Microwave carrier aggregation control and configuration method

Abstract

The invention discloses a microwave carrier aggregation control and configuration method, which relates to the technical field of microwave carrier aggregation control and configuration, and comprises the following steps of: acquiring time sequence data in a carrier switching process from a microwave carrier aggregation system in real time through a monitoring system and a network log tool, establishing an original data source, and establishing an original data source; and basic data is provided for subsequent data collection and analysis. Carrier scheduling abnormity is intelligently evaluated through the pre-trained machine learning model, when the mismatching risk is detected, the buffer areas of the sending end and the receiving end are dynamically adjusted according to the severity, data are pre-stored before switching, and the data waiting time is prolonged after switching, so that millisecond-level signal interruption is filled up, and the switching efficiency is improved. Data transmission continuity and system stability are remarkably improved, seamless communication is guaranteed in high-delay sensitive scenes such as financial transactions, intelligent power grid remote control and automatic driving V2X, and safety and economic risks caused by transient signal loss are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of microwave carrier aggregation control and configuration, and in particular to a microwave carrier aggregation control and configuration method. Background Art

[0002] Microwave carrier aggregation control and configuration refers to the process of improving data transmission capacity by aggregating multiple carrier channels in microwave communication systems, and dynamically controlling the allocation, adjustment and management of these carriers. Its core goal is to optimize bandwidth utilization, improve data throughput, and enhance the system's anti-interference ability and reliability. During the configuration process, multiple factors need to be considered, including available spectrum resources, link quality, interference level, and traffic demand, and the combination of carriers is dynamically adjusted through intelligent algorithms (such as adaptive spectrum allocation, power control, and channel equalization). In addition, microwave carrier aggregation usually involves collaborative optimization of the physical layer and the MAC layer to ensure that multiple carriers can work together efficiently and achieve stable multi-carrier transmission, while reducing link congestion and latency and improving overall communication efficiency.

[0003] The prior art has the following deficiencies:

[0004] Currently, many microwave carrier aggregation systems use intelligent dynamic scheduling mechanisms to flexibly adjust carrier allocation based on real-time network traffic and channel quality. However, during the scheduling process, if the switching timing does not match, short signal interruptions at the millisecond level may be difficult to be captured by traditional QoS monitoring mechanisms. In scenarios with extremely low latency requirements (such as financial trading systems and smart grid remote control), this tiny loss may lead to transaction interruptions, equipment mis-triggering, or abnormal data calculation; in autonomous driving V2X communications, it may cause a momentary interruption of the vehicle-mounted microwave link, which in turn causes the vehicle to misjudge environmental information, ultimately leading to unpredictable traffic accidents.

[0005] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0006] The purpose of the present invention is to provide a microwave carrier aggregation control and configuration method, which collects, integrates and preprocesses the switching timing data of the microwave carrier aggregation system in real time, extracts key indicators using feature engineering, and generates a switching timing mismatch risk coefficient with the help of a pre-trained machine learning model to intelligently evaluate carrier scheduling anomalies. When a mismatch risk is detected, the sender and receiver buffers are dynamically adjusted according to the severity, data is pre-stored before switching, and the data waiting time is extended after switching to fill the millisecond signal interruption, significantly improving the continuity of data transmission and system stability, ensuring seamless communication in high-delay sensitive scenarios such as financial transactions, smart grid remote control, and autonomous driving V2X, and reducing the safety and economic risks caused by short-term signal loss, so as to solve the problems in the above-mentioned background technology.

[0007] In order to achieve the above object, the present invention provides the following technical solution: a microwave carrier aggregation control and configuration method, comprising the following steps:

[0008] Through the monitoring system and network log tools, the timing data of the carrier switching process is collected in real time from the microwave carrier aggregation system to establish the original data source and provide basic data for subsequent data collection and analysis;

[0009] After collecting scattered time series data, integrate various data sources into a unified data set;

[0010] After preprocessing the switching timing data in the data set, the key indicators that can reflect the switching timing mismatch are extracted from the preprocessed data through feature engineering methods, and the analyzed key features are input into the pre-trained machine learning model to conduct real-time evaluation of the potential switching timing mismatch risk in the carrier scheduling process;

[0011] Based on the matching risk index output by the machine learning model, it can intelligently sense whether there is a switching timing mismatch in the current carrier allocation and scheduling process;

[0012] When a risk of switching timing mismatch is detected, the buffer settings of the sender and receiver are dynamically adjusted according to the predicted severity. Before the switch, the sender cache size is increased in advance to store more data; after the switch, the data waiting time is extended at the receiver to ensure the continuity and integrity of the data flow.

[0013] Preferably, the specific steps of collecting carrier switching timing data in real time in a microwave carrier aggregation system are as follows:

[0014] Deploy monitoring agents and log collection tools on each key node to ensure that detailed system operation logs and events can be recorded;

[0015] The log data generated by each node is aggregated in real time through a centralized monitoring platform, and all data are timestamped using high-precision clock synchronization to ensure the time accuracy of the data;

[0016] Configure monitoring rules and data collection strategies to automatically identify and capture key carrier switching events, and parse key switching events into structured data;

[0017] The collected time series data is stored in a database or time series database for subsequent analysis.

[0018] Preferably, in the preprocessed data, key indicators that can reflect the switching timing mismatch are extracted by feature engineering methods, and the specific steps are as follows:

[0019] The key features reflecting the switching timing mismatch are extracted from the pre-processed switching timing data, wherein the extracted key features include the timing overlap degree between the old carrier and the new carrier in the switching window when the old carrier has not been completely withdrawn and the new carrier has begun to be established, and the degree of disorder of the data packet sequence before and after the switching. After analyzing the extracted key features through feature engineering technology, the switching overlap distortion factor and the data packet disorder rate factor are generated respectively. The switching overlap distortion factor and the data packet disorder rate factor are used as key indicators reflecting the switching timing mismatch. The timing mismatch risk in the carrier switching process is preliminarily quantified, and the abnormal connection degree and data flow synchronization deviation in the switching window are evaluated, so as to identify potential switching abnormality risks.

