Method for processing service message based on multi-band network, server and medium

Abstract

This application relates to the field of computer communication technology, and provides a service message processing method, server, and medium based on a multi-band network. The method includes: acquiring a mixed primary and secondary frequency signal stream formed by receiving primary and secondary frequency signals at a communication station; performing reception behavior logic parsing processing on the mixed primary and secondary frequency signal stream, extracting parallel detection state transition sequences and time-division frequency selection dwell decision sequences, and generating a reception behavior logic descriptor; performing behavior trajectory classification processing on the reception behavior logic descriptor through a station behavior clustering discrimination model, dividing the communication station into a station cluster and generating a station cluster affiliation identifier; determining the scheduling priority order of cluster-level service messages and the cluster-level multi-band resource occupancy coordination strategy based on the station cluster affiliation identifier; generating service message distribution control instructions based on the scheduling priority order and occupancy coordination strategy and pushing them to the communication station to adjust its service message transmission and reception window.

Description

Technical Field

[0001] This application belongs to the field of computer communication technology, specifically relating to a service message processing method, server, and medium based on a multi-band network. Background Technology

[0002] In multi-band communication networks, communication stations are typically configured to work collaboratively on primary and secondary frequency bands. During the network deployment phase, the primary frequency band is generally used to complete the entire network topology construction. During the service communication phase, stations receive mixed primary and secondary frequency signals across the entire bandwidth based on their physical layer capabilities. Stations with parallel detection capabilities can demodulate signals from both frequency bands simultaneously, while stations without this capability need to evaluate the signal quality of both bands before selecting one to receive. This results in two distinct physical layer processing behavior modes at the receiving end: parallel detection and time-division multiplexing. Furthermore, different stations exhibit significant heterogeneity in frequency band switching delay tolerance and channel dwell time preferences due to differences in hardware architecture, firmware version, and deployment environment.

[0003] Existing technologies typically employ a uniform scheduling priority rule and static resource allocation template across the entire network when scheduling service packets for multiple sites, failing to identify the actual receiving behavior and logical preferences of each site. When there are a large number of sites with heterogeneous behavioral preferences in the network, the aforementioned coarse-grained scheduling method can easily lead to a mismatch between the service packet transmission and reception windows of some sites and their physical layer processing characteristics, resulting in insufficient utilization of channel resources and decreased service packet distribution efficiency. Summary of the Invention

[0004] This application provides a service packet processing method, server, and medium based on a multi-band network, which enables the service packet distribution and control instructions to not only meet the different tendencies of different clusters in multi-band parallel detection and time-division frequency selection, but also dynamically balance the allocation ratio of primary and secondary frequency band channel resources among the clusters. Under the premise of ensuring the fairness of the whole network scheduling, it improves the service packet distribution efficiency and spectrum resource utilization balance in multi-band fusion coverage scenarios.

[0005] This application provides a service packet processing method based on a multi-band network, applied to a service packet processing server, the method comprising: The system acquires a mixed signal stream of primary and secondary frequency bands formed by the primary frequency band signals and secondary frequency band signals received by the communication station within a preset monitoring period. The mixed signal stream of primary and secondary frequency bands includes multiple continuously captured mixed signal frame units, and each mixed signal frame unit carries a beacon frame identifier and a set of physical layer capability declaration fields. The receiving behavior logic parsing process is performed on the main and auxiliary frequency mixed signal stream to extract the parallel detection state transition sequence and time-division frequency selection dwell decision sequence when the communication station responds to the main frequency band signal and the auxiliary frequency band signal, and to generate a receiving behavior logic descriptor that reflects the physical layer processing behavior preference of the corresponding communication station. The receiving behavior logical descriptor is classified into behavior trajectories by a preset site behavior clustering discrimination model, and the communication site is divided into a site cluster whose corresponding receiving behavior similarity meets the preset aggregation criterion, and a site cluster affiliation identifier code is generated. The scheduling priority order of cluster-level service messages corresponding to the communication site and the cluster-level multi-band resource occupation coordination strategy are determined based on the site cluster affiliation identifier code. Based on the scheduling priority order of the cluster-level service packets and the cluster-level multi-band resource occupancy coordination strategy, a service packet distribution control instruction is generated for the communication station, and the service packet distribution control instruction is pushed to the communication station through the main frequency band signal or the auxiliary frequency band signal to adjust its service packet transmission and reception window.

[0006] This application provides a service message processing server, which includes a processor and a memory, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the above-described method.

[0007] This application provides a computer-readable storage medium including a computer program. When the computer program is run on a service message processing server, the computer program is used to cause the service message processing server to perform the steps of the above-described method.

[0008] This application analyzes the parallel detection state transition sequence and time-division frequency selective dwell decision sequence of communication sites from the primary and secondary frequency mixed signal stream to generate a receiving behavior logic descriptor to finely characterize the physical layer processing behavior preferences of the sites. It then uses a site behavior clustering discrimination model to group sites with similar receiving behavior logic patterns into the same site cluster, uniformly outputting a site cluster affiliation identifier at the cluster level. This transforms single-site behavior deviation noise into cluster statistical characteristics for scheduling decisions. Based on the differentiated site cluster affiliation identifier, the scheduling priority order of cluster-level service packets and the cluster-level multi-band resource allocation coordination strategy are determined. This ensures that the service packet distribution control instructions not only match the differentiated tendencies of different clusters in multi-band parallel detection and time-division frequency selective dwell, but also dynamically balance the allocation ratio of primary and secondary frequency band channel resources among the clusters. While ensuring the fairness of network-wide scheduling, this improves the service packet distribution efficiency and spectrum resource utilization balance in multi-band fusion coverage scenarios. Attached Figure Description

[0009] Figure 1 This is a flowchart illustrating a service packet processing method based on a multi-band network, as provided in an embodiment of this application.

[0010] Figure 2 This is a schematic diagram of the structure of a service message processing server provided in an embodiment of this application.

[0011] Figure 3 This is a functional block diagram of a service message processing server provided in an embodiment of this application. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this application. Obviously, the described embodiments are only some embodiments of the technical solutions of this application, and not all embodiments. Based on the embodiments recorded in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the technical solutions of this application.

[0013] See Figure 1 This is a service packet processing method based on a multi-band network provided in the embodiments of this application. The method can be applied to a service packet processing server, and the specific process is as follows: steps 110-150.

[0014] The service packet processing method based on a multi-band network disclosed in this application is executed by a service packet processing server. This service packet processing server maintains a bidirectional communication connection with communication stations in the network through primary and secondary frequency band channels. During operation, the communication station continuously captures wireless signals in space. When the communication station is simultaneously within the coverage area of ​​both primary and secondary frequency band signals, its receiving link will form a mixed primary and secondary frequency signal stream. This mixed signal stream is periodically reported to the service packet processing server, which then performs subsequent operations such as receiving behavior parsing, station clustering, scheduling strategy generation, and distribution and control command push.

[0015] Step 110: Obtain the main and auxiliary frequency mixed signal stream formed by the main frequency band signal and the auxiliary frequency band signal received by the communication station within a preset monitoring period. The main and auxiliary frequency mixed signal stream contains multiple continuously captured mixed signal frame units, and each mixed signal frame unit carries a beacon frame identifier and a set of physical layer capability declaration fields.

[0016] The acquisition of the primary and secondary frequency hybrid signal stream originates from the continuous scanning of the receiving link of the communication station across the entire frequency band. The communication station continuously samples and performs frame synchronization detection on the air interface signal according to a preset monitoring period. When the communication station captures signal components from both the primary and secondary frequency band carrier channels within the same monitoring window, the analog-to-digital converter module of the receiving link converts the analog baseband waveforms of the two frequency bands into digital sampling sequences. The physical layer frame parsing module then performs frame delimiting and demapping operations on the digital sampling sequences to extract the hybrid signal frame unit that completely carries the Media Access Control layer protocol data unit.

[0017] Each hybrid signal frame unit includes a frame header and a payload. The frame header carries a beacon frame identifier, which indicates the basic service set identifier, primary band channel number, and secondary band channel number to which the current hybrid signal frame unit belongs. The payload contains a set of physical layer capability declaration fields, which consists of several fixed-length capability indication bits. Each capability indication bit occupies a preset number of bits and is used to declare the characteristics of the communication station in terms of physical layer processing capabilities such as multi-band parallel reception, band switching delay tolerance, and channel dwell time preference.

[0018] The service message processing server retrieves multiple consecutive mixed signal frame units in batches from the protocol stack management entity of the communication site through the backhaul link. These mixed signal frame units are arranged in the order of the timestamps in their respective frame control fields to form a main and auxiliary frequency mixed signal stream, which serves as the raw input data for the receiving behavior logic parsing and processing.

[0019] Step 120: Perform reception behavior logic parsing processing on the main and auxiliary frequency mixed signal stream, extract the parallel detection state transition sequence and time-division frequency selection dwell decision sequence when the communication station responds to the main frequency band signal and the auxiliary frequency band signal, and generate a reception behavior logic descriptor that reflects the physical layer processing behavior preference of the corresponding communication station.

[0020] The core of the receiving behavior logic parsing process is to convert the frequency band occupancy records contained in the primary and secondary frequency mixed signal stream into the receiving strategy behavior sequence actually adopted by the communication site.

[0021] The service message processing server first parses the beacon frame identifier carried by each mixed signal frame unit in the primary and secondary frequency mixed signal stream to determine the basic service set identifier to which each mixed signal frame unit belongs and the corresponding primary frequency band channel number and secondary frequency band channel number. Then, it extracts the values ​​of several key capability indication bits from the physical layer capability declaration field set, including multi-band parallel reception capability indication bit, frequency band switching delay tolerance level indication bit, and channel dwell time preference indication bit.

[0022] The multi-band parallel reception capability indicator bit is used to determine whether the communication station has the hardware processing capability to perform synchronous demodulation of the main frequency band signal and the auxiliary frequency band signal; the frequency band switching delay tolerance level indicator bit is used to quantify the minimum time interval tolerance required for the communication station to switch between parallel detection state and time-division frequency selective dwell state; the channel dwell time preference indicator bit reflects the strength of the communication station's tendency to maintain continuous reception operation on a single frequency band channel.

[0023] After extracting the above parameters, the service message processing server unfolds the primary and secondary frequency mixed signal stream along the time axis into a continuous logical time sequence. For each logical time, it performs state marking processing based on the changes in the channel number of the adjacent mixed signal frame units before and after that time. If only the primary frequency band channel number changes between adjacent logical times while the secondary frequency band channel number remains unchanged, it is marked as a parallel detection state; if both the primary and secondary frequency band channel numbers change between adjacent logical times, it is marked as a time-division frequency selective camping state. The state marking results of all continuous logical times are serialized and integrated to obtain the parallel detection state transition sequence and the time-division frequency selective camping decision sequence, and combined with the previously extracted capability indicator bit values ​​to generate a receiver behavior logical descriptor.

[0024] The encoding format of the receive behavior logic descriptor includes a behavior tendency identifier field and a dwell feature description field. The behavior tendency identifier field records the quantized value of the communication station's selection bias between parallel detection and time-division frequency selection, and the dwell feature description field records the quantized value of the communication station's stickiness to the primary frequency band signal channel or the secondary frequency band signal channel in the time-division frequency dwell state.

