning, thus avoiding the ambiguity of principle caused by directly summing the dB values.

[0119] This embodiment eliminates dimensional errors and confusion of physical meaning by introducing a modified ray tracing model and linearized interference spillover calculation. It can accurately identify malicious beamforms that, while improving coverage in the local area, severely interfere with neighboring areas, providing key constraints that conform to physical facts for subsequent multi-objective optimization.

[0120] Example 5:

[0121] Methods for calculating the optimal antenna weight configuration include:

[0122] Initialize a random population of antenna weights, each containing amplitude weights and phase weights;

[0123] Substitute the antenna weights of each group into the preset antenna pattern synthesis formula to generate the corresponding virtual beamform.

[0124] Calculate the target coverage gain of the virtual beamform in the three-dimensional traffic heatmap, and calculate the fitness function value that is positively correlated with the target coverage gain and negatively correlated with the interference spillover amount according to the preset weighting logic.

[0125] Based on the fitness function value, the antenna weight population is selected, crossovered and mutated to generate a new generation of antenna weight population;

[0126] Repeat the iteration until the fitness function value converges or the preset number of iterations is reached, and select the set of antenna weights with the highest fitness function value as the optimal antenna weight configuration.

[0127] This embodiment details the specific logic for calculating the optimal antenna weight configuration. This process uses a heuristic algorithm to find the global optimal solution in the parameter space. According to an intelligent base station antenna system in Embodiment 4, the method for calculating the optimal antenna weight configuration includes: initializing a set of random antenna weight populations, each set of antenna weights including amplitude weights and phase weights; substituting each set of antenna weights into a preset antenna pattern synthesis formula to generate the corresponding virtual beamform; the specific synthesis formula is as follows:

[0128]

[0129] in, For the first The weighted values ​​of each array element, i.e. For amplitude, For phase, This refers to the index of an element in an antenna array, with a value range of... ,for A rectangular planar array, index Traverse all array elements in row-major order; simultaneously, and In actual output, discretization and quantization processing are required based on the step accuracy of the hardware phase shifter and attenuator. To ensure the physical computability of the simulation model, this embodiment uses the general element model approximation defined in 3GPP 38.901, specifically the single radiating element pattern model from Table 7.3-1 of that standard, for mathematical expression. Assuming the horizontal direction is omnidirectional radiation, where,

[0130]

[0131] For wave number, The coordinates of the array element positions are given; the target coverage gain of the virtual beamform in the three-dimensional traffic heatmap is calculated, and this gain... Defined as a weighted spatial integral of beam power and traffic density:

[0132]

[0133] in, 3D mesh Traffic density value, For grid The spatial angle is determined; and according to the preset weighted logic, the fitness function value that is positively correlated with the target coverage gain and negatively correlated with the interference spillover is calculated; to ensure the correctness of the physical meaning and eliminate the optimization bias caused by the difference in dimensions, the gain and interference terms are normalized and a minimum value is introduced. To prevent the denominator from being zero, and also to prevent the target from covering the gain. Setting the value to 0 causes logarithmic operations to overflow; therefore, set... :

[0134]

[0135] in, These two parameters, representing the historical gain extremes, are dynamically updated using a sliding window mechanism as the system operates to adapt to fluctuations in call traffic characteristics at different times; if no historical data is available, the system is initialized by default. This value is set based on the physical gain limit of the antenna array; ; This is a reference value for the maximum permissible interference overflow, in milliwatts (mW).

[0136] The fitness function is constructed as follows:

[0137]

[0138] in, The target coverage gain of the normalized virtual beamform in the 3D traffic heatmap; This represents the normalized interference overflow. To cover the weighting coefficient, the value ranges from 0.6 to 0.8. The specific value is determined using a lookup table method based on the network deployment scenario. The lookup table configuration logic is as follows: In a high-density call traffic scenario in a CBD, set... Prioritize coverage gain; in suburban or wide-coverage scenarios, set... To enhance the suppression of interference from neighboring cells; This is the interference weighting coefficient, with a value ranging from 0.2 to 0.4. The interference term is squared here to address the logical flaw in linear weighted logic where high gain and high interference may outperform low gain and low interference. A nonlinear penalty forced algorithm is used to avoid the high interference region. Based on this, the antenna weight population is selected, crossovered, and mutated according to the fitness function value to generate a new generation of antenna weight population.