[0020] Preferably, the analyzed switching overlap distortion factor and data packet disorder rate factor are input into a pre-trained machine learning model, and a switching timing mismatch risk coefficient is generated based on the machine learning model. The switching timing mismatch risk coefficient is used to perform a real-time evaluation of the potential switching timing mismatch risk in the carrier scheduling process, and the potential switching timing mismatch risk in the carrier scheduling process is intelligently predicted.

[0021] Preferably, the switching timing mismatch risk coefficient generated when the mismatch risk in the carrier scheduling process is predicted by the machine learning model is compared and analyzed with the pre-set risk coefficient reference threshold, and intelligent perception is performed on whether there is a switching timing mismatch in the current carrier allocation scheduling process. The specific steps are as follows:

[0022] If the switching timing mismatch risk coefficient is greater than the risk coefficient reference threshold, the current carrier allocation scheduling process is judged as having a switching timing mismatch; if the switching timing mismatch risk coefficient is less than or equal to the risk coefficient reference threshold, the current carrier allocation scheduling process is judged as having no switching timing mismatch.

[0023] Preferably, after analyzing the timing overlap degree of the old carrier in the switching window when the old carrier has not been completely withdrawn and the new carrier has begun to be established by feature engineering technology, a switching overlap distortion factor is generated. The specific steps are as follows:

[0024] Define the time overlap weight function to quantify the proportion of the old carrier and the new carrier in the switching window. The expression of the time overlap weight function is:

[0025] ,

[0026] Where: Φ overlap It indicates the timing overlap degree of the old carrier and the new carrier in the switching window, and measures the mutual interference degree of the old carrier and the new carrier in the switching window. old (ξ) is the normalized representation of the old carrier signal power at time point ξ, 0≤P old (ξ)≤1,P new (ξ) is the normalized representation of the new carrier signal power at time point ξ, P new (ξ)≤1, Ω is the time range of the switching window, that is, the carrier switching time interval expected by the system, and γ is the adjustment factor used to control the influence of time overlap on the final index;

[0027] After determining the timing overlap degree of the old carrier and the new carrier within the switching window, the final switching overlap distortion factor is calculated by further combining the signal change rate and link state change. The calculation expression is:

[0028] ,

[0029] Where: S odi is the switching overlap distortion factor, which is used to quantify the degree of timing mismatch during carrier switching. λ is the index adjustment factor, which controls the weight of time overlap in the total index. is the instantaneous rate of change of the old carrier power over time, which measures the attenuation speed of the old carrier signal. is the instantaneous rate of change of the new carrier power over time, which measures the growth rate of the new carrier signal. β is a control parameter used to control the influence of the power change rate. η is the adjustment ratio weight, which controls the influence between different switching rates.

[0030] Preferably, after analyzing the disorder degree of the extracted data packet sequence before and after the switching by feature engineering technology, a data packet disorder rate factor is generated, and the specific steps are as follows:

[0031] For each data packet, calculate its "sequence offset" during the switching process, that is, the offset of the current data packet relative to its original sequence position. By comparing the actual receiving order of each data packet with the expected order, the sequence offset of each data packet is obtained. Suppose the expected order and actual receiving order of the i-th data packet are p and p, respectively. i and r i , i represents the index of the data packet, then the calculation expression of the sequence offset is:

[0032] Δp i =|r i -p i |,

[0033] Among them, Δp i represents the sequence offset of the ith data packet,

[0034] After obtaining the order deviation of each data packet, the data packet disorder rate factor of the entire data stream is calculated. The calculation of the data packet disorder rate factor integrates the deviation degree of all data packets and weightedly considers the relative position of the data packets. The specific calculation expression is:

[0035] ,

[0036] Among them, PO ri is the packet disorder rate factor, N is the total number of packets, α and μ are weight coefficients that adjust the degree of disorder.

[0037] Preferably, when it is detected that there is a risk of mismatch in the switching timing, the buffer settings of the transmitting end and the receiving end are dynamically adjusted according to the predicted severity. The specific steps are as follows:

[0038] When a switching timing mismatch risk is detected, the amplitude of the buffer adjustment is first determined. In order to ensure the continuity of the data flow, the sender buffer compensation amount and the receiver buffer extension amount are dynamically adjusted according to the size of the risk coefficient, so that they can fill the short data flow interruption caused by the switching timing mismatch. The adjustment amplitude matches the switching timing mismatch risk coefficient and is affected by the risk coefficient reference threshold. The calculation formula is as follows:

[0039] ,

[0040] Among them: B tx is the buffer compensation amount at the sender, that is, the amount of data that needs to be additionally buffered before switching to fill the gap during switching. rx The amount of buffer extension at the receiving end, that is, the buffered data time needs to be extended after switching to absorb the data misalignment problem caused by switching delay. risk is the risk factor of switch timing mismatch, calculated by the machine learning model.th is the reference threshold of the risk factor, γ1 and γ2 are buffer adjustment coefficients, reflecting the sensitivity of the system to the switching timing mismatch, and β1 and β2 are buffer nonlinear adjustment indexes, which are used to enhance the compensation capability for timing mismatch.

[0041] Preferably, after calculating the sender cache compensation amount, the sender will dynamically adjust the cache strategy and store more data before switching to ensure that the receiving end can still process data normally even if the data stream is temporarily lost during the switching process. The optimized sender cache filling formula is as follows:

[0042] ,

[0043] Where: C tx is the optimized sender cache size, is the original sender buffer size, the default value in normal switching, λ1 is the sender buffer increment adjustment coefficient, which determines the increase in buffer size, δ sync is the synchronization error between the new and old carriers, indicating the degree of deviation of the current carrier switching synchronization. η1 is the synchronization error attenuation factor, which controls the impact of the synchronization error on the sender buffer.