[0025] Step 121: Parse the beacon frame identifier carried by each mixed signal frame unit in the primary and secondary frequency mixed signal stream, and determine the basic service set identifier to which the mixed signal frame unit belongs and the corresponding primary frequency band channel number and secondary frequency band channel number.

[0026] In the mixed signal stream of primary and secondary frequencies, the header region of each mixed signal frame unit contains a beacon frame identifier. The encoding rules of the beacon frame identifier follow a preset bit mapping definition table. The service message processing server maintains a global channel number index table, which establishes a triplet mapping association between the basic service set identifier, the primary frequency band channel number, and the secondary frequency band channel number.

[0027] During the parsing process, the service message processing server performs a bitmask extraction operation on the beacon frame identifier of each mixed-signal frame unit. It performs a bitwise AND operation between the predefined identifier mask and the beacon frame identifier to separate the original bit sequences of the basic service set identifier field, the primary band channel number field, and the secondary band channel number field. Then, each original bit sequence is converted into an integer channel number value according to the decoding rules of the corresponding field. For mixed-signal frame units where identifier parsing fails, the service message processing server marks them as incomplete frame units and removes them from the subsequent status mark sequence to ensure the consistency and reliability of the received behavior logic parsing results.

[0028] Step 122: Extract the multi-band parallel reception capability indicator bit, the band switching delay tolerance level indicator bit, and the channel dwell time preference indicator bit from the physical layer capability declaration field set carried by the mixed signal frame unit, and determine whether the communication site supports the synchronous demodulation operation of the main frequency band signal and the auxiliary frequency band signal based on the multi-band parallel reception capability indicator bit.

[0029] The physical layer capability declaration field set consists of several sets of fixed-length capability indicator bits, which are arranged sequentially in a predefined order within the payload area of ​​the mixed-signal frame unit. The service message processing server parses the structure definition descriptor of the physical layer capability declaration field set to obtain the offset address and bit width parameters of the multi-band parallel reception capability indicator bits, the offset address and bit width parameters of the band switching delay tolerance level indicator bits, and the offset address and bit width parameters of the channel dwell time preference indicator bits.

[0030] Based on their respective offset addresses and bit width parameters, the service message processing server uses bit truncation operations to separate the original bit values ​​of each indicator bit from the bit stream of the physical layer capability declaration field set, and then maps the original bit values ​​to the corresponding capability level values ​​according to their respective quantization encoding tables. For multi-band parallel reception capability indicator bits, if the mapped capability level value is greater than or equal to the preset minimum parallel reception capability threshold, it is determined that the communication station supports synchronous demodulation operations of the primary frequency band signal and the secondary frequency band signal; if the capability level value is less than the minimum parallel reception capability threshold, it is determined that the communication station does not support synchronous demodulation operations.

[0031] Step 123: Obtain the state transition tolerance time parameter required for the communication station to transition from parallel detection state to time-division frequency selective dwell state when receiving concurrent primary frequency band signals and secondary frequency band signals, based on the frequency band switching delay tolerance level indicator bit.

[0032] The quantization encoding table of the band switching delay tolerance level indicator defines multiple discrete delay levels, each corresponding to a nominal delay time interval. The service message processing server reads the delay level corresponding to the band switching delay tolerance level indicator reported by the communication station, retrieves the upper and lower bounds of the nominal delay time interval corresponding to that delay level from the delay level mapping database, and takes the median of the nominal delay time interval as the state transition tolerance time parameter of the communication station.

[0033] The state transition tolerance time parameter is in milliseconds. The smaller the value, the higher the response speed requirement of the communication station to concurrent dual-band signals, and the more inclined it is to quickly transition to the time-division frequency-selective dwell state of a single band after detecting parallel signals. The larger the value, the longer the parallel processing tolerance capability of the communication station, and the longer it can maintain the parallel detection state under the same conditions.

[0034] Step 124: Determine the duration parameter of the continuous dwell tendency of the communication station to maintain continuous reception operation on the main frequency band signal channel in the time-division frequency-selective dwell state based on the channel dwell duration preference indicator bit.

[0035] The quantization encoding table for the channel dwell time preference indicator also defines a set of graded values ​​corresponding to the dwell tendency strength. The service message processing server looks up the corresponding dwell tendency grade identifier from the dwell tendency mapping relationship library based on the bit value of this indicator. Different dwell tendency grade identifiers establish a mapping relationship with a baseline dwell time interval.

[0036] Based on the median of the baseline dwell time interval, adaptive adjustment is performed in conjunction with the channel quality indicator (CMI) assessment value of the channel in the current main frequency band of the communication site: if the CMI assessment value is higher than the preset high-quality channel judgment threshold, the median is multiplied by an extension factor greater than 1.0 to obtain the continuous dwell tendency duration parameter; if the CMI assessment value is lower than the preset low-quality channel judgment threshold, the median is multiplied by a shortening factor less than 1.0 to obtain the continuous dwell tendency duration parameter.

[0037] The physical meaning of the persistent dwell tendency duration parameter is the length of time that a communication station is expected to continuously dwell on the main frequency band signal channel in a single time-division frequency-selective dwell state. This parameter directly affects the duration marker value of the main frequency band dwell segment in the subsequent time-division frequency-selective dwell decision sequence.

[0038] Step 125: Combine the state marking results of the continuous logical moments of the main and auxiliary frequency mixed signal streams, and the correlation results of the duration parameter of the continuous dwell tendency and the duration parameter of the parallel detection state, to generate a receiving behavior logical descriptor that reflects the physical layer processing behavior preference of the communication station.

[0039] The generation process of the receive behavior logic descriptor integrates the temporal behavior characteristics in the state marking results with the tendency parameters extracted from the physical layer capability indicator bits. First, the main and auxiliary frequency mixed signal streams are sequentially marked with logical states in time order. Then, statistical analysis and parameter correlation operations are performed on the marked state sequence to finally construct a composite descriptor structure that includes a behavior tendency identifier field and a dwell characteristic description field.

[0040] Step 1251: Perform logical state marking processing on the main and auxiliary frequency mixed signal streams in chronological order. If the beacon frame identifier of an adjacent mixed signal frame unit changes the main frequency band channel number while the auxiliary frequency band channel number remains unchanged, mark the current logical state as parallel detection state and record the corresponding main frequency band channel feature code change information. If the beacon frame identifier of an adjacent mixed signal frame unit changes both the main frequency band channel number and the auxiliary frequency band channel number, mark the current logical state as time-division frequency-selective camping state and record the corresponding dual-frequency band channel linkage feature code change information.

[0041] Status marking processing depends on the change pattern of beacon frame identifiers between adjacent mixed signal frame units in the primary and secondary frequency mixed signal stream.

[0042] The service message processing server traverses the entire primary and secondary frequency mixed signal stream and performs channel number differential comparison on each pair of temporally adjacent mixed signal frame units: the primary frequency band channel number of the next frame unit is subtracted from the primary frequency band channel number of the previous frame unit to obtain the primary frequency band differential value, and the secondary frequency band channel number of the next frame unit is subtracted from the secondary frequency band channel number of the previous frame unit to obtain the secondary frequency band differential value.

[0043] When the absolute value of the primary frequency band differential value is greater than 0 and the absolute value of the secondary frequency band differential value is equal to 0, the logical moment between the adjacent frame units is marked as a parallel detection state. At the same time, the primary frequency band channel number of the next frame unit is concatenated with the primary frequency band channel number of the previous frame unit to form a two-byte primary frequency band channel feature code change information and recorded in the auxiliary field of the state mark entry.

[0044] When the absolute value of the difference in the main frequency band is greater than 0 and the absolute value of the difference in the auxiliary frequency band is also greater than 0, the logical time between the adjacent frame units is marked as a time-division frequency-selective dwell state. The main frequency band channel number and the auxiliary frequency band channel number of the next frame unit are combined with the corresponding channel number of the previous frame unit to generate dual-frequency band channel linkage feature code change information. The encoding method of the dual-frequency band channel linkage feature code change information is to perform an XOR operation on the two sets of channel number pairs respectively and then concatenate them in order.

[0045] Step 1252: Integrate the state marking results of continuous logic moments to generate a state transition sequence that includes parameters such as the duration of parallel detection state, the triggering time information of time-division frequency-selective dwell state, and the transition interval parameter between adjacent states.

[0046] After completing the state marking between all adjacent mixed signal frame units, the service message processing server performs continuous state segment merging on the marking results. It iterates through the state marking result sequence, merging multiple logical moments with the same state marking value and adjacent time into a single state segment. The starting logical moment of each state segment is marked as the segment start moment, and the ending logical moment is marked as the segment end moment. The difference between the segment end moment and the segment start moment is recorded as the duration parameter of that state segment.

[0047] If the flag value of a state segment is parallel detection state, its duration parameter is recorded as the parallel detection state duration parameter; if the flag value of a state segment is time-division frequency-selective dwell state, its duration parameter is recorded as the time-division frequency-selective dwell state duration parameter, and the start time of the segment is recorded as the time-division frequency-selective dwell state trigger time information.

[0048] The difference between the start time of a pair of adjacent parallel detection state segments and a time-division frequency-selective dwell state segment is recorded as the adjacent state transition interval parameter. The type, duration parameter, trigger time information, and adjacent state transition interval parameter of all state segments are stored sequentially according to the order in which the segments occur, forming a state transition sequence.

[0049] Step 1253: Analyze the correlation logic between the transition frequency of the parallel detection state transition to the time-division frequency-selective dwell state in the statistical state transition sequence and the state transition tolerance time parameter, and generate a probability distribution descriptor to characterize the behavior tendency of the communication station to maintain the parallel detection state or to trigger the time-division frequency-selective dwell state under different channel load conditions.

[0050] The service message processing server performs transition frequency statistics on each pair of adjacent state segments in the state transition sequence. If the subsequent adjacent segment of a parallel detection state segment is a time-division frequency-selective resident state segment, it is counted as a positive transition event; if the subsequent adjacent segment of a time-division frequency-selective resident state segment is a parallel detection state segment, it is counted as a negative transition event. Within the same preset monitoring period, the channel load index corresponding to the transition times of all positive transition events is summarized. This channel load index is calculated by querying the combined weighted value of the service message queuing depth and channel busy-idle ratio of the primary and secondary frequency band channels of the basic service set in which the communication station is located.

[0051] Using the normalized results of the channel load index value at the transition moment and the state transition tolerance time parameter as binary condition variables, the conditional probability estimation method is used to calculate the conditional probability distribution of positive transition events occurring in different channel load index intervals and different transition tolerance time intervals, respectively, to obtain the behavior trend probability distribution descriptor. The behavior trend probability distribution descriptor is a set of discrete key-value pairs, each key-value pair consisting of a joint key of a load interval identifier code and a transition interval identifier code and the corresponding positive transition probability value.

[0052] Step 1254: The relative proportion between the persistent dwell tendency duration parameter and the parallel detection state duration parameter is correlated, and a receiving behavior logical descriptor reflecting the physical layer processing behavior preference of the corresponding communication station is generated by combining the behavior tendency probability distribution descriptor. The receiving behavior logical descriptor includes a behavior tendency identifier field and a dwell feature description field.

[0053] The final generation of the receiving behavior logic descriptor associates the duration parameter of the persistent dwell tendency obtained in step 124 with the duration parameter of the parallel detection state obtained in step 1252.

