Matrix switching method for 5G communication signals

Abstract

The invention discloses a matrix switching method for 5G communication signals. The method comprises the following steps: collecting and monitoring 5G network signal data in real time; evaluating a 5G network signal, and constructing a sustainable learning prediction model; dynamically adjusting a matrix switching strategy through prediction content provided by the prediction model, and carrying out sustainable optimization on the model; through the model, a switching path is predicted and selected, and signal switching is realized. The invention aims to reduce switching delay of 5G communication signals in a matrix switching system, reduce signal deterioration and improve flexibility of network slice configuration.

Description

Technical Field

[0001] The present invention belongs to the field of signal switching, especially the matrix switching method for signal switching and the signal switching optimization method based on machine learning. Background Art

[0002] 5G has diverse application scenarios nowadays and is inseparable from our lives, such as autonomous driving, industrial Internet, telemedicine, etc., and has relatively high requirements for the reliability and real-time performance of handover.

[0003] A matrix switching system is a device used for signal routing between multiple input and output ports. In a communication system, a matrix switch can achieve flexible distribution of signals, enabling any input signal to be routed to any output port, or simultaneously routed to multiple output ports. With the popularization of 5G technology, the requirements for the transmission speed and quality of communication signals are increasing day by day. When the existing matrix switching system processes 5G communication signals, it also faces multiple challenges such as handover delay and signal deterioration.

[0004] According to the Chinese patent authorization document CN 115866680A, this patent compresses communication baseband data to obtain 5G communication baseband data, modulates the 5G communication baseband data, establishes a bandwidth transmission method based on the coherence bandwidth amount and the coherence time amount, and establishes a transmission characteristic model of 5G communication signals. For the 5G signal transmission carried out by this patent, the problem of signal abnormality during the signal transmission process is not handled, and signal transmission abnormalities and other situations will still occur. Summary of the Invention

[0005] The object of the present invention is to provide a matrix switching system and method to solve problems such as handover delay and signal deterioration and improve stability according to the deficiencies still existing in the above-mentioned 5G signals during the handover process.

[0006] Another object of the present invention is to perform intelligent slicing and machine learning management on this model after solving the signal switching problem, continuously import historical data into this model, and perform sustainable optimization.

[0007] A matrix switching method for 5G communication signals includes the following steps: Collect and monitor 5G network signal data in real time; Evaluate 5G network signals and construct a sustainable learning prediction model; Dynamically adjust the matrix switching strategy according to the prediction content provided by the prediction model and perform sustainable optimization on this model; Predict and select a handover path through this model to achieve signal handover.

[0008] Collect 5G network signal data and monitor various indicators in real time, including signal strength or transmission rate. Preprocess the data, clean unnecessary data, and normalize the data.

[0009] The constructed prediction model for sustainable learning is continuously updated by supplementing with subsequent data.

[0010] The construction of the network prediction model for sustainable learning is based on the following steps: Collect signal data from places such as 5G base stations; extract relevant information from historical data, conduct correlation analysis and importance assessment on it, and select appropriate signals; sort the signals in a time series to form a time series, and based on time series analysis, select an appropriate model to capture the dynamic characteristics of the time series; divide the data set into a training set and a test set, train the model with the training set, optimize the prediction performance, and adjust the performance parameters.

[0011] Furthermore, based on the prediction results, dynamically adjust the matrix switching strategy and select the switching path.

[0012] The specific steps for dynamically adjusting the matrix switching strategy based on the prediction results are as follows: Based on the prediction results, dynamically adjust the matrix switching strategy, conduct real-time monitoring of the signals, and perform prediction analysis through the previously constructed network model; based on the network state prediction, simultaneously evaluate the performance of multiple paths and assign weights to the performance indicators of each path; based on different path selection algorithms, find the path with the minimum delay and the path with the most abundant bandwidth; dynamically adjust the path selection according to the real-time changes of the network state; deploy the selected path strategy to the network to ensure that data transmission is carried out according to the calculated path.

[0013] Furthermore, after the handover, the system should continuously monitor the network performance, collect feedback data to evaluate the effect of the handover strategy, and continuously optimize the prediction model and the handover algorithm to adapt to the dynamic changes of the network.

[0014] Furthermore, update and reference the intelligent slicing management mechanism for matrix switching. This mechanism dynamically adjusts the slicing resource allocation according to the network load and user requirements to ensure that each slice can obtain the best performance. Adopt the intelligent slicing management mechanism to slice, manage, and analyze the data, and solve the network slicing configuration and the handover delay decision between systems.