[0044] Preferably, when the receiving end detects the mismatch risk, it is necessary to extend the data cache time to absorb the disorder or loss of data packets caused by the switching. The adjusted receiving end data waiting time calculation formula is as follows:

[0045] ,

[0046] Where: T rx is the optimized receiving end data waiting time, is the original receiving end data waiting time, the default value under normal circumstances, λ2 is the receiving end buffer adjustment coefficient, controlling the adjustment amplitude, Δ jitter is the jitter change of the data stream before and after the switch, reflecting the stability of the data stream during the switch. η2 is the jitter attenuation factor, which controls the impact of data jitter on buffer adjustment.

[0047] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0048] The present invention collects, integrates and pre-processes the switching timing data of each key node (such as base station, transmission equipment and QoS monitoring tool) in the microwave carrier aggregation system in real time, and uses feature engineering technology to extract key indicators such as the overlap between the exit of the old carrier and the establishment of the new carrier and the disorder of the data packet, and then inputs these analyzed indicators into the pre-trained machine learning model, so as to intelligently generate the switching timing mismatch risk coefficient, and evaluate the potential anomalies in the carrier scheduling process in real time based on this risk coefficient. When the switching timing mismatch risk is detected, the buffer settings of the transmitter and the receiver are further dynamically adjusted according to the severity of the risk, the buffer amount of the transmitter is increased in advance before the switching, more data is pre-stored, and the data waiting time of the receiver is extended after the switching, so as to effectively fill the millisecond signal interruption caused by the carrier switching mismatch, thereby significantly improving the continuity of data transmission and the overall stability of the system, ensuring efficient and seamless data transmission in application scenarios with extremely high latency requirements such as financial transactions, smart grid remote control, and autonomous driving V2X communication, and greatly reducing the safety and economic risks caused by short-term signal loss. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0050] Figure 1 The present invention is a flowchart of the microwave carrier aggregation control and configuration method. DETAILED DESCRIPTION

[0051] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.

[0052] The present invention provides Figure 1 The microwave carrier aggregation control and configuration method shown includes the following steps:

[0053] Through the monitoring system and network log tools, the timing data of the carrier switching process is collected in real time from the microwave carrier aggregation system to establish the original data source and provide basic data for subsequent data collection and analysis;

[0054] The timing data during the carrier switching process includes the timestamps of carrier activation and release, real-time network traffic, channel quality indicators (such as signal-to-noise ratio, bit error rate, signal power), and scheduling decision time.

[0055] The specific steps for real-time collection of carrier switching timing data in a microwave carrier aggregation system are as follows: deploy monitoring agents and log collection tools on each key node (such as base stations, transmission equipment, and scheduling servers) to ensure that these devices can record detailed system operation logs and events; aggregate the log data generated by each node in real time through a centralized monitoring platform (such as ELK Stack or Prometheus), and use high-precision clock synchronization (such as PTP or NTP) to timestamp all data to ensure the time accuracy of the data; configure monitoring rules and data collection strategies to automatically identify and capture key switching events such as carrier activation, carrier release, and channel quality changes, and parse these events into structured data; store the collected time series data in a database or time series database for subsequent preprocessing, feature extraction, and further analysis.

[0056] After collecting scattered time series data, various data sources (such as base station logs, transmission equipment monitoring data, QoS reports, etc.) are integrated into a unified data set;

[0057] At this point, database or data warehouse technology can be used to store and index data according to time series and different carrier identifiers. Data aggregation can not only achieve centralized data management, but also facilitate subsequent data preprocessing and feature engineering. The main function of this step is to ensure data integrity and consistency, and provide a structured and standardized data foundation for subsequent analysis work.

[0058] After preprocessing the switching timing data in the data set, the key indicators that can reflect the switching timing mismatch are extracted from the preprocessed data through feature engineering methods, and the analyzed key features are input into the pre-trained machine learning model to conduct real-time evaluation of the potential switching timing mismatch risk in the carrier scheduling process;

[0059] Preprocessing mainly includes data cleaning (eliminating errors, duplications and abnormal data), noise removal, missing value completion, data standardization and time synchronization correction. The role of preprocessing is to make the integrated data set high-quality and consistent, laying a solid foundation for subsequent feature extraction and machine learning model training.

[0060] In the preprocessed data, the key indicators that can reflect the switching timing mismatch are extracted through feature engineering methods. The specific steps are as follows:

[0061] The key features reflecting the switching timing mismatch are extracted from the pre-processed switching timing data, wherein the extracted key features include the timing overlap degree between the old carrier and the new carrier in the switching window when the old carrier has not been completely withdrawn and the new carrier has begun to be established, and the degree of disorder of the data packet sequence before and after the switching. After analyzing the extracted key features through feature engineering technology, the switching overlap distortion factor and the data packet disorder rate factor are generated respectively. The switching overlap distortion factor and the data packet disorder rate factor are used as key indicators reflecting the switching timing mismatch. The timing mismatch risk in the carrier switching process is preliminarily quantified, and the abnormal connection degree and data flow synchronization deviation in the switching window are evaluated, so as to identify potential switching abnormality risks.

[0062] During the carrier switching process, the old carrier has not been completely withdrawn, while the new carrier has begun to be established, and the two have a high degree of timing overlap within the switching window, indicating that there is a risk of timing mismatch in the switching process. Under normal circumstances, carrier switching should follow the principles of precise synchronization and seamless connection to ensure that the new carrier can stably take over data transmission after the resources of the old carrier are completely released. However, when the old carrier is still active but attenuated in the switching window, and the new carrier has been involved in advance but has not yet been completely stable, it may cause multiple problems, such as: signal phase interference, power mutation, data flow disorder or short-term link blocking. This overlapping distortion may cause microsecond or even millisecond instantaneous signal conflicts, causing abnormal behavior of data packets at the transmission layer, such as burst packet loss, increased retransmission rate, and packet disorder, which in turn affects the throughput and delay control of the system. In high-delay sensitive scenarios (such as financial transactions, smart grids, and autonomous driving V2X communications), this switching anomaly may cause the loss or false triggering of key control signals. Therefore, a high degree of timing overlap is usually regarded as an important signal of carrier switching timing mismatch.