[0054] Specifically, the service message processing server obtains the continuous dwell tendency duration sequence and the parallel detection state duration sequence recorded by the communication station in multiple consecutive state transition sequences, performs time-series smoothing processing on the two duration sequences to eliminate instantaneous fluctuation interference, and generates a stable characterization value of dwell tendency duration and a benchmark reference characterization value of parallel detection state duration.

[0055] Step 12541: Obtain the duration sequence of the continuous dwell tendency and the duration sequence of the parallel detection state recorded by the communication station in multiple consecutive state transition sequences.

[0056] Each communication station generates a state transition sequence within each preset monitoring cycle. Each state transition sequence contains several records of continuous dwell tendency duration and parallel detection state duration. The continuous dwell tendency duration records belonging to the same communication station within N consecutive preset monitoring cycles are concatenated end-to-end in monitoring cycle order to form a continuous dwell tendency duration sequence. Similarly, the parallel detection state duration records within these N cycles are concatenated in the same order to form a parallel detection state duration sequence. The value of N is determined by the configuration parameters of the service message processing server.

[0057] Step 12542: Perform time-series smoothing on the duration sequence of the persistent dwell tendency and the duration sequence of the parallel detection state, respectively, to generate stable characterization values ​​of the dwell tendency duration and benchmark reference characterization values ​​of the duration of the parallel detection state.

[0058] The temporal smoothing process involves applying an exponentially weighted moving average function to the duration sequence of persistent dwell tendency. Higher smoothing weights are assigned to duration records closer to the current time, with these weights decreasing exponentially with increasing time. The resulting weighted average value is the stable representation of the dwell tendency duration. The same exponentially weighted moving average function with the same parameters is applied to the duration sequence of parallel detection states to obtain a benchmark reference value for the duration of parallel detection states.

[0059] Step 12543: Calculate the relative proportion between the dwell tendency duration stability characterization value and the benchmark reference characterization value, and input the relative proportion value into the preset dwell preference classification logic unit to obtain the dwell stickiness label used to describe the degree of stickiness of the communication station to the main frequency band signal channel.

[0060] The relative proportion value is obtained by dividing the stable dwell time characterization value by the reference characterization value. The range of the relative proportion value reflects the dependence of the communication site on the main frequency band signal channel during the time-division frequency-selective dwell period.

[0061] The business message processing server has a built-in dwell preference classification logic unit. This logic unit maps the relative proportion value to discrete dwell stickiness indicators based on a set of preset classification threshold boundaries. For example, when the relative proportion value is greater than the first classification threshold, the dwell stickiness indicator is a high stickiness level code; when the relative proportion value is between the second classification threshold and the first classification threshold, the dwell stickiness indicator is a medium stickiness level code; and when the relative proportion value is lower than the second classification threshold, the dwell stickiness indicator is a low stickiness level code.

[0062] Step 12544: Analyze the change trajectory of the behavior trend identifier field in the behavior trend probability distribution descriptor, and extract the change curve of the transition trigger logic sensitivity of the communication station from parallel detection state to time-division frequency selective dwell state during the continuous monitoring period.

[0063] The behavior trend identifier field in the behavior trend probability distribution descriptor stores the positive transition probability value corresponding to the joint key of each load interval identifier code and transition interval identifier code. These probability values ​​are arranged sequentially along the monitoring period time axis to form the change trajectory of the behavior trend identifier field.

[0064] The service message processing server uses this change trajectory as the original discrete sampling point sequence, and connects adjacent sampling points through a piecewise linear interpolation method to form a transition trigger logic sensitivity change curve that changes continuously on the time axis. This curve provides a dynamic trend estimate of the transition response of the communication site to the channel load change to the subsequent time window in each continuous monitoring period.

[0065] Step 12545: Perform segmented feature extraction processing on the transition trigger logic sensitivity change curve to obtain the segmented sensitivity difference distribution information under different combinations of main frequency band channel feature codes and auxiliary frequency band channel feature codes.

[0066] The segmented feature extraction process, based on different combinations of the primary band channel feature code and the secondary band channel feature code, extracts the actual feature code combinations that appear in the primary band channel feature code change information and the dual-band channel linkage feature code change information recorded in step 1251 as classification keys. For each classification key, the mean and standard deviation of the transition trigger logic sensitivity change curve segment within the associated time period are calculated. These statistical feature pairs are used as the segmented sensitivity representation of that feature code combination. All classification keys and their corresponding statistical feature pairs together constitute the segmented sensitivity difference distribution information output in this step.

[0067] Step 12546: Combining the residence stickiness identifier and the segmentation sensitivity difference distribution information, construct a behavior preference mapping vector with the residence tendency dimension and the state transition sensitivity dimension as coordinate axes using a predefined behavior preference classification mapping table.

[0068] The predefined behavioral preference classification mapping table is a lookup table structure. The row index of this lookup table is the discrete value of the dwell stickiness identifier, and the column index is the discretized interval code corresponding to each statistical feature quantity in the segmented sensitivity difference distribution information. Using the currently calculated dwell stickiness identifier as the row index and the sensitivity discretized interval code corresponding to the dominant frequency band feature code combination in the segmented sensitivity difference distribution information as the column index, a two-dimensional coordinate value is read from the lookup table. The first coordinate component of this two-dimensional coordinate value is the reference projection length of the dwell tendency dimension, and the second coordinate component is the reference projection length of the state transition sensitivity dimension. The behavioral preference mapping vector is composed of the above two reference projection lengths.

[0069] Step 12547: Project the behavior preference mapping vector along the residence tendency dimension using the principal component direction, and retain the feature projection component with the maximum behavior discernibility in the projection direction.

[0070] The process of principal component direction projection is as follows: calculate the covariance matrix for the set of behavioral preference mapping vectors belonging to different communication stations within the same station cluster, perform eigenvalue decomposition on the covariance matrix to obtain the eigenvector direction corresponding to the dwell tendency dimension, perform vector dot product operation on the behavioral preference mapping vector of the current communication station and the eigenvector direction, and the result of the dot product operation is the principal component direction projection value along the dwell tendency dimension. Take this projection value as the feature projection component. The value of the feature projection component represents the deviation level of the communication station from the average dwell tendency of all communication stations in the dwell tendency dimension.

[0071] Step 12548: Perform vector fusion processing on the feature projection components and the behavior preference mapping vector along the orthogonal components of the state transition sensitivity dimension to generate a receiving behavior logical descriptor that represents the physical layer processing behavior preference of the communication station. The behavior tendency identifier field in the receiving behavior logical descriptor stores the direction code of the feature projection components, and the dwell feature description field in the receiving behavior logical descriptor stores the amplitude code of the orthogonal components.

[0072] The orthogonal components of the behavior preference mapping vector along the state transition sensitivity dimension are obtained as follows: subtract the projection vector of the behavior preference mapping vector onto the principal component direction of the residence tendency dimension, and the magnitude of the residual vector is the magnitude of the orthogonal component.

[0073] The feature projection components are converted into direction codes with positive and negative signs. If the value of the feature projection component is positive, the direction code is a positive behavior trend code; if the value of the feature projection component is negative, the direction code is a negative behavior trend code. The amplitude of the orthogonal components is discretized by a quantizer and converted into a fixed-width amplitude code. The direction code and amplitude code are filled into the behavior trend identifier field and the resident feature description field of the receive behavior logic descriptor, respectively, and together they form the encoding body of the receive behavior logic descriptor.

[0074] Step 130: The receiving behavior logical descriptor is classified into behavior trajectories using a preset site behavior clustering discrimination model. The communication sites are divided into site clusters whose corresponding receiving behavior similarity meets the preset aggregation criteria, and a site cluster affiliation identifier code is generated.

[0075] The service message processing server loads the pre-trained site behavior clustering discrimination model. This model consists of three interconnected processing units: a feature encoding layer, a feature aggregation layer, and a cluster center distance calculation layer. The model receives the receive behavior logic descriptor generated in step 120 as input. Through sequential processing by the three processing units, it ultimately outputs a set of relative distance parameters between the communication station and each candidate site cluster, and completes the site cluster division based on preset aggregation criteria.

[0076] Step 131: Input the behavior trend identifier field and the dwell feature description field of the received behavior logical descriptor into the feature encoding layer of the site behavior clustering discrimination model. The feature encoding layer maps the behavior trend identifier field into a behavior trend feature vector and the dwell feature description field into a dwell feature vector.

[0077] The feature encoding layer comprises two parallel encoding modules: a behavior trend encoding module and a resident feature encoding module. The behavior trend encoding module consists of an embedding layer and a fully connected layer. The number of entries in the behavior trend identifier field of the embedding layer equals the total number of possible values ​​for the direction code. The embedding dimension is a pre-defined hyperparameter. The behavior trend identifier field is used to obtain an embedding vector through a lookup table in the embedding layer. The fully connected layer then performs a linear transformation and activation function rectification on the embedding vector to generate the behavior trend feature vector. The resident feature encoding module also consists of an embedding layer and a fully connected layer. The amplitude code of the resident feature description field is converted into an amplitude embedding vector by the embedding layer, and then mapped to the resident feature vector by the fully connected layer.

[0078] Step 132: The behavior trend feature vector and the residence feature vector are processed across dimensions by the feature aggregation layer of the site behavior clustering discrimination model to generate a joint representation feature tensor describing the communication site in the behavior trend dimension and the residence tendency dimension.

[0079] The feature aggregation layer concatenates the behavioral trend feature vector and the resident feature vector along the feature dimension axis. The length of the concatenated vector is equal to the sum of the feature dimensions of the two input vectors. The concatenated result is the joint representation feature tensor. Optionally, a batch normalization layer is added after the concatenation operation to perform a normalization operation on the joint representation feature tensor along the batch dimension. The normalization operation is used to eliminate the scale shift introduced by the different numerical distribution ranges of the behavioral trend feature vector and the resident feature vector.

[0080] Step 133: Invoke the cluster center distance calculation layer of the site behavior clustering discrimination model, and perform multi-dimensional spatial distance measurement calculation on the joint representation feature tensor and the cluster feature center vectors of multiple pre-established candidate site clusters to generate a set of relative distance parameters between the communication site and each candidate site cluster; wherein, each relative distance parameter in the set of relative distance parameters corresponds to a candidate site cluster and each relative distance parameter represents the degree of logical difference between the receiving behavior logic mode of the communication site and the typical receiving behavior logic mode of the corresponding candidate site cluster.

[0081] The cluster center distance calculation layer maintains a cluster feature center vector table, which stores the cluster feature center vectors of all candidate site clusters. The dimension of each cluster feature center vector is equal to the dimension of the joint representation feature tensor. The initial value is obtained by performing a dimension-wise arithmetic mean of the joint representation feature tensors of a small number of communication sites with known behavioral categories. For the current input joint representation feature tensor, the cluster center distance calculation layer calculates its Euclidean distance with each cluster feature center vector in the table. Before calculating the Euclidean distance, both are normalized using the L2 norm to unify the vector magnitude scale. The calculated Euclidean distance value is used as a relative distance parameter, and the set of relative distance parameters corresponding to all candidate site clusters constitutes the output of the cluster center distance calculation layer.

[0082] Step 134: Retrieve the smallest relative distance parameter in the set of relative distance parameters and compare the smallest relative distance parameter with the aggregation radius threshold specified by the preset aggregation criterion.