[0015] Furthermore, based on the handover optimization technology of machine learning and artificial intelligence, adopt advanced machine learning and artificial intelligence methods to intelligently predict and optimize the handover process.

[0016] Finally, after the collected signals are switched by matrix after the above steps are executed, they are output. Since the present invention mainly deals with problems such as abnormal signal transmission, the displayed result should be normal signal transmission between signals, and the signal switching delay and signal deterioration should be reduced or not generated.

[0017] The present invention aims to generate a prediction model for the matrix switching method, perform short-term prediction according to this model, and perform real-time switching paths, intelligent slicing, and updating of the machine learning model. Switch the signal data being processed to reduce the switching delay, improve the signal stability, and improve problems such as abnormalities in the switched signals.

[0018] The above description is only an overview of the technical solution of the present invention. In order to be able to understand the technical means of the present invention more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features, and advantages of the present invention more obvious and understandable, the specific embodiments of the present invention are specifically exemplified below. Brief Description of the Drawings

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0020] Figure 1 It is a flowchart of the present invention.

[0021] Figure 2 It is a flowchart of an embodiment of the present invention. Detailed Embodiments

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0023] The development process of 5G communication signal handover can be summarized into the following stages: In the early stage of 5G handover technology, around 2019, 5G began commercial use. The handover technology was mainly improved and extended based on the handover mechanism of 4G. The 5G network introduced higher frequencies and more complex network architectures, posing higher requirements for handover technology. From 2020 to 2021, as the 5G network was gradually improved, the handover technology began to develop towards a more efficient and stable direction. For example, a handover trigger mechanism based on parameters such as signal strength and carrier-to-interference ratio, as well as a handover prediction algorithm based on the user's moving speed and direction, were introduced. In 2022, 3GPP proposed two key mobility enhancement technologies in Release 16: DAPS (Dual Active Protocol Stack) and CHO (Conditional Handover). However, the 5G signal matrix handover method still follows the previous algorithm, and there are still problems such as handover delay and signal deterioration for signal handover.

[0024] In view of the above problems, the present invention provides an embodiment to solve the problems such as handover delay and signal deterioration of matrix handover.

[0025] According to the appendix Figure 2 A signal receiving module is provided to monitor the current network signal metrics, including signal strength, transmission rate, etc.

[0026] Combined with historical data, a sustainable learning network prediction model is constructed, which can be continuously updated through subsequent data supplementation. For example, using big data technology to analyze the core network performance metrics in the form of time series, and adopting support vector machine and association rule technologies to optimize the prediction process and results, improving the prediction accuracy. The specific steps are as follows: Data collection and preprocessing: Collect data such as signal strength, handover delay, and signal quality from 5G base stations or user equipment, and preprocess the collected data. For the preprocessing, first clean the unnecessary or missing data, handle missing values and outliers to ensure data quality, and normalize the data during the signal handover process.

[0027] The application of a digital signal processor (DSP) can be adopted. In the baseband processing module, the DSP is used to perform complex algorithm operations on digital signals, including signal modulation and demodulation, channel encoding and decoding, etc.

[0028] After preprocessing, feature extraction and selection of the collected signals are required. Extract features such as signal strength change rate, handover delay, and signal quality metrics from historical data; use correlation analysis and feature importance evaluation to select the features that have the greatest impact on signal handover.

[0029] For correlation analysis, the Pearson correlation coefficient is used to measure the linear correlation degree between two continuous variables. Its value range is from -1 to 1. The closer the absolute value is to 1, the stronger the correlation. For feature importance assessment, model-based feature importance is adopted. A random forest model is trained, and the importance of features is evaluated by calculating the average decrease in impurity when the features are split in the tree.

[0030] After feature evaluation, time series analysis is carried out. For the formation of the time series, the processed signals are collected and arranged in chronological order to form a time series.

[0031] Support Vector Machine (SVM) is a powerful supervised learning model, widely used in classification and regression problems. In time series prediction, SVM can effectively make predictions through its good generalization ability and the ability to handle high-dimensional data. The basic principle of the SVM prediction model is the hyperplane and support vectors. SVM finds an optimal hyperplane to separate samples of different classes. This hyperplane can maximize the margin of sample classification, and the support vectors are those sample points on the margin boundary; kernel function, SVM uses the kernel function to map the input space to a high-dimensional feature space to solve the problem of linear inseparability.