[0139] To ensure the convergence of the algorithm and the diversity of solutions, the specific operational strategy is as follows: Selection operation: A binary tournament selection method is used, randomly selecting two individuals from the population each time, and retaining the fitness function value. Larger individuals enter the next generation breeding pool; Crossover operation: Taking advantage of the fact that both amplitude and phase weights are continuous real numbers, simulated binary crossover is used, with a crossover probability set to 0.8 and a distribution index set to 20, to generate offspring weights with similar statistical characteristics near the parent weights; Mutation operation: Multinomial mutation or Gaussian mutation is used to superimpose random perturbations with a mean of 0 and a variance decreasing with the number of iterations onto the amplitude and phase of the offspring individuals, respectively, with a mutation probability set to 1 / N, to prevent the algorithm from getting trapped in local optima; Repeat the iteration until the fitness function value converges or reaches the preset number of iterations, and select the set of antenna weights with the highest fitness function value as the optimal antenna weight configuration;

[0140] This embodiment constructs a fitness function that includes positive gain and negative interference penalty, and modifies the normalization benchmark of the interference term to adapt to linear physical quantities, ensuring the physical self-consistency of the optimization objective; combined with the SBX crossover and polynomial mutation strategy in the genetic algorithm, it can quickly find the global optimal solution that balances coverage and interference in a huge parameter space.

[0141] Example 6:

[0142] Methods for driving reconfigurable antenna actuators to adjust radiation patterns include:

[0143] Analyze the optimal antenna weight configuration to separate the phase adjustment parameter and the amplitude adjustment parameter;

[0144] If the interference potential energy assessment result is a high interference potential energy marker, then based on the phase adjustment parameter, a radiation null point is generated in the physical direction of the corresponding interference sensitive area to form a null trap beam.

[0145] If the displacement of the spatial geometric centroid of the call hotspot area in the three-dimensional call heat map exceeds the preset threshold within the preset time window, it is determined to be a tidal distribution change. Based on the amplitude adjustment parameter, the beam main lobe width is dynamically compressed or widened to match the geometric envelope of the call hotspot area.

[0146] The generated null beam configuration or main lobe width configuration is sent to the phase shifter and attenuator network to complete the adjustment of the radiation pattern.

[0147] This embodiment details the specific logic of driving the reconfigurable antenna actuator to adjust the radiation pattern. This process achieves scene-based form-based maneuvering. The system analyzes the optimal antenna weight configuration and separates the phase adjustment parameters and amplitude adjustment parameters. If the interference potential energy assessment result is a high interference potential energy marker, then based on the phase adjustment parameters, a radiation null point is generated in the physical direction of the corresponding interference-sensitive area, forming a null beam. This is equivalent to creating an energy depression in the radiation pattern, causing the interference signals to cancel each other out in that direction. Its specific implementation is not a simple phase truncation, but rather uses a phase perturbation method to solve the null constraint: Let the target null angle be... The system calculates a set of phase perturbation vectors. So that the array factor satisfies ;in, The value of the zero-dimple depth constraint threshold is taken in this embodiment. This corresponds to approximately -60dB, ensuring effective interference suppression. Here, the gradient projection method is used to iteratively solve the aforementioned nonlinear constraints, constructing the objective function. To minimize the phase perturbation of the original beam while satisfying the null constraint, the specific iterative update formula is as follows:

[0148]

[0149] in, This represents the conjugate transpose of the Jacobian matrix. This indicates the operation of taking the real part, ensuring the phase perturbation amount It always remains a real physical quantity; Defined as array synthesis factor:

[0150]

[0151] The learning rate is initially set to 0.01, and decreases exponentially with the number of iterations. The Jacobian matrix is ​​the constraint condition. To ensure code reproducibility, it is explicitly defined. The The formula for calculating each element is: This step aims to minimize the phase perturbation while satisfying the null trap constraint, and solve for the... This is the final instruction sent to the phase shifter; if the displacement of the spatial geometric centroid of a traffic hotspot area in the 3D traffic heatmap exceeds a preset threshold within a preset time window, based on the base station coverage radius... If this is confirmed, the distribution is determined to be tidal, and the main lobe width of the beam is dynamically compressed or widened based on the amplitude adjustment parameters to match the geometric envelope of the traffic hotspot area. This step utilizes amplitude tapering technology, where the system maps the beamwidth requirement in the optimal antenna weight configuration to Taylor distribution parameters. The specific parameter mapping logic is as follows: calculate the standard deviation of the spatial distribution of the traffic hotspot area in the beam section direction. Set the target half-power beamwidth ;based on The sidelobe level parameters required to inversely solve the Taylor distribution And substitute into the following formula to calculate the first... The amplitude weight of each element :for A rectangular planar array, ultimately the first Amplitude weights of each physical element Separable distribution synthesis is employed, i.e. ,in and Apply the Taylor distribution formula described above based on the number of rows. With column number Calculate independently;

[0152]