[0083] A preset aggregation criterion defines an aggregation radius threshold, which is the optimal clustering radius determined by the site behavior clustering discrimination model during the training phase through grid search and silhouette coefficient evaluation. The business message processing server traverses all relative distance parameters in the relative distance parameter set, selects the minimum relative distance parameter through bubble comparison, and then compares the minimum relative distance parameter with the aggregation radius threshold.

[0084] Step 135: If the minimum relative distance parameter is less than the aggregation radius threshold, the communication site is assigned to the candidate site cluster corresponding to the minimum relative distance parameter. If the minimum relative distance parameter is greater than or equal to the aggregation radius threshold, a new site cluster is created with the joint representation feature tensor as the initial cluster feature center vector.

[0085] When the minimum relative distance parameter is less than the aggregation radius threshold, it indicates that the receiving behavior logic pattern of the communication station is highly similar to the typical pattern of the candidate station cluster corresponding to the minimum relative distance parameter, and it is determined that the two belong to the same station cluster. When the minimum relative distance parameter is greater than or equal to the aggregation radius threshold, it indicates that there is no existing station cluster with sufficiently similar receiving behavior logic pattern of the communication station, and it is determined that a new station cluster needs to be created. The cluster feature center vector of the new station cluster directly uses the joint representation feature tensor of the communication station as the initial value.

[0086] Step 136: Update the cluster feature center vector of the divided site cluster or the newly added site cluster. Based on the newly added joint representation feature tensor, perform weighted shift adjustment on the original cluster feature center vector to generate the updated cluster feature center vector for subsequent classification of the behavior trajectory of communication sites.

[0087] The weighted move adjustment process is as follows: If a communication site is assigned to an existing site cluster, the original cluster feature center vector and the number of communication sites already included in the cluster are read. The newly added joint representation feature tensor and the original cluster feature center vector are then used to calculate the updated cluster feature center vector according to the mathematical expectation update formula. This update formula is: the new center vector equals the original center vector multiplied by the original number of sites divided by the original number of sites plus 1, plus the result of the newly added feature tensor divided by the original number of sites plus 1, i.e., Cnew = (Cold × n + Fnew) / (n + 1), where Cnew represents the updated cluster feature center vector, Cold represents the original cluster feature center vector, Fnew represents the joint representation feature tensor of the newly added communication site, and n represents the original number of sites. If it is a newly created site cluster, the initial cluster feature center vector allocated in step 135 is directly used as the updated cluster feature center vector, and the number of sites in the cluster is recorded as 1.

[0088] Step 137: Generate a site cluster affiliation code based on the unique identifier of the site cluster to which the communication site ultimately belongs. The length of the site cluster affiliation code is consistent with the number of bits of the affiliation code predefined in the site behavior clustering discrimination model.

[0089] Each site cluster is assigned a globally unique identifier upon creation, which is a sequence of integer values. The service message processing server converts this integer sequence into a binary bit stream. If the total length of the binary bit stream is less than the predefined number of bits for the site cluster identifier, it is padded with zeros at the front. If the total length exceeds the number of bits, the lower-order bits are truncated, ultimately forming a fixed-length site cluster identifier.

[0090] Step 140: Determine the scheduling priority order of cluster-level service messages corresponding to the communication site and the cluster-level multi-band resource occupancy coordination strategy based on the site cluster affiliation identifier code.

[0091] The determination of scheduling priority and multi-band resource allocation coordination strategy relies on a pre-set mapping table and real-time channel status indicators, respectively. The service message processing server, based on the attribution classification results from step 130, retrieves scheduling preference configuration information and resource allocation threshold configuration information related to the site cluster from the cluster configuration parameter library. It further reads the real-time multi-band channel idleness indicators of the current basic service set, combines the static configuration information with the dynamic idleness indicators to calculate the core parameter set required for this scheduling decision.

[0092] Step 141: Input the site cluster affiliation identifier into the preset cluster scheduling characteristic mapping table for lookup and matching, and obtain the cluster behavior preference category label corresponding to the site cluster affiliation identifier. The cluster behavior preference category label includes the parallel detection tendency dominant category label and the time-division frequency selection tendency dominant category label.

[0093] The cluster scheduling characteristic mapping table is a key-value mapping storage structure. The key is the binary encoded value of the site cluster affiliation identifier, and the value is the enumerated identifier of the cluster behavior preference category label. The service message processing server uses the site cluster affiliation identifier generated in step 137 as the lookup key, and locates the corresponding table entry in the cluster scheduling characteristic mapping table through hash lookup or binary search, and reads the cluster behavior preference category label stored in the value field of the table entry. The value range of the cluster behavior preference category label consists of two predefined labels: one is the parallel detection tendency dominant category label, indicating that the communication sites in the site cluster generally tend to maintain the parallel detection reception state; the other is the time-division frequency selection tendency dominant category label, indicating that the communication sites in the site cluster trigger time-division frequency selection dwell operations more frequently.

[0094] Step 142: Parse the pre-configured scheduling priority sorting rule chain list. The scheduling priority sorting rule chain list defines the relative order of different cluster behavior preference category labels when sharing service message channel resources.

[0095] The scheduling priority sorting rule linked list is an ordered linked list data structure. Each node in the linked list stores a cluster behavior preference category label and a priority sorting number. The business message processing server traverses the scheduling priority sorting rule linked list, compares the cluster behavior preference category label obtained in step 141 with the label values ​​of each node in the linked list, and reads the priority sorting number of the matching node.

[0096] Step 143: Determine the scheduling priority order of cluster-level service messages in the site cluster to which the communication site is located based on the position order of the cluster behavior preference category label in the scheduling priority sorting rule chain. The scheduling priority order of cluster-level service messages has a positive correspondence with the position order.

[0097] A higher priority sorting number indicates a higher queuing position for the corresponding site cluster when accessing shared service message channel resources. The service message processing server directly uses the read priority sorting number as the scheduling priority order of cluster-level service messages. The scheduling priority order will directly affect the construction rules of the subsequent service message sorting and distribution planning table. Service message frames to be sent by communication sites within the cluster with higher priority will be allocated a priority for transmission.

[0098] Step 144: Based on the cluster behavior preference category label, query the preset multi-band resource cooperation and occupancy mapping table to obtain the first bearer ratio threshold value of the primary frequency band signal service packet and the second bearer ratio threshold value of the secondary frequency band signal service packet corresponding to the cluster behavior preference category label.

[0099] The multi-band resource collaboration and occupancy mapping table is a two-dimensional lookup table. The row index is an enumerated value of the cluster behavior preference category label, and the column indexes are the first bearer ratio field and the second bearer ratio field, respectively. The query process is similar to the lookup and matching operation in step 141. The matching row element is located through key-value mapping, and then the first bearer ratio threshold value and the second bearer ratio threshold value are read from that row element. The first bearer ratio threshold value limits the upper limit ratio of the service packet bearer resources that the site cluster can occupy on the main frequency band signal channel relative to the total available resources of the main frequency band signal channel. The second bearer ratio threshold value limits the upper limit ratio of the service packet bearer resources that can be occupied on the secondary frequency band signal channel relative to the total available resources of the secondary frequency band signal channel.

[0100] Step 145: Determine the main frequency band resource quota data allocated to the site cluster to which the communication site belongs, based on the first bearer ratio threshold and the actual main frequency band channel idleness index of the basic service set to which the current communication site belongs.

[0101] The actual main frequency band channel idleness index is the ratio of the number of time slots on the main frequency band signal channel that were not occupied by any site within the most recent statistical time window to the total number of time slots. The service message processing server uses the mathematical product of the first bearer proportion threshold and the actual main frequency band channel idleness index as the ratio of main frequency band channel resources that can be occupied to be allocated to the site cluster. Then, the ratio of main frequency band channel resources that can be occupied is converted into the absolute number of total main frequency band channel time slot resources to obtain the main frequency band resource quota data. The main frequency band resource quota data includes at least two parameters: main frequency band resource occupation time quota and main frequency band transmission opportunity quota.

[0102] Step 146: Determine the secondary frequency band resource quota data allocated to the site cluster to which the communication site belongs, based on the second bearer ratio threshold and the actual secondary frequency band channel idleness index of the basic service set to which the current communication site belongs.

[0103] The processing procedure for this step is similar to that of step 145, except that the first bearer ratio threshold is replaced with the second bearer ratio threshold, and the actual primary band channel idleness index is replaced with the actual secondary band channel idleness index. The service message processing server calculates the product of the second bearer ratio threshold and the actual secondary band channel idleness index to obtain the secondary band channel available resource ratio, which is then converted into secondary band resource quota data. The secondary band resource quota data includes at least the secondary band resource occupancy time quota and the secondary band transmission opportunity quota.

[0104] Step 147: Integrate the primary frequency band resource quota data and the secondary frequency band resource quota data to generate a set of multi-band resource occupancy restriction parameters describing the site cluster where the communication site is located on the primary frequency band signal channel and the secondary frequency band signal channel.

[0105] The multi-band resource occupancy limit parameter set is encapsulated in a structured object format. This format contains two substructures: a primary band resource quota object and a secondary band resource quota object. The primary band resource quota object stores numerical pairs of primary band resource occupancy time quota and primary band transmission opportunity quota, respectively. The secondary band resource quota object stores numerical pairs of secondary band resource occupancy time quota and secondary band transmission opportunity quota, respectively.

[0106] Step 148: The multi-band resource occupancy limit parameter set and the scheduling priority order of cluster-level service packets are logically encapsulated to generate a cluster-level multi-band resource occupancy coordination strategy. The cluster-level multi-band resource occupancy coordination strategy is referenced in the subsequent generation of service packet distribution and control instructions for communication sites.

[0107] The logical encapsulation process embeds the multi-band resource occupancy limit parameter set and scheduling priority order as two independent data fields into a unified data container. The data container serializes the two data fields using a type-length-value encoding format and appends a policy version number and a policy effective time range field to the container header. The resulting binary encoded block is the cluster-level multi-band resource occupancy coordination policy. This occupancy coordination policy is referenced in the subsequent service message distribution and control instruction generation process, serving as the constraint basis for selecting distribution frequency bands, dividing transmission time windows, and allocating transmission opportunities.

[0108] Step 150: Based on the scheduling priority order of cluster-level service messages and the cluster-level multi-frequency band resource occupation coordination strategy, generate service message distribution control instructions for communication sites, and push the service message distribution control instructions to communication sites through main frequency band signals or auxiliary frequency band signals to adjust their service message sending and receiving windows.

[0109] After obtaining all scheduling policy parameters in step 140, the service message processing server retrieves all service message frames currently stored for transmission to the communication site from the service message buffer queue, selects frequency bands and arranges the transmission order according to the scheduling priority, and constrains the arrangement result according to the quota upper limit of each frequency band limited by the multi-frequency band resource occupancy limit parameters.

[0110] After the arrangement is completed, the sorting and distribution scheme, together with the resource occupation time quota and the transmission opportunity quota, are encapsulated into a service message distribution control instruction message. After real-time evaluation of the channel quality of the main frequency band signal channel and the auxiliary frequency band signal channel, the frequency band with better channel quality is selected as the bearer frequency band. The service message distribution control instruction is sent to the communication station through the media access control layer unicast management frame of the bearer frequency band.

[0111] Step 151: Input the site cluster affiliation identifier into the preset cluster scheduling characteristic mapping table for lookup and matching, and obtain the cluster behavior preference category label corresponding to the site cluster affiliation identifier.