[0063] After analyzing the timing overlap between the old carrier and the new carrier in the switching window when the old carrier has not been completely withdrawn and the new carrier has begun to be established through feature engineering technology, the switching overlap distortion factor is generated. The specific steps are as follows:

[0064] During the carrier switching process, the time period when the old carrier signal has not completely decayed may overlap with the time period when the new carrier signal has been established in advance but has not yet stabilized, resulting in signal interference and data flow dislocation. Therefore, it is necessary to first define the time overlap weight function to quantify the proportion of the old carrier and the new carrier in the switching window. The expression of the time overlap weight function is:

[0065] ,

[0066] Where: Φ overlap It indicates the timing overlap degree of the old carrier and the new carrier in the switching window, and measures the mutual interference degree of the old carrier and the new carrier in the switching window. old(ξ) is the normalized representation of the old carrier signal power at time point ξ, 0≤P old (ξ)≤1,P new (ξ) is the normalized representation of the new carrier signal power at time point ξ, P new (ξ)≤1, Ω is the time range of the switching window, that is, the carrier switching time interval expected by the system, γ is the adjustment factor, which is not specifically limited here and can be set according to actual needs to control the influence of time overlap on the final index. γ>1 makes the larger overlap weight more obvious;

[0067] The system-estimated carrier switching time interval refers to the theoretical time range of carrier switching that is calculated in advance by the microwave carrier aggregation system during dynamic scheduling based on factors such as channel quality, network load, and signal interference. This interval is usually determined by a network scheduling algorithm or prediction model to guide the system to complete the release of the old carrier and the establishment of the new carrier at the appropriate time point to achieve seamless switching. This time interval not only takes into account the trend of channel quality fluctuations, but also combines historical switching data and estimated system delays to ensure that the switching process is as stable as possible. For example, in high-reliability communication scenarios (such as smart grid remote control and autonomous driving V2X communication), this interval needs to be accurate to microseconds to avoid short signal interruptions or data transmission anomalies caused by switching timing mismatches. Therefore, the system-estimated carrier switching time interval is essentially a theoretical optimal switching window. Completing the switching within this time period can minimize signal jitter, packet loss, and service quality degradation.

[0068] The purpose of this step is to calculate the relative power distribution of the old and new carriers in the switching window and measure the degree of their timing overlap. When the proportion of the old and new carriers in the switching window is high, it means that the old carrier still occupies a large signal energy in the switching window, while the new carrier has been involved in advance, which may cause the switching timing to be mismatched.

[0069] After determining the timing overlap degree of the old carrier and the new carrier within the switching window, the final switching overlap distortion factor is calculated by further combining the signal change rate and link status change. The calculation expression is:

[0070] ,

[0071] Where: S odi is the switching overlap distortion factor, which is used to quantify the degree of timing mismatch during carrier switching. The larger the value, the more serious the timing mismatch. λ is the exponential adjustment factor, which controls the weight of time overlap in the total index (usually 1<λ<3). is the instantaneous rate of change of the old carrier power over time, which measures the attenuation speed of the old carrier signal. is the instantaneous rate of change of the new carrier power over time, which measures the growth rate of the new carrier signal. β is a control parameter used to control the degree of influence of the power change rate. β>1 makes the mutation more significant. η is the adjustment ratio weight, which controls the influence between different switching rates.

[0072] The purpose of this step is to quantify the mismatch between the old carrier attenuation rate and the new carrier establishment rate. That is, if the old carrier attenuation rate does not match the new carrier establishment rate (the difference is large), packet loss and uneven signal transition are likely to occur during the switching process. The final switching overlap distortion factor combines the time overlap and signal change rate to accurately quantify the timing mismatch within the switching window, providing a reference for subsequent optimization.

[0073] It can be seen from the switching overlap distortion factor that after analyzing the timing overlap degree of the old carrier and the new carrier in the switching window when the old carrier has not been completely withdrawn and the new carrier has begun to be established through feature engineering technology, the switching overlap distortion factor is generated. The larger the performance value of the switching overlap distortion factor, the greater the risk of switching timing mismatch during carrier switching, and vice versa. The switching overlap distortion factor quantifies the timing overlap degree of the old carrier and the new carrier in the switching window, reflecting the relative power overlap degree between the new and old carriers during the switching process. If the switching overlap distortion factor is high, it means that when the new carrier has not been fully established, the old carrier still occupies the signal transmission resources to a large extent, resulting in the overlap of the two in the same time period, which may cause signal interference, data packet misordering or signal attenuation, thereby increasing the risk of switching timing mismatch. On the contrary, if the switching overlap distortion factor is low, it means that the timing of the two is more coordinated during the switching process, the signal transition is smooth, and the risk of switching timing mismatch is small. Therefore, the switching overlap distortion factor is an effective indicator for evaluating the stability and timing matching degree of carrier switching.

[0074] During the carrier switching process, the order of data packets before and after the switch is highly chaotic, which usually means that there is a greater risk of switching timing mismatch. During the carrier switching process, the order of data packets is the key to ensuring the integrity and correctness of the data stream. Under normal circumstances, the switching process should be smooth to ensure that after the data transmission of the old carrier is completed, the new carrier immediately takes over and continues to transmit seamlessly. However, if the switching timing does not match, that is, the old carrier fails to be released in time or the new carrier fails to start at the scheduled time point, it may cause the order of data packets to be chaotic, that is, the data packets are out of order or lost. This chaos reflects the instability of the signal link, which may cause data flow disruption, increased transmission delay, and even seriously affect the system performance and reliability of scenarios with high latency requirements (such as financial transactions, autonomous driving communications, etc.). Therefore, a high degree of chaos in the order of data packets is usually a direct sign of timing mismatch.