[0153] in, The total number of array elements; These are Taylor coefficients, calculated based on the sidelobe level parameter. and normal sidelobe control parameters ,in, The preset number of side lobes in the normal accessory lobe region is typically set to a value of [value missing]. ; The target sidelobe level, in dB, is typically a negative value. (Parameter) and The mapping relationship is , thus calculating Substitute the value into the subsequent formula;

[0154] The specific formula is as follows:

[0155] Among them, symbols This indicates a multiplication operation, and the loop variable is modified to... To distinguish, among which, scaling factor Defined as:

[0156]

[0157] For example, when traffic is dispersed in a nearby square and the beam needs to be widened, the system reduces the amplitude weights of the array elements at both ends through the attenuator network, making them present a Taylor distribution or Chebyshev distribution with high values ​​in the middle and low values ​​at both ends, thereby increasing the main lobe width. Conversely, when traffic is concentrated in a distant high-rise building and the beam needs to be compressed, the amplitude weights are adjusted to tend towards a uniform distribution in order to obtain the narrowest beam width and the maximum directivity coefficient. The system sends the generated null beam configuration or main lobe width configuration to the phase shifter and attenuator network to complete the adjustment of the radiation pattern.

[0158] This embodiment achieves two advanced functions through decoupling control of phase and amplitude: using phase interference to create nulls in a specific direction to accurately eliminate interference, and using amplitude weighting to dynamically change the beam width to adapt to tidal traffic; this reconfiguration capability gives the antenna system environmental adaptability similar to that of a living organism.

[0159] Example 7:

[0160] Methods for real-time monitoring and adjustment of network performance metrics to form closed-loop control include:

[0161] Within the preset observation window after adjusting the radiation pattern, re-acquire measurement report data;

[0162] Calculate the adjusted average signal-to-noise ratio for the entire network and the edge user rate;

[0163] Calculate the performance gain difference before and after the adjustment;

[0164] If the performance gain difference is less than the preset minimum positive return threshold, it is determined that there is a deviation in the current environment model, and a model correction instruction is triggered.

[0165] In response to the model correction command, the multipath fading factor in the spatiotemporal topology reconstruction module is adjusted, and the disturbance potential energy analysis and intelligent decision optimization process is re-executed.

[0166] This embodiment details the specific logic of real-time monitoring of adjusted network performance indicators to form closed-loop control, a process that endows the system with self-diagnostic capabilities. Within a preset observation window after adjusting the radiation pattern, the system re-collects measurement report data. The system calculates the adjusted average signal-to-noise ratio (SNR) and edge user rate, and calculates the performance gain difference before and after adjustment. Since SNR and user rate belong to different physical dimensions, direct calculation lacks physical meaning. This embodiment uses the Z-score normalization method based on historical statistical standard deviation to calculate the dimensionless comprehensive performance gain. The calculation formula is as follows:

[0167]

[0168] in, This represents the performance gain difference. The signal-to-noise ratio is the average signal-to-noise ratio of the entire network before and after adjustment. For edge user rates before and after adjustment; The historical statistical standard deviation of the average signal-to-noise ratio of the entire network; The historical statistical standard deviation of edge user rates; This is a weighting coefficient. Since the dimensional differences have been eliminated through standard deviation, this coefficient is only used to adjust the relative priority of coverage and capacity, and defaults to equal weight. If the performance gain difference is less than the preset minimum positive benefit threshold, i.e., the overall improvement is required to reach at least 0.2 standard deviations, the system determines that the current environmental model has a deviation and triggers a model correction command. This situation usually means that there is a deviation between theoretical calculations and the actual environment, such as changes in the reflection characteristics of buildings. In response to the model correction command, the system adjusts the multipath fading factor in the spatiotemporal topology reconstruction module and re-executes the interference potential energy analysis and intelligent decision optimization process; the multipath fading factor here... Specifically, this relates to the ray tracing model in Example 4, as the reflection coefficient. The global correction item, i.e. During system initialization, the multipath fading factor... The default value is set to When performance is not up to standard, adjust accordingly. The value changes the intensity of the multipath signal in the model, thus forcing the optimization algorithm to find the beam solution under more stringent disturbance constraints; the adjustment strategy adopts a discretized PID feedback control algorithm.

[0169]

[0170] in, As a multipath fading factor, This refers to the performance gain deviation, i.e., the preset minimum positive return threshold. Difference from the currently calculated performance gain difference, The system recommends setting the proportional and integral coefficients as follows: This parameter combination was obtained by tuning the multipath fading channel model after performing a step response test using the Ziegler-Nichols method. For discrete sampling time index, To control the cycle; to ensure that the model parameters conform to physical laws, the updated... Apply boundary constraints: If but ,like but Through this feedback loop, when the actual performance is lower than expected, i.e., the error is... The system automatically increases the parameters of the reflection model, making the interference assessment in the virtual environment closer to the real physical environment or more conservative, thereby improving the robustness of the system.