[0112] The cluster scheduling feature mapping table is a key-value mapping storage structure. The key is the binary encoded value of the site cluster affiliation identifier, and the value is the enumerated identifier of the cluster behavior preference category label. The business message processing server uses the site cluster affiliation identifier generated in step 137 as the lookup key, locates the corresponding table entry in the cluster scheduling feature mapping table through hash lookup, and reads the cluster behavior preference category label stored in the value field of that table entry.

[0113] Step 152: Parse the pre-configured scheduling priority sorting rule linked list and determine the scheduling priority order of cluster-level service messages of the site cluster to which the communication site is located based on the cluster behavior preference category label.

[0114] The scheduling priority sorting rule linked list is an ordered linked list data structure. Each node in the linked list stores a cluster behavior preference category label and a priority sorting number. The service message processing server traverses the scheduling priority sorting rule linked list, compares the cluster behavior preference category label obtained in step 151 with the label values ​​of each node in the linked list, and reads the priority sorting number of the matching node. The higher the priority sorting number, the earlier the queuing position of the site cluster when sharing service message channel resources. This priority sorting number is used as the scheduling priority order of cluster-level service messages.

[0115] Step 153: Based on the cluster behavior preference category label, query the preset multi-band resource cooperation and occupancy mapping table to obtain the first bearer ratio threshold value of the primary frequency band signal service packet and the second bearer ratio threshold value of the secondary frequency band signal service packet corresponding to the cluster behavior preference category label.

[0116] The multi-band resource collaboration and occupancy mapping table is a two-dimensional lookup table. The row indexes are enumerated values ​​of the cluster behavior preference category labels, and the column indexes are the first bearer ratio field and the second bearer ratio field, respectively. The service message processing server locates the matching row element through key-value mapping and then reads the first bearer ratio threshold value and the second bearer ratio threshold value from that row element. The first bearer ratio threshold value limits the upper limit proportion of resources that the site cluster can occupy on the primary frequency band signal channel, and the second bearer ratio threshold value limits the upper limit proportion of resources that can be occupied on the secondary frequency band signal channel.

[0117] Step 154: Based on the first bearer ratio threshold and the actual main frequency band channel idleness index of the basic service set to which the current communication site belongs, determine the main frequency band resource occupation time quota and main frequency band transmission opportunity quota allocated to the site cluster to which the communication site is located.

[0118] The actual main frequency band channel idleness index is the ratio of the number of time slots on the main frequency band signal channel that were not occupied by any site in the most recent statistical time window to the total number of time slots. The service message processing server uses the product of the first bearer proportion threshold and the actual main frequency band channel idleness index as the main frequency band channel available resource ratio, and then converts the main frequency band channel available resource ratio into the absolute number of main frequency band channel time slot resources to determine the main frequency band resource occupancy time quota and the main frequency band transmission opportunity quota, respectively.

[0119] Step 155: Based on the second bearer ratio threshold and the actual secondary band channel idleness index of the basic service set to which the current communication site belongs, determine the secondary band resource occupancy time quota and secondary band transmission opportunity quota allocated to the site cluster to which the communication site is located.

[0120] The actual secondary band channel idleness index is the ratio of the number of time slots in the secondary band signal channel that were not occupied by any site within the most recent statistical time window to the total number of time slots. The service message processing server uses the product of the second bearer proportion threshold and the actual secondary band channel idleness index as the secondary band channel available resource ratio, and then converts the secondary band channel available resource ratio into the absolute number of total secondary band channel time slot resources to determine the secondary band resource occupancy time quota and the secondary band transmission opportunity quota, respectively.

[0121] Step 156: Based on the main frequency band resource occupation time quota, the main frequency band transmission opportunity quota, the secondary frequency band resource occupation time quota, and the secondary frequency band transmission opportunity quota, construct a service message distribution control instruction and select the carrying frequency band for transmission to adjust the service message transmission and reception window of the communication station.

[0122] After obtaining all resource quota parameters from steps 154 and 155, the service message processing server begins to construct the complete message payload of the service message distribution control instruction. After the instruction is constructed, it evaluates the channel quality of the primary frequency band signal channel and the secondary frequency band signal channel to select the carrying frequency band.

[0123] Step 1561: Based on the main frequency band resource occupancy time quota, the main frequency band transmission opportunity quota, the secondary frequency band resource occupancy time quota, and the secondary frequency band transmission opportunity quota, generate the basic parameter payload of the service message distribution control instruction for the communication site.

[0124] The overall structure of the service message distribution and control instruction consists of two connected parts: the basic parameter payload and the scheduling planning payload. The basic parameter payload is used to transmit resource occupancy limit parameters. The service message processing server encodes the primary frequency band resource occupancy time quota obtained in step 154 ​​into a fixed-width binary value and fills it into the primary frequency band time quota field of the basic parameter payload. It also encodes the primary frequency band transmission opportunity quota and fills it into the primary frequency band opportunity quota field. The secondary frequency band resource occupancy time quota and secondary frequency band transmission opportunity quota obtained in step 155 are encoded and filled into the secondary frequency band time field and secondary frequency band opportunity quota field, respectively. An additional instruction sequence number field is reserved in the basic parameter payload to identify the version of the current distribution and control instruction.

[0125] Step 1562: Read the length parameter and service quality level flag of each service message frame to be sent in the service message buffer queue to be sent to the communication station. Generate a service message sorting and distribution plan table for the main frequency band signal channel based on the main frequency band resource occupation time quota and the main frequency band transmission opportunity quota. Generate a service message sorting and distribution plan table for the secondary frequency band signal channel based on the secondary frequency band resource occupation time quota and the secondary frequency band transmission opportunity quota.

[0126] The service message processing server maintains two independent sorting queues for the buffer queue of service messages to be sent, corresponding to the primary frequency band signal channel and the secondary frequency band signal channel, respectively. For each service message frame to be sent, its frame length parameter and service quality level (SQL) flag are extracted. The frames are initially sorted according to the priority sorting rule from high to low based on the SQL flag. If the SQL flags are the same, they are then sorted again according to the order of the frame length parameter from small to large.

[0127] After sorting, the service message processing server sequentially adds the frame length parameter of each service message frame to be sent to the allocated primary frequency band time quota variable or the allocated secondary frequency band time quota variable. When the accumulated time quota variable exceeds the corresponding resource occupancy time quota, the frame and subsequent frames are removed from the current frequency band's sorting and distribution plan table. The frames that are not removed and their sorting order constitute the service message sorting and distribution plan table for that frequency band. The primary frequency band transmission opportunity quota and the secondary frequency band transmission opportunity quota serve as the maximum allowed number of frames in the primary frequency band sorting and distribution plan table and the secondary frequency band sorting and distribution plan table, respectively.

[0128] Step 1563: After completing the duplicate transmission avoidance processing of the service messages to be transmitted in the main frequency band signal channel and the auxiliary frequency band signal channel, the service message sorting and distribution planning table is assembled into the scheduling planning payload part of the distribution control instruction.

[0129] The duplicate transmission avoidance is achieved by comparing the frame sequence number field of the same service message frame to be transmitted in the primary frequency band sorting and distribution planning table and the secondary frequency band sorting and distribution planning table. If a frame sequence number appears in the sorting and distribution planning tables of both frequency bands, the sorting and distribution entry of the frequency band in which it is located in the scheduling priority order is retained and it is deleted from the sorting and distribution planning table of the other frequency band.

[0130] After deduplication, the two sorting and distribution planning tables are structured and encoded according to the frame sequence number, the offset of the proposed transmission time window, and the identifier of the proposed frequency band. They are then merged and arranged into the scheduling planning payload. The header of the scheduling planning payload records the number of entries in the main frequency band planning table and the number of entries in the secondary frequency band planning table, followed by the encoded records of each planning table entry in sequence.

[0131] Step 1564: Detect the channel quality indicator evaluation values ​​of the current primary frequency band signal channel and the secondary frequency band signal channel, and select the signal frequency band whose channel quality indicator evaluation value meets the preset reliability judgment conditions as the instruction bearer frequency band. Through the selected signal frequency band, encapsulate the service message distribution and control instruction containing the basic parameter payload and scheduling planning payload into a unicast management frame and send it to the media access control layer of the communication station, so as to control the communication station to adjust its service message transmission and reception window according to the pre-planned time window index and frequency band allocation instruction in the service message distribution and control instruction.

[0132] The acquisition of channel quality indication (CQI) assessment values ​​is completed by the physical layer within each channel state measurement cycle, and the measurement results are reported to the channel state information cache table of the media access control layer. The service message processing server reads the CQI assessment values ​​of the primary frequency band signal channel and the secondary frequency band signal channel in the most recent measurement cycle, and compares the two CQI assessment values ​​with the minimum channel quality threshold specified by the preset reliability evaluation conditions.

[0133] If the channel quality indication evaluation values ​​of both signal frequency bands meet the minimum channel quality threshold, the frequency band with the higher channel quality indication evaluation value is selected as the command carrying frequency band; if only one signal frequency band meets the threshold, the frequency band that meets the threshold is selected; if neither meets the threshold, the judgment operation of this step is re-executed after waiting for the next measurement cycle.

[0134] After selecting the command bearer frequency band, the basic parameter payload and scheduling planning payload are filled into the frame body field according to the frame format specification of the media access control layer unicast management frame. The destination address field of the unicast management frame is filled with the media access control layer address of the communication station. Then, it is delivered to the physical layer and completed radio frequency transmission via the selected command bearer frequency band.

[0135] After successfully receiving and demodulating the unicast management frame, the media access control layer of the communication site configures its own media access control layer's transmit / receive window parameters according to the time window index specified in the basic parameter payload and the frequency band allocation indication specified in the scheduling planning payload. The actual transmission and reception of subsequent service messages are all constrained by these transmit / receive window parameters.

[0136] As another embodiment, after pushing the service message distribution control command to the communication station via the main frequency band signal or the auxiliary frequency band signal, the method further includes: Step 210: Obtain the actual dwell behavior record sequence generated by the communication station when executing the service message distribution control command in multiple consecutive service message transmission and reception windows. The actual dwell behavior record sequence contains several dwell status markers arranged in time sequence and the corresponding signal frequency band occupancy duration information.

[0137] After the service message distribution control command takes effect on the communication site side, the communication site will strictly follow the time window index planning and frequency band allocation instructions specified in the command to execute each round of transmission and reception operations. During this period, the protocol stack inside the communication site will record the frequency band occupancy behavior log in real time within each time window. The log entries include information such as the window start time, window end time, selected camping frequency band number, and the actual duration of frequency band occupancy within the window.

[0138] Every feedback acquisition cycle, the service message processing server requests the communication station to upload all frequency band occupancy behavior logs since the end of the previous acquisition cycle. The communication station packages the log entries into log report frames in chronological order of the window start time and sends them back to the service message processing server via the control channel. The service message processing server extracts the raw log records from the log report frames, sorts them in ascending order according to the window start time, extracts the stationed frequency band number of each log entry to form a stationed status marker sequence, and extracts the occupancy duration within each group of windows to form signal frequency band occupancy duration information. Together, these two constitute the actual stationed behavior record sequence output in this step.