[0075] After analyzing the disorder degree of the extracted data packet sequence before and after the switch through feature engineering technology, the data packet disorder rate factor is generated. The specific steps are as follows:

[0076] For each data packet, calculate its "sequence offset" during the switching process, that is, the offset of the current data packet relative to its original sequence position. By comparing the actual receiving order of each data packet with the expected order, the sequence offset of each data packet is obtained. Suppose the expected order and actual receiving order of the i-th data packet are p and p, respectively. i and r i , i represents the index of the data packet, then the calculation expression of the sequence offset is:

[0077] Δp i =|r i -p i |

[0078] , where Δp i It indicates the order deviation of the ith data packet. The larger the order deviation, the higher the degree of disorder of the data packet.

[0079] The purpose of this step is to capture subtle changes in the order of data packets by quantifying the amount of disorder of each data packet, and further evaluate the potential risks caused by switching timing mismatch.

[0080] After obtaining the order deviation of each data packet, the data packet disorder rate factor of the entire data stream is calculated. The calculation of the data packet disorder rate factor integrates the deviation degree of all data packets and weightedly considers the relative position of the data packets. The specific calculation expression is:

[0081] ,

[0082] Among them, PO ri is the packet disorder rate factor, N is the total number of packets, α and μ are weight coefficients, which adjust the degree of disorder, and can generally be adjusted according to actual application requirements;

[0083] This formula is based on the relative offset reflects the relative disorder of each data packet relative to its expected position, while the denominator The higher position packets (usually key packets) are given more weight to ensure that the system is more sensitive to out-of-order. The role of this step is to integrate the out-of-order degree of all packets and weight them according to the importance of the packets, thereby generating a packet out-of-order rate factor that fully reflects the degree of switching timing mismatch.

[0084] It can be seen from the packet disorder rate factor that the packet disorder rate factor is generated after analyzing the disorder degree of the extracted packet sequence before and after the switching through feature engineering technology. The larger the performance value of the packet disorder rate factor, the higher the risk of switching timing mismatch during carrier switching, and vice versa, it indicates that the switching timing is relatively stable. The packet disorder rate factor is calculated based on the order deviation of the data packet and the relative disorder weight. When the timing mismatch occurs during the switching process (such as the new carrier is not fully established and the old carrier is released prematurely), the receiving order of the data packet will deviate greatly from its expected order, resulting in aggravated data flow disorder and an increase in the calculated value of the packet disorder rate factor. A higher packet disorder rate factor value means that the data flow has a significant disorder or dislocation within the switching window, which may cause problems such as data reassembly failure at the receiving end, increased packet loss rate or sudden increase in delay, thereby affecting business continuity. A lower packet disorder rate factor value indicates that the packet order remains basically stable, the switching timing match is good, and the data flow transmission is relatively stable. Therefore, the packet disorder rate factor can be used as an effective indicator to quantify the degree of switching timing mismatch.

[0085] The analyzed switching overlap distortion factor and data packet disorder rate factor are input into the pre-trained machine learning model, and the switching timing mismatch risk coefficient is generated based on the machine learning model. The switching timing mismatch risk coefficient is used to evaluate the potential switching timing mismatch risk in the carrier scheduling process in real time, and the potential switching timing mismatch risk in the carrier scheduling process is intelligently predicted.

[0086] A pre-trained machine learning model refers to an algorithm model that is fixed and embedded in the system after optimizing and tuning the model parameters using a large amount of historical data and annotated data through offline learning before actual deployment. In this system, the model mainly uses the handover overlap distortion factor and packet disorder rate factor collected historically as input features, and models the handover timing mismatch that occurs during carrier scheduling through supervised learning methods. During the training process, developers collect a large amount of handover data in different scenarios, and manually or automatically annotate whether each handover event has a serious mismatch risk to form a sample library; then, through feature engineering, key indicators that can reflect handover timing mismatch are extracted, and these features are modeled using algorithms such as random forests, support vector machines, neural networks (such as LSTM or CNN), generating a model that can accurately predict risks. This process includes steps such as data preprocessing, feature selection, model training, cross-validation, and parameter tuning, and ultimately obtains a stable model that can output the handover timing mismatch risk coefficient in real time after inputting new feature data. Pre-trained machine learning models have strong generalization and real-time response capabilities. After deployment, they can perform risk assessment and prediction based on current real-time input data without the need to train from scratch each time, greatly improving the efficiency and response speed of the system.

[0087] Once the pre-trained machine learning model is trained, it can be directly deployed to the microwave carrier aggregation system, and in actual operation, it receives key feature parameters extracted from the switching timing data and processed by feature engineering in real time. After the model receives these input parameters, it will output a quantified switching timing mismatch risk coefficient through internal calculation logic and weight mapping. The switching timing mismatch risk coefficient reflects the potential mismatch risk level in the current carrier scheduling process. Based on this, the system can evaluate possible anomalies in the carrier switching process in real time.

[0088] The pre-trained machine learning model not only significantly shortens the response time of real-time risk prediction, but also continuously optimizes its own performance through continuous feedback and online learning, ensuring high accuracy and robustness in complex and changing communication environments, thereby providing intelligent and precise risk warning support for the carrier scheduling process.

[0089] The machine learning model is not specifically limited here, and can achieve the switching overlap distortion factor S odi and the packet out-of-order rate factor PO ri Perform comprehensive analysis to generate the switch timing mismatch risk factor Switch risk In order to realize the technical solution of the present invention, the present invention provides a specific implementation method; Switch timing mismatch risk coefficient Switch riskThe generated calculation formula is: Switch risk =α1*S odi +α2*PO ri , where α1 and α2 are the switching overlap distortion factors S odi and the packet out-of-order rate factor PO ri The preset proportional coefficient is the preset proportional coefficient given to different characteristic indicators (switching overlap distortion factor S odi and the packet out-of-order rate factor PO ri ) are used to adjust the influence of each indicator on the final switching timing mismatch risk coefficient. These preset proportional coefficients are usually pre-set based on historical data analysis, expert experience or machine learning training to ensure that the calculated switching timing mismatch risk coefficient can reasonably reflect the actual switching timing mismatch risk. For example, if experiments or model training show that switching overlap distortion has a greater impact on the system, α1>α2 can be set to emphasize the contribution of the switching overlap distortion factor; conversely, if the packet disorder rate has a more significant impact on system stability, the weight of α2 may be adjusted to be higher. By pre-setting the proportional coefficient, the needs of different systems can be flexibly adapted to make the calculated risk coefficient more accurate and in line with the actual situation.