[0032] After the time series is formed, the kernel function is selected. Linear kernel: suitable for linearly separable data sets. If the data is linearly separable in the feature space, the linear kernel can be selected; Nonlinear kernel: Polynomial kernel: suitable for cases where the data has a polynomial relationship. Radial Basis Function (RBF) kernel: suitable for most nonlinear problems, which can map the data to a high-dimensional space to make it linearly separable; Other kernel functions: such as Sigmoid kernel, etc., select the appropriate kernel function according to the specific problem.

[0033] After the kernel function is selected, model training is carried out. First, the data set is provided and divided into a training set and a test set.

[0034] The training set is used for model training. The training of the Support Vector Machine (SVM) is an optimization problem to find the optimal hyperplane. Here, the Lagrange multiplier method is adopted to transform the original optimization problem into a Lagrange dual problem, and the optimal solution is found by solving the dual problem. The quadratic programming (QP) algorithm is used to solve the dual problem to obtain the optimal Lagrange multiplier α.

[0035] According to the model evaluation results, the data after training is optimized to improve the accuracy.

[0036] Validation set evaluation: Use the validation set to evaluate the performance of the model, such as indicators like accuracy, recall rate, F1 score, etc.; Parameter adjustment: Adjust the model parameters according to the evaluation results, such as the penalty coefficient C and the kernel function parameter γ, to improve the generalization ability of the model.

[0037] Based on the prediction results, dynamically adjust the matrix switching strategy and select the optimal switching path.

[0038] Specifically, first, based on the monitoring data and prediction results, evaluate the health status and potential problems of the current network, such as congestion, interference, or fault points, and then conduct path evaluation and selection. That is, fully consider factors such as path delay, bandwidth, and stability, and based on the predicted network state, calculate the expected performance of each path and select the path that best meets the current and future requirements.

[0039] First, conduct network state evaluation by continuously monitoring key performance indicators of the network, such as delay, bandwidth, packet loss rate, signal strength, etc.

[0040] Utilize the previously constructed network prediction model to predict the future state of the network, including the possibility of congestion, interference, or fault points.

[0041] After the evaluation, perform path performance calculation, which includes path delay, path bandwidth, and path stability.

[0042] For the calculation of path delay, calculate the expected delay of each path, which refers to the time required for a signal to propagate from the start point to the end point of the path. Considering the impact of the predicted network state on the delay, first, determine the path from the start point to the end point. For each component on the path (such as logic gates, routers, switches, etc.), calculate its delay, and accumulate the delays of all components on the path to obtain the total path delay.

[0043] For the evaluation of path bandwidth, the available bandwidth refers to the maximum data transfer rate that the path can support without causing significant congestion. Evaluate the available bandwidth of each path to ensure that the path can meet the data transfer requirements. The method adopted is the minimum bandwidth method. On a path, the available bandwidth is usually determined by the link with the smallest bandwidth in the path. This is because the data transfer rate cannot exceed the slowest link in the path. The specific calculation is as follows: Path available bandwidth = min(link 1 bandwidth, link 2 bandwidth,..., link n bandwidth) Path available bandwidth = min(link 1 bandwidth, link 2 bandwidth,..., link n bandwidth).

[0044] For the analysis of path stability, analyze the stability of the path, including signal quality, interference level, and fault probability.

[0045] After calculating the path performance, evaluate multiple paths. Specifically, based on the network state prediction, evaluate the performance of multiple paths simultaneously, including delay, bandwidth, and stability.

[0046] And perform weight assignment to assign weights to the performance indicators of each path to reflect their importance in path selection.

[0047] Among them, for the algorithm of path evaluation, the shortest path algorithm, Dijkstra's algorithm, is adopted to find the path with the minimum delay; Dijkstra's algorithm is an algorithm used to find the single-source shortest path in a weighted graph, proposed by the Dutch computer scientist Edsger W. Dijkstra in 1956. This algorithm is applicable to graphs without negative-weight edges and can effectively find the shortest paths from a source point to all other vertices in the graph.

[0048] The bandwidth optimization algorithm, the maximum flow algorithm, is used to find the path with the most abundant bandwidth; the maximum flow algorithm is an important problem in network flow theory, aiming to find the maximum flow from the source point to the sink point. A network flow graph is a directed graph, and each edge has a capacity limit, indicating the maximum flow that the edge can pass through. The goal of the maximum flow problem is to find the maximum flow from the source point to the sink point without violating the capacity limits of the edges.

[0049] The comprehensive evaluation algorithm combines delay, bandwidth, and stability, and uses the weighted scoring method or a machine learning model to select the optimal path.