[0139] Step 220: Based on the site cluster affiliation identifier code, retrieve the standard dwell behavior reference sequence corresponding to the site cluster to which the communication site belongs from the preset cluster behavior template library; perform frequency band dwell behavior offset comparison processing on the standard dwell behavior reference sequence in a unified time dimension to obtain a frequency band behavior offset descriptor containing the main frequency band dwell time offset sequence and the auxiliary frequency band dwell time offset sequence.

[0140] The cluster behavior template library stores the standard dwell behavior baseline sequence corresponding to each site cluster affiliation identifier code. The standard dwell behavior baseline sequence is defined by performing time window-by-time average duration statistics on the actual dwell behavior record sequence reported by all communication sites belonging to the same site cluster within the same scheduling performance observation window, forming a combination value pair of the standard dwell duration of the main frequency band and the standard dwell duration of the auxiliary frequency band under each time window.

[0141] The service message processing server reads the corresponding standard dwell time baseline sequence from the cluster behavior template library based on the site cluster affiliation identifier code of the current communication site, and aligns its time window index with the time window index of the actual dwell time record sequence one by one. For each aligned time window, the difference between the main frequency band occupation time in the actual dwell time record sequence and the standard dwell time of the main frequency band in the standard dwell time baseline sequence is used as the main frequency band dwell time offset, and the difference between the secondary frequency band occupation time in the actual dwell time record sequence and the standard dwell time of the secondary frequency band in the standard dwell time baseline sequence is used as the secondary frequency band dwell time offset. The main frequency band dwell time offsets of multiple consecutive time windows are concatenated to obtain the main frequency band dwell time offset sequence, and the same concatenation is used to obtain the secondary frequency band dwell time offset sequence. The two sets of sequences constitute the frequency band behavior offset descriptor of this step.

[0142] Step 230: Perform abnormal deviation pattern recognition on the frequency band behavior offset descriptor to generate a deviation trend feature code that characterizes the degree and direction of deviation of the communication station's dwell behavior in multiple consecutive service message transmission and reception windows.

[0143] The implementation method for anomaly deviation pattern recognition involves calculating the root mean square statistic and mean sign direction for both the main frequency band dwell time offset sequence and the secondary frequency band dwell time offset sequence. The square root of the squared value of each element in the main frequency band dwell time offset sequence is obtained by taking the mean of the squared values. A positive arithmetic mean of the main frequency band dwell time offset sequence indicates a positive overdosing direction, while a negative mean indicates a negative time-lapse direction. The calculation method for the secondary frequency band offset is the same.

[0144] The calculated root mean square value of the main frequency band offset is mapped to a deviation level code through threshold comparison. The deviation direction of the main frequency band is converted into a direction code. The two are concatenated to generate the main frequency band deviation trend component. The same operation is performed on the secondary frequency band to generate the secondary frequency band deviation trend component. The two components are concatenated again to obtain the deviation trend feature code.

[0145] Step 240: Correct the basic parameter payload of the service message distribution control instruction according to the deviation trend feature code, generate the service message distribution control correction instruction, and push the service message distribution control correction instruction to the communication station through the main frequency band signal or the auxiliary frequency band signal to calibrate its subsequent service message transmission and reception window.

[0146] The combination of the main frequency band deviation level code and the main frequency band deviation direction code, which are parsed from the deviation trend feature code, reflects the severity and direction of the resource occupation deviation of the communication site on the main frequency band.

[0147] The specific logic for the correction process is as follows: If the main frequency band deviation direction code indicates a positive over-occupancy direction and the main frequency band deviation level code exceeds the preset correction trigger level, then the main frequency band resource occupancy time quota in the basic parameter payload is multiplied and reduced by a preset quota reduction factor, and the main frequency band transmission opportunity quota is reduced by the same factor; if the main frequency band deviation direction code indicates a negative time reduction direction and the deviation level code exceeds the correction trigger level, then it is multiplied and increased by a preset quota expansion factor. The correction process for the secondary frequency band basic parameter payload follows the same logic.

[0148] The corrected basic parameter payload is reassembled with the original scheduling planning payload into a service message distribution control correction instruction. The transmission frequency band selection of the service message distribution control correction instruction is the same as in step 1564. It is pushed to the communication station through the same unicast management frame encapsulation to cover the previous service message distribution control instruction and readjust its subsequent service message transmission and reception window.

[0149] As another embodiment, after pushing the service message distribution control command to the communication station via the main frequency band signal or the auxiliary frequency band signal, the method further includes: Step 310: Within the preset scheduling performance observation window, collect the scheduling performance feedback record sequence formed by all communication stations on the main frequency band signal channel and the auxiliary frequency band signal channel in response to their respective service message distribution control commands. The scheduling performance feedback record sequence includes the service message distribution completion rate, channel access delay and message retransmission frequency corresponding to each signal frequency band channel.

[0150] After the preset scheduling performance observation window ends, the service message processing server reads the accumulated scheduling performance feedback records from the protocol stack statistical counters of each communication station in the network in batches, and summarizes them into a scheduling performance feedback record sequence. Each communication station maintains three statistical counters for each signal frequency band channel: a service message distribution completion counter, a channel access total delay accumulation counter, and a message retransmission count counter.

[0151] The service message processing server reads the instantaneous values ​​of each counter at the start and end of the scheduling performance observation window, calculates the difference between the two instantaneous values ​​to obtain the cumulative value within the window, divides the cumulative value of the number of completed service message distributions by the cumulative value of the number of service message distribution attempts to obtain the service message distribution completion rate, divides the cumulative amount of total channel access delay by the number of service message distribution attempts to obtain the average channel access delay, and directly records the cumulative value of message retransmission count as the message retransmission frequency field.

[0152] Step 320: Based on the site cluster affiliation identifier of each communication site, the scheduling performance feedback record sequence is processed by sub-cluster aggregation to generate cluster-level scheduling performance feedback aggregation information corresponding to each site cluster.

[0153] The cluster-based aggregation processing uses the site cluster affiliation identifier as the grouping key, traversing the scheduling performance feedback records of all communication sites, and grouping the scheduling performance feedback records of communication sites with the same site cluster affiliation identifier into the same group. For each group, the arithmetic mean of the service message distribution completion rate of all communication sites in the group is calculated as the cluster-level service message distribution completion rate; the arithmetic mean of the average channel access delay of all communication sites in the group is calculated as the cluster-level average channel access delay; and the arithmetic mean of the message retransmission frequency of all communication sites in the group is calculated as the cluster-level average message retransmission frequency. These three cluster-level indicators constitute the cluster-level scheduling performance feedback aggregation information of that site cluster.

[0154] Step 330: Couple and parse the cluster-level scheduling performance feedback aggregation information with the cluster behavior preference category label of the site cluster to generate frequency band load pressure distribution characteristics. The frequency band load pressure distribution characteristics characterize the differences in the response of site clusters with different behavior preference categories to scheduling resources on the main frequency band signal channel and the auxiliary frequency band signal channel.

[0155] The implementation of coupled analysis involves constructing a two-sample difference analysis process with cluster behavior preference category labels as hierarchical variables. The cluster-level scheduling performance feedback aggregation information matrix of site clusters with dominant parallel detection tendency labels on the main frequency band signal channel is used as the first sample group. The aggregation information matrix of site clusters with dominant time-division frequency selection tendency labels on the main frequency band signal channel is used as the second sample group. The difference between the mean of cluster-level service message distribution completion rate in the first sample group and the corresponding mean in the second sample group is calculated, and the significance of this difference is tested.

[0156] The same processing method is used to perform difference analysis on the aggregated information matrix of the secondary frequency band signal channel. The difference direction and magnitude values ​​between the main frequency band and the secondary frequency band are integrated into a multi-dimensional vector, which is the frequency band load pressure distribution characteristic.

[0157] Step 340: Based on the frequency band load pressure distribution characteristics, the policy parameters of the primary frequency band resource occupation time quota and the secondary frequency band resource occupation time quota in the cluster-level multi-frequency band resource occupation coordination strategy are reconfigured to generate a multi-frequency band resource occupation coordination strategy update instruction. The multi-frequency band resource occupation coordination strategy update instruction and subsequent service message distribution control instructions are encapsulated and transferred to update the resource scheduling strategy of the entire site cluster.

[0158] The strategy parameter reconfiguration process triggers quota adjustments based on the magnitude of differences in various dimensions of the frequency band load pressure distribution characteristics. If the magnitude of the difference in the frequency band load pressure distribution characteristics in the primary frequency band dimension exceeds a preset reconfiguration trigger threshold, then for the site clusters belonging to the behavioral preference category tags that are at a disadvantage in the primary frequency band, the proportion of primary frequency band resource occupation time quota in their cluster-level multi-frequency band resource occupation coordination strategy is increased, with the increase being directly proportional to the magnitude of the difference. Simultaneously, the corresponding quota proportion for the site clusters belonging to the opposite behavioral preference category tags is adjusted to maintain the total primary frequency band quota of the entire network unchanged. The adjustment method for secondary frequency band resource occupation time quota is similar. The adjusted primary frequency band resource occupancy time quota and secondary frequency band resource occupancy time quota are rewritten into the cluster-level multi-frequency band resource occupancy coordination strategy, forming a multi-frequency band resource occupancy coordination strategy update instruction. This update instruction is broadcast to all communication sites in the network via the beacon frame of the primary frequency band channel, and is also referenced by the subsequent service message distribution and control instruction generation process of each communication site, thus completing the closed-loop update of the cluster resource scheduling strategy for all sites.

[0159] In the above embodiments, before the site behavior clustering discrimination model is put into application, the main and auxiliary frequency mixed signal streams reported by each communication site in the network in multiple consecutive historical monitoring periods are collected, and the receiving behavior logic parsing processing described above is performed on the mixed signal frame unit sequence in each monitoring period to obtain the sample receiving behavior logic descriptor of each site, thereby constructing a training sample set.

[0160] The behavioral category label for each training sample is determined by a combination of semi-automatic annotation and manual review. In the semi-automatic annotation stage, the sample is pre-labeled as either the parallel detection tendency-dominant class or the time-division frequency selection tendency-dominant class based on the comparison between the direction code of the behavioral tendency identifier field and the threshold of the positive transition probability value in the behavioral tendency probability distribution descriptor. For samples whose direction code and transition probability fall within a preset dilemma range, the manual review process is initiated, and the final category is determined by the human based on the frequency band dwell time distribution map of the site in multiple monitoring cycles.

[0161] Furthermore, in the feature encoding layer of the site behavior clustering discrimination model, the number of embedding entries in the behavior trend encoding module is set to the total number of possible values ​​for the direction code. The embedding dimension hyperparameter typically ranges from 16 to 64. The output dimension of the fully connected layer is consistent with the output dimension of the fully connected layer of the resident feature encoding module to ensure dimensionality symmetry between the behavior trend feature vector and the resident feature vector when performing cross-dimensional connections in the feature aggregation layer. After the concatenation operation, the feature aggregation layer is connected to a batch normalization layer. The momentum parameter of the batch normalization layer is selected between 0.1 and 0.3, a value determined through multiple rounds of cross-validation to be a balance between convergence speed and generalization performance.

[0162] In the cluster center distance calculation layer, the cluster feature center vector of the candidate site cluster is pre-calculated from the labeled training sample set using the K-means clustering algorithm during the initial deployment of the model. The initial number of clusters is determined by the silhouette coefficient evaluation criterion, and the aggregation radius threshold associated with this number of clusters is also taken from the upper bound of the average intra-cluster distance corresponding to the maximum value of the silhouette coefficient.