[0090] It can be seen from the switching timing mismatch risk coefficient that after analyzing the degree of timing overlap between the old carrier and the new carrier in the switching window when the old carrier has not been completely withdrawn and the new carrier has begun to be established through feature engineering technology, a switching overlap distortion factor is generated. The larger the performance value of the switching overlap distortion factor, the larger the performance value of the packet disorder rate factor is generated after analyzing the degree of confusion of the extracted data packet sequence before and after the switching through feature engineering technology. That is, when the switching timing mismatch risk in the carrier scheduling process is predicted by the machine learning model, the larger the performance value of the switching timing mismatch risk coefficient generated is, indicating that the risk of switching timing mismatch in the carrier switching process is higher, and vice versa, it indicates that the risk of switching timing mismatch in the carrier switching process is lower.

[0091] Based on the matching risk index output by the machine learning model, it can intelligently sense whether there is a switching timing mismatch in the current carrier allocation and scheduling process;

[0092] The switching timing mismatch risk coefficient generated when the mismatch risk in the carrier scheduling process is predicted by the machine learning model is compared and analyzed with the pre-set risk coefficient reference threshold, so as to intelligently sense whether there is a switching timing mismatch in the current carrier allocation scheduling process. The specific steps are as follows:

[0093] If the switching timing mismatch risk coefficient is greater than the risk coefficient reference threshold, the current carrier allocation scheduling process is judged as having a switching timing mismatch; if the switching timing mismatch risk coefficient is less than or equal to the risk coefficient reference threshold, the current carrier allocation scheduling process is judged as having no switching timing mismatch.

[0094] When a mismatch risk in the switching sequence is detected, the buffer settings of the sender and receiver are dynamically adjusted according to the predicted severity. Before the switch, the sender cache is increased in advance to store more data; after the switch, the data waiting time is extended at the receiver to ensure the continuity and integrity of the data flow.

[0095] This dynamic buffer mechanism needs to be precisely controlled within the microsecond time scale to prevent data interruption and not introduce excessive latency. The purpose of this step is to smooth the brief signal interruptions that may occur during the switching process through buffer adjustment, ensuring the stability and reliability of data transmission of the entire system and meeting the strict requirements of low-latency applications.

[0096] When a mismatch risk is detected in the switching timing, the buffer settings of the transmitter and receiver are dynamically adjusted according to the predicted severity. The specific steps are as follows:

[0097] When a switching timing mismatch risk is detected, the buffer adjustment range is first determined. To ensure the continuity of the data flow, the sender buffer compensation amount and the receiver buffer extension amount are dynamically adjusted according to the risk coefficient to fill the short data flow interruption caused by the switching timing mismatch. The adjustment range should match the switching timing mismatch risk coefficient and be affected by the risk coefficient reference threshold. The calculation formula is as follows:

[0098] ,

[0099] Among them: B tx is the buffer compensation amount at the sender, that is, the amount of data that needs to be additionally buffered before switching to fill the gap during switching. rx The amount of buffer extension at the receiving end, that is, the buffered data time needs to be extended after switching to absorb the data misalignment problem caused by switching delay. risk is the risk factor of switch timing mismatch, calculated by the machine learning model. th is the reference threshold of the risk factor, indicating the maximum acceptable degree of timing mismatch of the system; γ1 and γ2 are buffer adjustment coefficients, reflecting the sensitivity of the system to switching timing mismatch. The larger the value, the more obvious the adjustment; β1 and β2 are buffer nonlinear adjustment exponents, which are used to enhance the compensation capability for larger timing mismatches, so that the adjustment amount increases exponentially when the risk increases;

[0100] The core function of this step is to quantitatively adjust the needs of the sender and receiver caches according to the current switching timing mismatch risk of the system, and provide data support for subsequent dynamic adjustments.

[0101] After calculating the sender's buffer compensation amount, the sender will dynamically adjust the buffer strategy and store more data before switching to ensure that the receiver can still process data normally even if the data stream is temporarily lost during the switching process. The optimized sender buffer filling formula is as follows:

[0102] ,

[0103] Where: C tx is the optimized sender cache size, is the original sender buffer size, the default value in normal switching, λ1 is the sender buffer increment adjustment coefficient, which determines the increase in buffer size, δ sync is the synchronization error between the new and old carriers, indicating the degree of deviation of the current carrier switching synchronization; η1 is the synchronization error attenuation factor, controlling the influence of the synchronization error on the sender buffer;

[0104] The synchronization error between the new and old carriers refers to the time deviation between the establishment time of the new carrier and the release time of the old carrier during the carrier switching process of the microwave carrier aggregation system, which makes it impossible for the data stream to be seamlessly connected within the switching window. This error may come from multiple factors, including the drift of the hardware clock, the delay of the carrier scheduling strategy, the fluctuation of the link delay, and the dynamic change of the channel quality. When the synchronization error is large, the new carrier may not be fully established, and the old carrier may be released, resulting in a short gap in data transmission, causing signal loss or instantaneous link disconnection; conversely, if the new and old carriers overlap for a certain period of time, phase misalignment, signal interference or data packet duplication may occur. Therefore, during the carrier switching process, accurate control of the synchronization error is a key factor in ensuring the stability of the data stream and reducing the risk of timing mismatch.

[0105] This step ensures that the sender actively increases the data buffer before the switch occurs to reduce the data flow interruption problem caused by timing mismatch. At the same time, the exponential decay function is used to balance the impact of synchronization error on cache increment, so that no additional storage resources will be wasted when the synchronization is good.