[0171] The closed-loop control mechanism established in this embodiment can detect negative gain and automatically correct environmental model parameters when unforeseen changes in the environment cause the optimization strategy to fail. This avoids the continued effectiveness of erroneous strategies and ensures the long-term robustness of network performance in dynamic environments.

[0172] The system also includes an energy efficiency management module, used for:

[0173] Based on the 3D call heat map, the total call demand at the current moment is calculated.

[0174] Compare the total call volume demand with the preset energy-saving activation threshold;

[0175] If the total traffic demand is less than the energy-saving start threshold, a channel shutdown command is generated to lock some antenna array channels and recalculate the optimal antenna weight configuration of the remaining channels to maintain the basic coverage waveform.

[0176] If the total call demand is greater than or equal to the energy-saving activation threshold, then the full-channel operation mode will be maintained.

[0177] This embodiment details the specific logic of the energy efficiency management module, which achieves deep energy saving based on load. The system uses a three-dimensional traffic heatmap to calculate the total traffic demand at the current moment. The system compares the total traffic demand with a preset energy-saving threshold. If the total traffic demand is less than the energy-saving threshold, the system generates a channel shutdown command, locking some antenna array channels. The specific locking strategy uses a sparse array masking mechanism, i.e., generating a mask of length... binary mask vector ,in, Indicates that the switch is off. This indicates that the array is enabled; to prevent the grating lobe effect and maintain the main lobe directivity, this embodiment adopts a random sparse sampling strategy, that is, while keeping the array elements at both ends enabled, i.e. Under the premise that the remaining array elements are randomly set to 0 with a 50% probability; or a fixed-interval shutdown strategy is adopted, that is, let And recalculate the optimal antenna weight configuration for the remaining channels, that is, the number of array elements in the optimization algorithm. Corrected to the number of activated array elements And in the beamforming formula, only the following is accumulated. The corresponding items are used to maintain the basic coverage waveform; for example, some channels are shut down during off-peak hours at night, but the key is that the weights of the remaining channels will be re-optimized after the shutdown to prevent coverage black holes; in response to the total call demand being greater than or equal to the energy-saving start threshold, the system will maintain full-channel operation mode.

[0178] This embodiment achieves deep energy saving based on traffic load while ensuring basic coverage. Unlike the traditional simple hard shutdown, this system re-optimizes the weights of the remaining channels after shutting down some channels, ensuring that energy saving is not reduced and significantly reducing the operating energy consumption of the base station.

[0179] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A smart base station antenna system, characterized in that, include: The multi-source sensing fusion module is used to collect the current operating parameter data of the base station antenna and the measurement report data from the terminal in real time. It parses and cleans the measurement report data, extracts the signal strength, signal-to-noise ratio, angle of arrival and time advance data, and aligns the current operating parameter data with the extracted data in time sequence to generate a multi-dimensional sensing dataset. The spatiotemporal topology reconstruction module is used to invert the multipath propagation environment in virtual space based on multidimensional sensing datasets and combined with geographic information system data, and to construct a three-dimensional traffic heat map that reflects the distribution of user density in three-dimensional space. The interference potential energy analysis module is used to simulate the signal overflow situation of the current beam pattern at the edge of the neighboring cell based on the three-dimensional traffic heat map, calculate the overlap coverage, and generate interference potential energy assessment results accordingly. The intelligent decision optimization module is used to construct a multi-objective optimization function with the goal of maximizing the signal strength of the hotspot area of ​​the cell and minimizing the interference potential energy of the neighboring cell. It performs a full-domain search on the three-dimensional traffic heat map and calculates the optimal antenna weight configuration. The adaptive reconfiguration execution module is used to convert the optimal antenna weight configuration into beamforming commands or electrically adjustable downtilt commands, drive the reconfigurable antenna actuator to adjust the radiation pattern, and monitor the adjusted network performance indicators in real time to form a closed-loop control.

2. The intelligent base station antenna system according to claim 1, characterized in that, Methods for generating multidimensional sensing datasets include: Obtain scheduling information for each transmission time interval recorded by the base station to determine the effective service period; Sampling points that fall within the valid business hours are selected from the measurement report data; Based on a preset statistical deviation threshold or interquartile range rule, outlier removal is performed on the selected sampling points to obtain valid sampling points. The coordinate system is transformed between the angle of arrival data corresponding to the effective sampling point and the normal direction of the base station antenna to obtain the absolute spatial angle; The absolute spatial angle, signal strength, and signal-to-noise ratio are mapped to the current engineering parameter data according to the timestamp index, thus completing the generation of the multidimensional sensing dataset.

