A method and system for optimizing communication resource scheduling for smart highways

Abstract

The present invention discloses a communication resource scheduling optimization method and system for smart highways, which relates to the field of communication resource data scheduling and processing technology. The method includes: starting a multi-source data acquisition module to perform real-time sensing on the target highway to obtain a high-speed real-time operation data set; determining a traffic scene information set, performing cluster analysis, and formulating an operation resource classification list based on the analysis results; determining multiple communication resource demand dynamic coefficients, performing deep reinforcement learning, and constructing a resource scheduling model; generating a communication resource scheduling allocation strategy; performing scheduling judgment, and feedback-adjusting the communication resource scheduling allocation strategy based on the judgment results to perform intelligent scheduling optimization. The present invention solves the technical problems that the existing technology cannot meet the low-latency communication requirements of smart highway scenarios and the resource utilization efficiency is low, and achieves the technical effect of realizing intelligent scheduling optimization of smart highway communication resources and improving the rationality and effectiveness of resource allocation.

Description

Technical Field

[0001] The present invention relates to the technical field of communication resource data scheduling and processing, and in particular to a method and system for optimizing communication resource scheduling for intelligent high-speed communication. Background Art

[0002] With the rapid development of intelligent transportation systems, smart highways have become a key development direction in the transportation sector. Within smart highway scenarios, large amounts of data transmission are required between vehicles and infrastructure, and between vehicles themselves, to enable functions such as real-time road condition monitoring, automated driving assistance, and emergency information dissemination. This places extremely high demands on the rational allocation of communication resources. However, current smart highway communication resource allocation faces numerous challenges. For one thing, traditional communication resource allocation methods are mostly based on fixed rules or empirical evidence, lacking the ability to dynamically perceive and respond to changes in real-time traffic conditions and communication needs. During peak traffic periods, communication demands between vehicles and between vehicles and infrastructure increase dramatically. Fixed resource allocation strategies are unable to adjust resource allocation in a timely manner, often leading to communication congestion, data transmission delays, and even interruptions, seriously impacting the operational efficiency and service quality of smart highway systems. Furthermore, the actual demand for communication resources varies significantly across different traffic scenarios, but existing technologies are unable to effectively distinguish and accurately adapt them. On lightly trafficked sections, communication resources may be overallocated, resulting in idle and wasted resources. On the other hand, on heavily trafficked or accident-prone sections, insufficient resource allocation fails to meet data transmission needs.

[0003] Existing technologies cannot meet the low-latency communication requirements of smart high-speed scenarios and have technical problems such as low resource utilization efficiency. Summary of the Invention

[0004] The present application provides a communication resource scheduling optimization method and system for smart highways, which is used to solve the technical problems in the existing technology that the low-latency communication requirements of smart highway scenarios cannot be met and the resource utilization efficiency is low.

[0005] In view of the above problems, the present application provides a method and system for optimizing communication resource scheduling for smart highways.

[0006] A first aspect of the present application provides a method for optimizing communication resource scheduling for smart highways, the method comprising:

[0007] Start the multi-source data acquisition module to perform real-time sensing on the target highway to obtain a high-speed real-time operation data set; perform multi-scenario analysis on the target highway based on the high-speed real-time operation data set to determine a traffic scenario information set, perform cluster analysis on the high-speed real-time operation data set according to the traffic scenario information set, and formulate an operation resource classification list based on the analysis results; traverse the operation resource classification list to perform dynamic calculation of communication resource requirements, determine multiple communication resource requirement dynamic coefficients, perform deep reinforcement learning based on the multiple communication resource requirement dynamic coefficients, and construct a resource scheduling model; synchronize the high-speed real-time operation data set to the resource scheduling model for resource scheduling, and generate a communication resource scheduling allocation strategy; simulate the execution of the communication resource scheduling allocation strategy to perform scheduling judgment, feedback and adjust the communication resource scheduling allocation strategy according to the judgment result, and generate the communication resource scheduling allocation optimization strategy to perform intelligent scheduling optimization of the communication resources of the smart highway.