[0106] When the receiving end detects the risk of mismatch, it needs to extend the data cache time to absorb the disorder or loss of data packets caused by switching. The calculation formula for the adjusted receiving end data waiting time is as follows:

[0107] ,

[0108] Where: T rxis the optimized receiving end data waiting time, is the original receiving end data waiting time, the default value under normal circumstances, λ2 is the receiving end buffer adjustment coefficient, controlling the adjustment amplitude, Δ jitter is the jitter change of the data stream before and after the switch, reflecting the stability of the data stream during the switch; η2 is the jitter attenuation factor, controlling the impact of data jitter on buffer adjustment;

[0109] The jitter change of the data stream before and after the switch refers to the change in the degree of fluctuation of the packet arrival time before and after the switch during the carrier switching process. It is used to measure the impact of the switch on the stability of the data stream. Jitter is essentially a random fluctuation in the delay of packet transmission, usually caused by factors such as channel quality fluctuations, queue congestion, and changes in network load. When a carrier switch occurs, if the timing of the new carrier is not synchronized with the old carrier, or the queue management of the data packet is unstable during the switching process, it may cause a significant increase in jitter, making the packet interval at the receiving end uneven, and even causing short-term data accumulation or gaps, affecting real-time applications (such as VoIP, V2X communications, and financial transactions). The jitter change is quantified by calculating the jitter mean difference or standard deviation change before and after the switch. If Δ jitter If it is too high, it means that the switching has a great impact on the stability of the data stream, and the buffering strategy of the receiving end needs to be adjusted to absorb jitter and reduce the impact of data confusion.

[0110] The main function of this step is to adjust the buffer time intelligently so that the receiving end can absorb data jitter and disorder after detecting the switching mismatch, ensuring that the data flow remains stable during the switching. By weighted adjustment of the jitter change of the data flow before and after the switching, the buffer time can be automatically increased in the case of large data jitter, reducing the risk of data misalignment.

[0111] The present invention collects, integrates and pre-processes the switching timing data of each key node (such as base station, transmission equipment and QoS monitoring tool) in the microwave carrier aggregation system in real time, and uses feature engineering technology to extract key indicators such as the overlap between the exit of the old carrier and the establishment of the new carrier and the disorder of the data packet, and then inputs these analyzed indicators into the pre-trained machine learning model, so as to intelligently generate the switching timing mismatch risk coefficient, and evaluate the potential anomalies in the carrier scheduling process in real time based on this risk coefficient. When the switching timing mismatch risk is detected, the buffer settings of the transmitter and the receiver are further dynamically adjusted according to the severity of the risk, the buffer amount of the transmitter is increased in advance before the switching, more data is pre-stored, and the data waiting time of the receiver is extended after the switching, so as to effectively fill the millisecond signal interruption caused by the carrier switching mismatch, thereby significantly improving the continuity of data transmission and the overall stability of the system, ensuring efficient and seamless data transmission in application scenarios with extremely high latency requirements such as financial transactions, smart grid remote control, and autonomous driving V2X communication, and greatly reducing the safety and economic risks caused by short-term signal loss.

[0112] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0113] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0114] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A microwave carrier aggregation control and configuration method, characterized in that: The following steps are involved: Through the monitoring system and network log tools, the timing data of the carrier switching process is collected in real time from the microwave carrier aggregation system to establish the original data source and provide basic data for subsequent data collection and analysis; After collecting scattered time series data, integrate various data sources into a unified data set; After preprocessing the switching timing data in the data set, the key indicators that can reflect the switching timing mismatch are extracted from the preprocessed data through feature engineering methods, and the analyzed key features are input into the pre-trained machine learning model to conduct real-time evaluation of the potential switching timing mismatch risk in the carrier scheduling process; Based on the matching risk index output by the machine learning model, it can intelligently sense whether there is a switching timing mismatch in the current carrier allocation and scheduling process; When a risk of switching timing mismatch is detected, the buffer settings of the sender and receiver are dynamically adjusted according to the predicted severity. Before the switch, the sender cache size is increased in advance to store more data; after the switch, the data waiting time is extended at the receiver to ensure the continuity and integrity of the data flow.

2. The microwave carrier aggregation control and configuration method according to claim 1, characterized in that: The specific steps for real-time acquisition of carrier switching timing data in a microwave carrier aggregation system are as follows: Deploy monitoring agents and log collection tools on each key node to ensure that detailed system operation logs and events can be recorded; The log data generated by each node is aggregated in real time through a centralized monitoring platform, and all data are timestamped using high-precision clock synchronization to ensure the time accuracy of the data; Configure monitoring rules and data collection strategies to automatically identify and capture key carrier switching events, and parse key switching events into structured data; The collected time series data is stored in a database or time series database for subsequent analysis.

3. The microwave carrier aggregation control and configuration method according to claim 1, characterized in that: In the preprocessed data, the key indicators that can reflect the switching timing mismatch are extracted through feature engineering methods. The specific steps are as follows: The key features reflecting the switching timing mismatch are extracted from the pre-processed switching timing data, wherein the extracted key features include the timing overlap degree between the old carrier and the new carrier in the switching window when the old carrier has not been completely withdrawn and the new carrier has begun to be established, and the degree of disorder of the data packet sequence before and after the switching. After analyzing the extracted key features through feature engineering technology, the switching overlap distortion factor and the data packet disorder rate factor are generated respectively. The switching overlap distortion factor and the data packet disorder rate factor are used as key indicators reflecting the switching timing mismatch. The timing mismatch risk in the carrier switching process is preliminarily quantified, and the abnormal connection degree and data flow synchronization deviation in the switching window are evaluated, so as to identify potential switching abnormality risks.

4. The microwave carrier aggregation control and configuration method according to claim 3, characterized in that: The analyzed switching overlap distortion factor and data packet disorder rate factor are input into the pre-trained machine learning model, and the switching timing mismatch risk coefficient is generated based on the machine learning model. The switching timing mismatch risk coefficient is used to evaluate the potential switching timing mismatch risk in the carrier scheduling process in real time, and the potential switching timing mismatch risk in the carrier scheduling process is intelligently predicted.