3. The intelligent base station antenna system according to claim 2, characterized in that, Methods for constructing a 3D call traffic heatmap include: The coverage area is divided into three-dimensional grid cells of preset size; Based on the time lead data, the radial distance between the user and the base station is estimated, and combined with the absolute spatial angle, the spatial projection position of each effective sampling point in the three-dimensional grid cell is determined. Density clustering analysis is performed on the effective sampling points within each three-dimensional grid cell to calculate the traffic density value of each three-dimensional grid cell. Kriging interpolation or Gaussian process regression is used to smoothly fill sparse regions, generating a continuous three-dimensional traffic heatmap.

4. The intelligent base station antenna system according to claim 3, characterized in that, Methods for generating disturbance potential assessment results include: Obtain the location information of neighboring base stations and the geographical boundary information of the neighboring cell edges; Mark the interference-sensitive areas that belong to the coverage area of ​​the neighboring cell in the 3D traffic heat map; The power spectral density distribution of the current beam configuration in the interference-sensitive region was simulated using a ray tracing model. Calculate the cumulative received power in the interference-sensitive area and compare it with the preset interference tolerance threshold; If the accumulated received power is greater than the interference tolerance threshold, a high interference potential energy marker is generated, and the interference overflow of the excess portion is calculated. If the cumulative received power is less than or equal to the interference tolerance threshold, a low interference potential energy marker is generated. High or low disturbance potential energy markers are used as the results of disturbance potential energy assessment.

5. The intelligent base station antenna system according to claim 4, characterized in that, Methods for calculating the optimal antenna weight configuration include: Initialize a random population of antenna weights, each containing amplitude weights and phase weights; Substitute the antenna weights of each group into the preset antenna pattern synthesis formula to generate the corresponding virtual beamform. Calculate the target coverage gain of the virtual beamform in the three-dimensional traffic heatmap, and calculate the fitness function value that is positively correlated with the target coverage gain and negatively correlated with the interference spillover amount according to the preset weighting logic. Based on the fitness function value, the antenna weight population is selected, crossovered and mutated to generate a new generation of antenna weight population; Repeat the iteration until the fitness function value converges or the preset number of iterations is reached, and select the set of antenna weights with the highest fitness function value as the optimal antenna weight configuration.

6. The intelligent base station antenna system according to claim 5, characterized in that, Methods for driving reconfigurable antenna actuators to adjust radiation patterns include: Analyze the optimal antenna weight configuration to separate the phase adjustment parameter and the amplitude adjustment parameter; If the interference potential energy assessment result is a high interference potential energy marker, then based on the phase adjustment parameter, a radiation null point is generated in the physical direction of the corresponding interference sensitive area to form a null trap beam. If the displacement of the spatial geometric centroid of the call hotspot area in the three-dimensional call heat map exceeds the preset threshold within the preset time window, it is determined to be a tidal distribution change. Based on the amplitude adjustment parameter, the beam main lobe width is dynamically compressed or widened to match the geometric envelope of the call hotspot area. The generated null beam configuration or main lobe width configuration is sent to the phase shifter and attenuator network to complete the adjustment of the radiation pattern.

7. The intelligent base station antenna system according to claim 1, characterized in that, Methods for real-time monitoring and adjustment of network performance metrics to form closed-loop control include: Within the preset observation window after adjusting the radiation pattern, re-acquire measurement report data; Calculate the adjusted average signal-to-noise ratio for the entire network and the edge user rate; Calculate the performance gain difference before and after the adjustment; If the performance gain difference is less than the preset minimum positive return threshold, it is determined that there is a deviation in the current environment model, and a model correction instruction is triggered. In response to the model correction command, the multipath fading factor in the spatiotemporal topology reconstruction module is adjusted, and the disturbance potential energy analysis and intelligent decision optimization process is re-executed.

8. The intelligent base station antenna system according to claim 1, characterized in that, The system also includes an energy efficiency management module for: Based on the 3D call heat map, the total call demand at the current moment is calculated. Compare the total call volume demand with the preset energy-saving activation threshold; If the total traffic demand is less than the energy-saving start threshold, a channel shutdown command is generated to lock some antenna array channels and recalculate the optimal antenna weight configuration of the remaining channels to maintain the basic coverage waveform. If the total call demand is greater than or equal to the energy-saving activation threshold, then the full-channel operation mode will be maintained.

Images