[0163] The optimization objective during model training is to minimize the Euclidean distance loss between the joint representation feature tensor and the feature center vector of its respective cluster. An adaptive moment estimation optimizer is used, with an initial learning rate ranging from 0.001 to 0.0005 and a weight decay coefficient set to 1 × 10⁻⁶. -4The training iterations continue until the sum of squared distances within the cluster converges to a stable interval. After training, the cluster feature center vector table in the cluster center distance calculation layer is continuously updated with the incremental learning mechanism. Specifically, whenever a new communication station is confirmed to be included in one of the station clusters, the joint representation feature tensor of the new station and the original cluster feature center vector are adjusted online using a sample-weighted moving average to maintain the continuous representativeness of the cluster feature centers on the behavioral preferences within the cluster.

[0164] This application analyzes the parallel detection state transition sequence and time-division frequency selective dwell decision sequence of communication sites from the primary and secondary frequency mixed signal stream to generate a receiving behavior logic descriptor to finely characterize the physical layer processing behavior preferences of the sites. It then uses a site behavior clustering discrimination model to group sites with similar receiving behavior logic patterns into the same site cluster, uniformly outputting a site cluster affiliation identifier at the cluster level. This transforms single-site behavior deviation noise into cluster statistical characteristics for scheduling decisions. Based on the differentiated site cluster affiliation identifier, the scheduling priority order of cluster-level service packets and the cluster-level multi-band resource allocation coordination strategy are determined. This ensures that the service packet distribution control instructions not only match the differentiated tendencies of different clusters in multi-band parallel detection and time-division frequency selective dwell, but also dynamically balance the allocation ratio of primary and secondary frequency band channel resources among the clusters. While ensuring the fairness of network-wide scheduling, this improves the service packet distribution efficiency and spectrum resource utilization balance in multi-band fusion coverage scenarios.

[0165] In practical implementation, those skilled in the art can combine general deep learning model building techniques to construct the behavior tendency encoding module and the resident feature encoding module into a standard structure of an embedding layer plus a fully connected layer. The number of entries in the embedding layer is directly taken as the number of discrete values ​​of the direction code or amplitude code. The embedding dimension is set between 16 and 64 according to the model capacity requirements. The output of the fully connected layer uses linear transformation plus activation function rectification. The feature aggregation layer uses vector concatenation followed by batch normalization layer. The momentum parameter of batch normalization is between 0.1 and 0.3. All of the above operations can be directly implemented through the built-in layers and default parameters in TensorFlow or PyTorch. The initial cluster centers of the cluster center distance calculation layer can be pre-calculated by the K-means clustering algorithm on the joint representation feature tensor of the labeled samples. The number of clusters and the aggregation radius threshold are automatically optimized on the validation set by the silhouette coefficient. The model training uses an adaptive moment estimation optimizer with an initial learning rate in the range of 0.001 to 0.0005 and a weight decay coefficient set to 1×10. -4 Training continues until the sum of squared distances within the cluster converges.

[0166] For channel number differential comparison and state transition event statistics of mixed signal frame units, those skilled in the art can use a finite state machine to compare the channel number change patterns of adjacent frames bit by bit, and map the changes in the differential values ​​of the main frequency band and the auxiliary frequency band to a parallel detection state or a time-division frequency-selective dwell state through a state transition table. The smoothing of the persistent dwell tendency duration sequence and the parallel detection state duration sequence can be directly achieved using an exponentially weighted moving average algorithm, allocating attenuation weight coefficients according to time proximity. The product of the bearer ratio threshold and the channel idleness index is used to calculate the absolute amount of resource quota for each frequency band. The channel idleness index is obtained from time slot occupancy statistics. The ratio of the number of time slots to the total number of time slots is calculated following the conventional resource utilization quantification method in communication systems. The hierarchical mapping of dwell stickiness indicators relies on preset hierarchical threshold boundaries to discretize and classify the relative proportion values. The normalization of cosine similarity or Euclidean distance metrics can be achieved by normalizing the vector magnitude scale using the L2 norm before calculation.

[0167] Based on the same inventive concept, embodiments of this application also provide a service message processing server. See also... Figure 2 As shown, this is a schematic diagram of the structure of a possible service message processing server provided in an embodiment of this application. Figure 2 In this system, the service message processing server 200 includes a processor 210 and a memory 220. The processor 210 and the memory 220 are interconnected via a communication bus. The memory 220 stores computer programs that can be executed by the processor 210. By executing the instructions stored in the memory 220, the processor 210 can perform the steps of the aforementioned service message processing method based on a multi-band network.

[0168] Based on the same inventive concept, embodiments of this application provide a computer-readable storage medium including a computer program. When the computer program runs on a service message processing server, it causes the service message processing server to execute the steps of the aforementioned service message processing method based on a multi-band network. In some possible implementations, various aspects of the service message processing method based on a multi-band network provided in this application can also be implemented as a program product including a computer program. When the program product runs on a service message processing server, the computer program causes the service message processing server to execute the steps of the aforementioned service message processing method based on a multi-band network. For example, the service message processing server can execute... Figure 1The steps are shown in the diagram. The computer-readable storage medium includes volatile or non-volatile or a combination thereof, and may be removable or non-removable. Examples of computer-readable storage media include, but are not limited to, phase-change random access memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random-access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), flash memory or other memory technologies, CD-ROM, Digital Video Disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium.

[0169] like Figure 3 As shown, this is a functional block diagram of the service message processing server provided in this embodiment of the application. The service message processing server includes a service message processing device, which includes: The mixed signal acquisition module is used to acquire the main and auxiliary frequency mixed signal stream formed by the main frequency band signal and the auxiliary frequency band signal received by the communication station within a preset monitoring period. The main and auxiliary frequency mixed signal stream includes multiple continuously captured mixed signal frame units, and each mixed signal frame unit carries a beacon frame identifier and a set of physical layer capability declaration fields. The behavior logic parsing module is used to perform receiving behavior logic parsing processing on the main and auxiliary frequency mixed signal stream, extract the parallel detection state transition sequence and time-division frequency selection dwell decision sequence when the communication station responds to the main frequency band signal and the auxiliary frequency band signal, and generate a receiving behavior logic descriptor that reflects the physical layer processing behavior preference of the corresponding communication station. The behavior trajectory classification module is used to classify the receiving behavior logical descriptor by a preset site behavior clustering discrimination model, divide the communication site into a site cluster whose corresponding receiving behavior similarity meets the preset aggregation criterion, and generate a site cluster affiliation identifier code. The cluster indicator determination module is used to determine the scheduling priority order of cluster-level service messages corresponding to the communication site and the cluster-level multi-frequency band resource occupation coordination strategy based on the site cluster affiliation identifier code. The control instruction generation module is used to generate a control instruction for the distribution of service packets to the communication station based on the scheduling priority order of the cluster-level service packets and the occupancy coordination strategy of the cluster-level multi-frequency band resources, and push the control instruction for the distribution of service packets to the communication station through the main frequency band signal or the auxiliary frequency band signal to adjust its service packet transmission and reception window.

[0170] Accordingly, this application also provides a computer program product, which includes a computer program or instructions that, when executed by a processor, cause the processor to implement the steps in the above method embodiments. It should be understood that each step or combination of steps in the above method flow can be implemented by the computer program or instructions. Furthermore, these computer programs or instructions can be applied to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device, enabling the processor of the general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to function as an apparatus for implementing the corresponding functions in the above method embodiments.

[0171] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0172] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0173] Finally, it should be noted that the above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for processing service packets based on a multi-band network, characterized in that, The method includes: The system acquires a mixed signal stream of primary and secondary frequency bands formed by the primary frequency band signals and secondary frequency band signals received by the communication station within a preset monitoring period. The mixed signal stream of primary and secondary frequency bands includes multiple continuously captured mixed signal frame units, and each mixed signal frame unit carries a beacon frame identifier and a set of physical layer capability declaration fields. The receiving behavior logic parsing process is performed on the main and auxiliary frequency mixed signal stream to extract the parallel detection state transition sequence and time-division frequency selection dwell decision sequence when the communication station responds to the main frequency band signal and the auxiliary frequency band signal, and to generate a receiving behavior logic descriptor that reflects the physical layer processing behavior preference of the corresponding communication station. The receiving behavior logical descriptor is classified into behavior trajectories by a preset site behavior clustering discrimination model, and the communication site is divided into a site cluster whose corresponding receiving behavior similarity meets the preset aggregation criterion, and a site cluster affiliation identifier code is generated. The scheduling priority order of cluster-level service messages corresponding to the communication site and the cluster-level multi-band resource occupation coordination strategy are determined based on the site cluster affiliation identifier code. Based on the scheduling priority order of the cluster-level service packets and the cluster-level multi-band resource occupancy coordination strategy, a service packet distribution control instruction is generated for the communication station, and the service packet distribution control instruction is pushed to the communication station through the main frequency band signal or the auxiliary frequency band signal to adjust its service packet transmission and reception window.

2. The method according to claim 1, characterized in that, The receiving behavior logic parsing process of the mixed primary and secondary frequency signal stream extracts the parallel detection state transition sequence and time-division frequency selection dwell decision sequence when the communication station responds to the primary frequency band signal and the secondary frequency band signal, and generates a receiving behavior logic descriptor reflecting the physical layer processing behavior preference of the communication station, including: The beacon frame identifier carried by each hybrid signal frame unit in the primary and secondary frequency hybrid signal stream is parsed to determine the basic service set identifier to which the hybrid signal frame unit belongs and the corresponding primary frequency band channel number and secondary frequency band channel number. Extract the multi-band parallel reception capability indicator bit, the band switching delay tolerance level indicator bit, and the channel dwell time preference indicator bit from the physical layer capability declaration field set carried by the mixed signal frame unit, and determine whether the communication station supports the synchronous demodulation operation of the main frequency band signal and the auxiliary frequency band signal based on the multi-band parallel reception capability indicator bit. The state transition tolerance time parameter required for the communication station to transition from parallel detection state to time-division frequency selective dwell state when receiving concurrent main frequency band signals and auxiliary frequency band signals is obtained according to the frequency band switching delay tolerance level indicator bit. The duration of continuous reception of the main frequency band signal channel by the communication station in the time-division frequency-selective dwell state is determined based on the channel dwell time preference indicator bit. By combining the state marking results of the continuous logical moments of the main and auxiliary frequency mixed signal stream, and the correlation results of the continuous dwell tendency duration parameter and the parallel detection state duration parameter, a receiving behavior logical descriptor reflecting the physical layer processing behavior preference of the communication station is generated.