[0008] A second aspect of the present application provides a communication resource scheduling optimization system for smart highways, the system comprising:

[0009] A high-speed real-time operation data set acquisition module is used to start the multi-source data acquisition module to perform real-time sensing on the target highway and obtain a high-speed real-time operation data set; an operation resource classification list formulation module is used to perform multi-scenario analysis on the target highway based on the high-speed real-time operation data set, determine a traffic scenario information set, perform cluster analysis on the high-speed real-time operation data set according to the traffic scenario information set, and formulate an operation resource classification list based on the analysis results; a resource scheduling model construction module is used to traverse the operation resource classification list to dynamically calculate communication resource requirements, determine multiple communication resource requirement dynamic coefficients, perform deep reinforcement learning based on the multiple communication resource requirement dynamic coefficients, and construct a resource scheduling model; a communication resource scheduling allocation strategy generation module is used to synchronize the high-speed real-time operation data set to the resource scheduling model for resource scheduling and generate a communication resource scheduling allocation strategy; an intelligent scheduling optimization module is used to simulate the execution of the communication resource scheduling allocation strategy to make a scheduling decision, feedback and adjust the communication resource scheduling allocation strategy based on the decision result, and generate the communication resource scheduling allocation optimization strategy to perform intelligent scheduling optimization of the communication resources of the smart highway.

[0010] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0011] The multi-source data acquisition module is activated to perform real-time sensing on the target highway to obtain a real-time high-speed operation data set; a multi-scenario analysis is performed on the target highway to determine the traffic scenario information set, cluster analysis is performed, and an operation resource classification list is formulated based on the analysis results; the operation resource classification list is traversed to dynamically calculate communication resource requirements, determine multiple communication resource demand dynamic coefficients, and construct a resource scheduling model; the real-time high-speed operation data set is synchronized with the resource scheduling model to perform resource scheduling and generate a communication resource scheduling allocation strategy; the communication resource scheduling allocation strategy is feedback-adjusted based on the judgment results to generate the communication resource scheduling allocation optimization strategy to intelligently schedule and optimize the communication resources of the smart highway. This achieves the technical effect of realizing intelligent scheduling optimization of smart highway communication resources and improving the rationality and effectiveness of resource allocation. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0013] Figure 1 A flow chart of a method for optimizing the scheduling of communication resources for smart high-speed communication provided in an embodiment of the present application.

[0014] Figure 2 A schematic diagram of the structure of a communication resource scheduling optimization system for smart high-speed communication provided in an embodiment of the present application.

[0015] Explanation of the reference numerals: high-speed real-time operation data set acquisition module 10, operation resource classification list formulation module 20, resource scheduling model construction module 30, communication resource scheduling allocation strategy generation module 40, intelligent scheduling optimization module 50. DETAILED DESCRIPTION

[0016] This application provides a communication resource scheduling optimization method and system for smart highways, which is used to solve the technical problems in the existing technology that the low-latency communication requirements of smart highway scenarios cannot be met and the resource utilization efficiency is low.

[0017] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative work are within the scope of protection of this application.

[0018] Example 1, as Figure 1 As shown, the present application provides a method for optimizing communication resource scheduling for smart highways, the method comprising:

[0019] Step S100: Start the multi-source data acquisition module to perform real-time sensing on the target highway to obtain a high-speed real-time operation data set.

[0020] Specifically, in the construction of smart highways, a comprehensive, multi-layered highway data collection network has been established by activating a multi-source data collection module. This module integrates a variety of advanced sensor technologies and data collection methods. For vehicle density and speed monitoring, geomagnetic sensors and radar sensors deployed on different sections of the highway can sense vehicle traffic in real time, accurately calculate vehicle density, and precisely measure vehicle speeds. Meteorological conditions are monitored using weather stations located along the highway. These stations collect real-time meteorological information such as temperature, humidity, rainfall, wind speed, and visibility. This data is crucial for assessing the impact of weather on highway driving safety and communication stability. Furthermore, traffic flow is captured through network monitoring equipment, which monitors traffic flow at communication base stations along the highway in real time, recording traffic data for various services, including voice calls and data transmission. Through the collaborative work of these different types of sensors and monitoring equipment, the multi-source data acquisition module continuously performs real-time sensing of the target highway, integrating, preliminarily processing and summarizing the massive real-time data collected, such as vehicle density, driving speed, weather conditions, business flow, etc., and ultimately forming a high-speed real-time operation data set, providing comprehensive and accurate data support for subsequent in-depth analysis of the highway and the rational scheduling of communication resources.