5. The microwave carrier aggregation control and configuration method according to claim 4, characterized in that: The switching timing mismatch risk coefficient generated when the mismatch risk in the carrier scheduling process is predicted by the machine learning model is compared and analyzed with the pre-set risk coefficient reference threshold, so as to intelligently sense whether there is a switching timing mismatch in the current carrier allocation scheduling process. The specific steps are as follows: If the switching timing mismatch risk coefficient is greater than the risk coefficient reference threshold, the current carrier allocation scheduling process is judged as having a switching timing mismatch; if the switching timing mismatch risk coefficient is less than or equal to the risk coefficient reference threshold, the current carrier allocation scheduling process is judged as having no switching timing mismatch.

6. The microwave carrier aggregation control and configuration method according to claim 3, characterized in that: After analyzing the timing overlap between the old carrier and the new carrier in the switching window when the old carrier has not been completely withdrawn and the new carrier has begun to be established through feature engineering technology, the switching overlap distortion factor is generated. The specific steps are as follows: Define the time overlap weight function to quantify the proportion of the old carrier and the new carrier in the switching window. The expression of the time overlap weight function is: ， Where: Φ overlap It indicates the timing overlap degree of the old carrier and the new carrier in the switching window, and measures the mutual interference degree of the old carrier and the new carrier in the switching window. old (ξ) is the normalized representation of the old carrier signal power at time point ξ, 0≤P old (ξ)≤1,P new (ξ) is the normalized representation of the new carrier signal power at time point ξ, P new (ξ)≤1, Ω is the time range of the switching window, that is, the carrier switching time interval expected by the system, and γ is the adjustment factor used to control the influence of time overlap on the final index; After determining the timing overlap degree of the old carrier and the new carrier within the switching window, the final switching overlap distortion factor is calculated by further combining the signal change rate and link state change. The calculation expression is: ， Where: S odi is the switching overlap distortion factor, which is used to quantify the degree of timing mismatch during carrier switching. λ is the index adjustment factor, which controls the weight of time overlap in the total index. is the instantaneous rate of change of the old carrier power over time, which measures the attenuation speed of the old carrier signal. is the instantaneous rate of change of the new carrier power over time, which measures the growth rate of the new carrier signal. β is a control parameter used to control the influence of the power change rate. η is the adjustment ratio weight, which controls the influence between different switching rates.

7. The microwave carrier aggregation control and configuration method according to claim 3, characterized in that: After analyzing the disorder degree of the extracted data packet sequence before and after the switch through feature engineering technology, the data packet disorder rate factor is generated. The specific steps are as follows: For each data packet, calculate its "sequence offset" during the switching process, that is, the offset of the current data packet relative to its original sequence position. By comparing the actual receiving sequence and expected sequence of each data packet, the sequence offset of each data packet is obtained. Suppose the expected sequence and actual receiving sequence of the i-th data packet are p and p, respectively. i and r i , i represents the index of the data packet, then the calculation expression of the sequence offset is: Δp i =|r i -p i |, Among them, Δp i represents the sequence offset of the ith data packet, After obtaining the order deviation of each data packet, the data packet disorder rate factor of the entire data stream is calculated. The calculation of the data packet disorder rate factor integrates the deviation degree of all data packets and weightedly considers the relative position of the data packets. The specific calculation expression is: ， Among them, PO ri is the packet disorder rate factor, N is the total number of packets, α and μ are weight coefficients that adjust the degree of disorder.

8. The microwave carrier aggregation control and configuration method according to claim 5, characterized in that: When a mismatch risk is detected in the switching timing, the buffer settings of the transmitter and receiver are dynamically adjusted according to the predicted severity. The specific steps are as follows: When a switching timing mismatch risk is detected, the amplitude of the buffer adjustment is first determined. In order to ensure the continuity of the data flow, the sender buffer compensation amount and the receiver buffer extension amount are dynamically adjusted according to the size of the risk coefficient, so that they can fill the short data flow interruption caused by the switching timing mismatch. The adjustment amplitude matches the switching timing mismatch risk coefficient and is affected by the risk coefficient reference threshold. The calculation formula is as follows: ， Among them: B tx is the buffer compensation amount at the sender, that is, the amount of data that needs to be additionally buffered before switching to fill the gap during switching. rx The amount of buffer extension at the receiving end, that is, the buffered data time needs to be extended after switching to absorb the data misalignment problem caused by switching delay. risk is the risk factor of switch timing mismatch, calculated by the machine learning model. th is the reference threshold of the risk factor, γ1 and γ2 are buffer adjustment coefficients, reflecting the sensitivity of the system to the switching timing mismatch, and β1 and β2 are buffer nonlinear adjustment indexes, which are used to enhance the compensation capability for timing mismatch.

9. The microwave carrier aggregation control and configuration method according to claim 8, characterized in that: After calculating the sender's buffer compensation amount, the sender will dynamically adjust the buffer strategy and store more data before switching to ensure that the receiver can still process data normally even if the data stream is temporarily lost during the switching process. The optimized sender buffer filling formula is as follows: ， Where: C tx is the optimized sender cache size, is the original sender buffer size, the default value in normal switching, λ1 is the sender buffer increment adjustment coefficient, which determines the increase in buffer size, δ sync is the synchronization error between the new and old carriers, indicating the degree of deviation of the current carrier switching synchronization, and η is the synchronization error attenuation factor, which controls the impact of the synchronization error on the sender buffer.

10. The microwave carrier aggregation control and configuration method according to claim 9, characterized in that: When the receiving end detects the risk of mismatch, it needs to extend the data cache time to absorb the disorder or loss of data packets caused by switching. The calculation formula for the adjusted receiving end data waiting time is as follows: ， Where: T rx is the optimized receiving end data waiting time, is the original receiving end data waiting time, the default value under normal circumstances, λ2 is the receiving end buffer adjustment coefficient, controlling the adjustment amplitude, Δ jitte is the jitter change of the data stream before and after the switch, reflecting the stability of the data stream during the switch. η2 is the jitter attenuation factor, which controls the impact of data jitter on buffer adjustment.

Images