[0050] The weighted scoring method is a commonly used decision analysis method. By assigning different weights to different decision factors and comprehensively considering these factors to make the best decision, the calculation steps of the weighted scoring method are as follows: Determine the evaluation indicators and identify all the decision factors to be considered. These factors can be any factors that affect the decision result, such as cost, risk, benefit, etc.; Allocate weights: Set weights for each decision factor. The weight represents the importance of the factor to the final decision result. The weight is usually a percentage and is allocated to each factor with a sum of 1; Score: Score each decision factor. The score can be qualitative (such as high, medium, low) or quantitative (such as 1 - 10 points), and the scoring criteria need to be set according to the specific situation; Calculate the weighted score and the final decision: Calculate the weighted score based on the score of each item and the corresponding weight. According to the comprehensive score, select the decision plan with the highest score as the best decision.

[0051] Dynamic path adjustment is divided into real-time adjustment, which dynamically adjusts the path selection according to the real-time changes of the network state to cope with sudden congestion or failures; Predictive adjustment: A strategy for optimizing and adjusting the network path in advance based on network prediction. Its core purpose is to take corresponding measures by anticipating possible problems in the network in advance, so as to avoid potential network failures, congestion, and other adverse situations and ensure the stable and efficient operation of the network.

[0052] To further address the issues of flexibility in network slice configuration, handover latency between systems, and decision-making accuracy, this technical solution introduces an intelligent slice management mechanism and handover optimization techniques based on machine learning and artificial intelligence.

[0053] The intelligent slice management mechanism is an important technology in modern network management, especially in 5G networks. It optimizes the allocation and management of network resources through intelligent means. The core concepts of intelligent slice management are network slicing and intelligent management.

[0054] Network slicing: Network slicing is a technology that divides a physical network into multiple virtual networks. Each slice can be independently configured and managed to meet the needs of different services.

[0055] Intelligent management: By introducing artificial intelligence and machine learning technologies, it realizes dynamic monitoring, optimization, and automated management of network slices.

[0056] The implementation mechanism is as follows: data analysis and prediction: Using big data analysis and machine learning algorithms, it conducts real-time analysis and prediction of network traffic, user behavior, and service requirements, thereby providing a basis for the dynamic adjustment of slices; Automated deployment and optimization: According to the analysis results, it automatically adjusts the resource allocation, topology, and performance parameters of slices to ensure the efficient operation of slices and the stability of services; Fault detection and self-healing: The intelligent system can monitor the operating status of network slices in real time, detect and handle faults in a timely manner, and even perform preventive adjustments before faults occur.

[0057] 5G networks shift from generalized services to personalized and customized services. Representative technologies are network slicing and edge computing. Network slicing provides customized, logically isolated, dedicated end-to-end virtual mobile networks and is the basic technical form for 5G to ensure service availability for vertical industries. Edge computing deploys network functions close to users, making it possible to achieve extremely low latency and local characteristic applications; adding an intelligent management mechanism to the matrix switching method mainly aims to dynamically adjust slice resource allocation according to network load and user requirements to ensure that each slice can obtain the best performance.

[0058] The specific implementation method is as follows: Use the network management system (NMS) and performance management system (PMS) to collect real-time data of network slices and monitor the load conditions and user requirements of each slice in real time; Based on the monitored data, intelligently analyze the demand and allocation of slice resources, and use advanced data analysis and machine learning technologies to optimize the method of network resource management.

[0059] Dynamically adjust the slice resource allocation strategy according to the analysis results to meet the needs of different users; AI can help analyze network data, predict user needs, and dynamically adjust network slice configurations according to the prediction results to achieve intelligent management.

[0060] Continuously optimize slice performance and improve the utilization efficiency of network resources; by real-time sensing network status, including node load, link bandwidth, user needs, etc., it is possible to more accurately predict future network traffic change trends.

[0061] Machine learning is a subfield of artificial intelligence that enables computers to learn and improve from data through algorithms. The core of machine learning lies in enabling computers to automatically identify patterns and regularities in data without explicit programming. Switching optimization technologies based on machine learning and artificial intelligence have important application values in modern communication networks. Especially in 5G networks, these technologies can help improve user experience and network performance.