3. The method according to claim 2, characterized in that, The process combines the state marking results of the continuous logical moments of the main and auxiliary frequency hybrid signal streams, and the correlation results of the continuous dwell tendency duration parameter and the parallel detection state duration parameter, to generate a receiving behavior logical descriptor reflecting the physical layer processing behavior preference of the communication station, including: The main and auxiliary frequency hybrid signal streams are sequentially processed by logical state marking in chronological order. If the beacon frame identifier of an adjacent hybrid signal frame unit changes the main frequency band channel number while the auxiliary frequency band channel number remains unchanged, the current logical state is marked as parallel detection state and the corresponding main frequency band channel feature code change information is recorded. If the beacon frame identifier of an adjacent hybrid signal frame unit changes both the main frequency band channel number and the auxiliary frequency band channel number, the current logical state is marked as time-division frequency selective camping state and the corresponding dual-frequency band channel linkage feature code change information is recorded. Integrate the state marking results of continuous logical moments to generate a state transition sequence that includes parameters such as the duration of parallel detection states, the triggering time of time-division frequency-selective dwell states, and the transition interval parameters between adjacent states. The correlation logic between the frequency of the parallel detection state transition to the time-division frequency selective dwell state in the state transition sequence and the state transition tolerance time parameter is statistically analyzed to generate a probability distribution descriptor for the behavior tendency of the communication station to maintain the parallel detection state or to trigger the time-division frequency selective dwell state under different channel load conditions. The relative proportion between the duration of the persistent dwell tendency parameter and the duration of the parallel detection state parameter is correlated, and a receiving behavior logical descriptor reflecting the physical layer processing behavior preference of the corresponding communication station is generated by combining the behavior tendency probability distribution descriptor. The receiving behavior logical descriptor includes a behavior tendency identifier field and a dwell feature description field.

4. The method according to claim 3, characterized in that, The association of the relative proportion between the persistent dwell tendency duration parameter and the parallel detection state duration parameter, and the generation of a receiving behavior logical descriptor reflecting the physical layer processing behavior preference of the communication station by combining the behavior tendency probability distribution descriptor, includes: Obtain the duration sequence of the continuous dwell tendency recorded by the communication station in multiple consecutive state transition sequences and the duration sequence of the parallel detection state; The duration sequence of the persistent dwell tendency and the duration sequence of the parallel detection state are respectively subjected to time-series smoothing processing to generate a stable characterization value of the dwell tendency duration and a benchmark reference characterization value of the duration of the parallel detection state. Calculate the relative proportion between the dwell tendency duration stability characterization value and the benchmark reference characterization value, and input the relative proportion value into a preset dwell preference classification logic unit to obtain a dwell stickiness identifier that describes the degree of adhesion of the communication station to the main frequency band signal channel. Analyze the change trajectory of the behavior trend identifier field in the behavior trend probability distribution descriptor, and extract the sensitivity change curve of the transition trigger logic of the communication station from parallel detection state to time-division frequency selective dwell state during the continuous monitoring period; The transition trigger logic sensitivity change curve is processed by segmented feature extraction to obtain the segmented sensitivity difference distribution information under different combinations of main frequency band channel feature codes and auxiliary frequency band channel feature codes. Combining the dwell stickiness identifier with the segmented sensitivity difference distribution information, a behavior preference mapping vector is constructed using a predefined behavior preference classification mapping table, with the dwell tendency dimension and the state transition sensitivity dimension as coordinate axes; The behavior preference mapping vector is projected along the dwell tendency dimension using the principal component direction, and the feature projection component with the maximum behavior discernibility in the projection direction is retained. The feature projection component and the behavior preference mapping vector are fused along the orthogonal components of the state transition sensitivity dimension to generate a receiving behavior logical descriptor that represents the physical layer processing behavior preference of the communication station. The behavior tendency identifier field in the receiving behavior logical descriptor stores the direction code of the feature projection component, and the dwell feature description field in the receiving behavior logical descriptor stores the amplitude code of the orthogonal component.

5. The method according to any one of claims 1-4, characterized in that, The step involves classifying the received behavior logical descriptors using a preset site behavior clustering discrimination model, assigning the communication sites to site clusters whose corresponding received behavior similarity satisfies a preset aggregation criterion, and generating a site cluster affiliation identifier code, including: The behavior trend identifier field and the dwell feature description field of the received behavior logical descriptor are input into the feature encoding layer of the site behavior clustering discrimination model. The behavior trend identifier field is mapped to a behavior trend feature vector and the dwell feature description field is mapped to a dwell feature vector through the feature encoding layer. The feature aggregation layer of the site behavior clustering discrimination model performs cross-dimensional connection processing on the behavior tendency feature vector and the dwelling feature vector to generate a joint representation feature tensor describing the communication site in the behavior tendency dimension and the dwelling tendency dimension. The cluster center distance calculation layer of the site behavior clustering discrimination model is invoked to perform multi-dimensional spatial distance measurement calculation on the joint representation feature tensor and the cluster feature center vectors of multiple pre-established candidate site clusters, generating a set of relative distance parameters between the communication site and each candidate site cluster; wherein, each relative distance parameter in the set of relative distance parameters corresponds to a candidate site cluster and each relative distance parameter represents the degree of logical difference between the receiving behavior logic mode of the communication site and the typical receiving behavior logic mode of the corresponding candidate site cluster; Retrieve the smallest relative distance parameter in the set of relative distance parameters and compare the smallest relative distance parameter with the aggregation radius threshold specified by the preset aggregation criterion; If the minimum relative distance parameter is less than the aggregation radius threshold, the communication station is assigned to the candidate station cluster corresponding to the minimum relative distance parameter. If the minimum relative distance parameter is greater than or equal to the aggregation radius threshold, a new station cluster is created with the joint representation feature tensor as the initial cluster feature center vector. Update the cluster feature center vector of the divided site cluster or newly added site cluster. Based on the newly added joint representation feature tensor, perform weighted shift adjustment on the original cluster feature center vector to generate the updated cluster feature center vector for subsequent classification of the behavior trajectory of communication sites. A site cluster affiliation identifier is generated based on the unique identifier of the site cluster to which the communication site ultimately belongs. The length of the site cluster affiliation identifier is consistent with the number of bits of the affiliation identifier code predefined in the site behavior clustering discrimination model.

6. The method according to claim 1, characterized in that, The step of determining the scheduling priority order of cluster-level service packets corresponding to the communication site and the cluster-level multi-band resource allocation coordination strategy based on the site cluster affiliation identifier includes: The site cluster affiliation identifier is input into a preset cluster scheduling characteristic mapping table for lookup and matching to obtain the cluster behavior preference category label corresponding to the site cluster affiliation identifier. The cluster behavior preference category label includes a parallel detection tendency dominant label and a time-division frequency selection tendency dominant label. Parse the pre-configured scheduling priority sorting rule chain, which defines the relative order of different cluster behavior preference category labels when sharing service message channel resources; Based on the position of the cluster behavior preference category label in the scheduling priority sorting rule chain, the scheduling priority order of the cluster-level service packets of the site cluster to which the communication site is located is determined, and the scheduling priority order of the cluster-level service packets has a positive correspondence with the position order. Based on the cluster behavior preference category label, query the preset multi-band resource cooperation and occupancy mapping table to obtain the first bearer ratio threshold value of the primary frequency band signal service packet and the second bearer ratio threshold value of the secondary frequency band signal service packet corresponding to the cluster behavior preference category label; Based on the first bearer ratio threshold and the actual main frequency band channel idleness index of the basic service set to which the current communication site belongs, the main frequency band resource quota data allocated to the site cluster to which the communication site is located is determined; Based on the second bearer ratio threshold and the actual secondary band channel idleness index of the basic service set to which the current communication site belongs, the secondary band resource quota data allocated to the site cluster to which the communication site is located is determined; Integrate the primary frequency band resource quota data and the secondary frequency band resource quota data to generate a set of multi-frequency band resource occupancy restriction parameters describing the site cluster where the communication site is located on the primary frequency band signal channel and the secondary frequency band signal channel; The set of multi-band resource occupancy limit parameters and the scheduling priority order of the cluster-level service packets are logically encapsulated to generate a cluster-level multi-band resource occupancy coordination strategy. This cluster-level multi-band resource occupancy coordination strategy is referenced when generating subsequent service packet distribution and control instructions for the communication site.

7. The method according to claim 1, characterized in that, Based on the scheduling priority order of the cluster-level service packets and the cluster-level multi-band resource allocation coordination strategy, a service packet distribution control instruction is generated for the communication station. This instruction is then pushed to the communication station via the primary frequency band signal or the secondary frequency band signal to adjust its service packet transmission and reception window, including: The site cluster affiliation identifier is entered into a preset cluster scheduling feature mapping table for a lookup and matching operation to obtain the cluster behavior preference category label corresponding to the site cluster affiliation identifier. Parse the pre-configured scheduling priority sorting rule linked list, and determine the scheduling priority order of cluster-level service packets of the site cluster to which the communication site is located based on the cluster behavior preference category label. Based on the cluster behavior preference category label, query the preset multi-band resource cooperation and occupancy mapping table to obtain the first bearer ratio threshold value of the primary frequency band signal service packet and the second bearer ratio threshold value of the secondary frequency band signal service packet corresponding to the cluster behavior preference category label; Based on the first bearer ratio threshold and the actual main frequency band channel idleness index of the basic service set to which the current communication site belongs, determine the main frequency band resource occupation time quota and main frequency band transmission opportunity quota allocated to the site cluster where the communication site is located. Based on the second bearer ratio threshold and the actual secondary band channel idleness index of the basic service set to which the current communication site belongs, determine the secondary band resource occupation time quota and secondary band transmission opportunity quota allocated to the site cluster to which the communication site is located; Based on the primary frequency band resource occupancy time quota, the primary frequency band transmission opportunity quota, the secondary frequency band resource occupancy time quota, and the secondary frequency band transmission opportunity quota, a service message distribution control instruction is constructed and a carrying frequency band is selected for transmission to adjust the service message transmission and reception window of the communication station.

8. The method according to claim 7, characterized in that, The step of constructing a service packet distribution and control instruction based on the primary frequency band resource occupancy time quota, the primary frequency band transmission opportunity quota, the secondary frequency band resource occupancy time quota, and the secondary frequency band transmission opportunity quota, and selecting the bearer frequency band for transmission, in order to adjust the service packet transmission and reception window of the communication station, includes: Based on the main frequency band resource occupancy time quota, the main frequency band transmission opportunity quota, the secondary frequency band resource occupancy time quota, and the secondary frequency band transmission opportunity quota, the basic parameter payload of the service message distribution control instruction for the communication station is generated. Read the length parameter and service quality level flag of each service message frame to be sent in the service message buffer queue to be sent to the communication station. Generate a service message sorting and distribution plan table for the main frequency band signal channel based on the main frequency band resource occupation time quota and the main frequency band transmission opportunity quota. Generate a service message sorting and distribution plan table for the secondary frequency band signal channel based on the secondary frequency band resource occupation time quota and the secondary frequency band transmission opportunity quota. After completing the duplicate transmission avoidance processing of the service packets to be transmitted in the main frequency band signal channel and the auxiliary frequency band signal channel, the service packet sorting and distribution planning table is assembled into the scheduling planning payload part of the distribution control instruction. The channel quality indicator (CMI) evaluation values ​​of the current primary frequency band signal channel and the secondary frequency band signal channel are detected, and the signal frequency band whose CMI evaluation value meets the preset reliability evaluation conditions is selected as the instruction bearer frequency band. The service message distribution and control instruction containing the basic parameter payload and the scheduling planning payload is encapsulated into a unicast management frame and sent to the media access control layer of the communication station through the selected signal frequency band, so as to control the communication station to adjust its service message transmission and reception window according to the pre-planned time window index and frequency band allocation indication in the service message distribution and control instruction.

9. A business message processing server, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program that, when executed by the processor, causes the processor to perform the steps of the service packet processing method based on any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a computer program that, when run on a service message processing server, causes the service message processing server to perform the steps of any of the service message processing methods based on multi-band networks as described in claims 1 to 8.

Images