[0062] Adopt advanced machine learning and artificial intelligence methods in the optimization of prediction models to perform intelligent prediction and optimization on the handover process. Specific measures include: Collect historical handover data and user behavior data, that is, including the behavior records of users switching between different pages, function modules or applications, as well as various operations of users in the application, such as browsing, clicking, searching, purchasing, etc.; Use machine learning algorithms to analyze and model the data. First, import relevant Python libraries, load the dataset, and conduct preliminary data exploration. Perform feature engineering, including feature selection and normalization. Use Keras to build a logistic regression model, train the model, and observe the performance of the model on the training set and validation set. Use the test set to evaluate the generalization ability of the model and calculate metrics and other data; Based on the model, predict future handover requirements and possible problems. By analyzing historical data, we can use various machine learning algorithms to predict users' possible future behaviors, such as page switching requirements. This not only helps optimize the user experience but also can solve possible problems in advance; Optimize the handover strategy and execution plan according to the prediction results. Through the prediction model, we can understand users' handover requirements and possible problems in advance, so as to optimize the handover strategy and execution plan and improve the user experience and system performance; Real-time monitor and adjust the handover process through artificial intelligence technology to ensure the accuracy and efficiency of the handover.

[0063] In the above two methods, during the handover process, there may be multiple different paths available for handover. At this time, the user's current usage scenario can be predicted by combining the user's needs and behaviors, so as to select a path suitable for the current usage scenario from the available handover paths for the user.

[0064] After the prediction model predicts the path, the path selection strategy is implemented. The strategy deployment deploys the selected path strategy to the network to ensure that data transmission is carried out along the optimal path; the performance monitoring continuously monitors the path performance to verify the effectiveness of the path selection strategy.

[0065] Finally, the handover operation is executed to ensure a fast and smooth transition of the signal. After the handover, the system should continuously monitor the network performance, collect feedback data, use it to evaluate the effect of the handover strategy, and continuously optimize the prediction model and handover algorithm to adapt to the dynamic changes of the network.

[0066] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: it is still possible to modify the specific implementation manners of the present invention or make equivalent replacements, and any modification or equivalent replacement that does not deviate from the spirit and scope of the present invention shall be covered by the protection scope of the claims of the present invention.

Claims

1. A matrix switching method for 5G communication signals, characterized in that, It includes the following steps: S01: Collect 5G network signal data and monitor it in real time; S02: Evaluate the 5G network signal and construct a sustainable learning prediction model; S03: Dynamically adjust the matrix switching strategy according to the prediction content provided by the prediction model, and continuously optimize this model; S04: Predict and select the switching path through this model to achieve signal switching.

2. The matrix switching method for 5G communication signals according to claim 1, wherein In step S01, collect 5G network signal data and monitor various indicators in real time, including signal strength or transmission rate, preprocess the data, clean unnecessary data, and normalize the data.

3. A matrix switching method for 5G communication signals according to claim 1, characterized in that The sustainable learning prediction model constructed in step S02 is continuously updated by supplementing subsequent data.

4. A matrix switching method for 5G communication signals according to claim 1 or 3, characterized in that The constructed sustainable learning model collects signal data from 5G base stations; extracts relevant information from historical data, conducts correlation analysis and importance evaluation on it, selects signals; sorts the signals in a time series to form a time series, and based on time series analysis, selects a model to capture the dynamic characteristics of the time series; Divide the data set into a training set and a test set, and use the training set to train the model, optimize the prediction performance, and adjust the performance parameters.

5. A matrix switching method for 5G communication signals according to claim 1, characterized in that In step S03, based on the prediction result, dynamically adjust the matrix switching strategy, evaluate and select the path, and select the switching path.

6. A matrix switching method for 5G communication signals according to claim 1 or 5, characterized in that, Based on the prediction result, dynamically adjust the matrix switching strategy, monitor the signal in real time, and conduct prediction analysis through the previously constructed network model; based on network state prediction, evaluate the performance of multiple paths simultaneously and assign weights to the performance indicators of each path; Based on different path selection algorithms, used to find the path with the minimum delay and the path with the most abundant bandwidth; dynamically adjust the path selection according to the real-time changes of the network state; deploy the selected path strategy to the network to ensure that data transmission is carried out according to the calculated path.

7. A matrix switching method for 5G communication signals according to claim 1 or 5, characterized in that Execute the switching operation, and the system should continuously monitor the network performance, collect feedback data, used to evaluate the effect of the switching strategy, and continuously optimize the prediction model and switching algorithm to adapt to the dynamic changes of the network.

8. A matrix switching method for 5G communication signals according to claim 1, characterized in that, Update the switching decision through reinforcement learning; by learning historical data and network behavior, the system predicts the switching timing and path.

9. A matrix switching method for 5G communication signals according to claim 1 or 5, characterized in that, Adopt an intelligent slicing management mechanism to slice, manage, and analyze data, and solve network slicing configuration and inter-system handover delay decision-making.

10. A matrix switching method for 5G communication signals according to claim 1 or 5, characterized in that Adopt advanced machine learning and artificial intelligence methods to intelligently predict and optimize the handover process.

Images