[0021] Step S200: Perform multi-scenario analysis on the target highway based on the high-speed real-time operation data set, determine a traffic scenario information set, perform cluster analysis on the high-speed real-time operation data set according to the traffic scenario information set, and formulate an operation resource classification list based on the analysis results.

[0022] Specifically, a multi-dimensional scenario analysis is conducted on the target highway using a real-time highway operation dataset. Information contained in the dataset, such as vehicle density, driving speed, weather conditions, and traffic volume, is deeply mined. For example, when vehicle density exceeds a certain threshold and driving speed remains consistently low, the road section is considered congested. Extremely low visibility in the weather conditions indicates a severe weather scenario. This approach identifies a set of traffic scenario information encompassing different conditions, such as congestion, smooth traffic, accidents, and severe weather. Subsequently, cluster analysis is performed on the real-time highway operation dataset based on this identified traffic scenario information set. Using key data such as vehicle density, driving speed, weather conditions, and traffic volume as clustering features, data points with similar characteristics are grouped together, forming multiple traffic operation cluster centers and, in turn, identifying multiple traffic operation clusters. For example, data from congested scenarios are grouped into one cluster, while data from smooth traffic scenarios are grouped into another. To develop a categorized list of operational resources, the historical operational resource demand dataset for the target highway is first retrieved and labels assigned to the different traffic operation clusters based on this dataset. Next, we analyze each cluster's resource requirements and determine resource requirement metrics, such as communication bandwidth, computing resources, and storage capacity. We prioritize resource requirements within each cluster based on their urgency and importance. Finally, by combining multiple cluster resource requirement metrics and their corresponding priorities, we construct a categorized list of operational resources, laying the foundation for subsequent accurate calculation of communication resource requirements and optimized resource scheduling.

[0023] Step S300: traverse the running resource classification list to perform dynamic calculation of communication resource requirements, determine multiple communication resource requirement dynamic coefficients, perform deep reinforcement learning according to the multiple communication resource requirement dynamic coefficients, and build a resource scheduling model.

[0024] Specifically, traversing the operational resource classification list is the primary task for subsequent work. This list, compiled based on different highway traffic scenarios and corresponding resource demand characteristics, encompasses a wealth of operational resource information. Each entry is associated with a specific operational point, which can be a different section of the highway, a key transportation hub, or a communication base station location. During this traversal process, communication resource demand is dynamically calculated for each operational point, taking into account multiple factors. Reference is made to historical operational resource demand datasets, incorporating information such as the current operational point's traffic scenario and resource demand priority to determine its basic communication demand. Furthermore, given the ever-changing traffic conditions, real-time communication demand data for each operational point is further calculated based on real-time data from multiple traffic operation clusters and the determined basic communication demand. To accurately reflect the dynamic relationship between communication resource demand and various influencing factors, a multivariate regression analysis is performed on this real-time communication demand data, resulting in multiple dynamic communication resource demand coefficients corresponding to each operational point. These coefficients not only consider traffic scenarios and resource priorities, but also incorporate factors such as real-time traffic changes, accurately reflecting the changing patterns of communication resource demand at different operational points at different times. Next, a resource scheduling model is constructed using the obtained dynamic coefficients of communication resource demands. Based on these coefficients, a comprehensive analysis of the communication nodes associated with each operating point is performed to plan multiple resource allocation paths that meet the requirements of different operating points. Furthermore, a resource scheduling state space is constructed based on the dynamic coefficients, accurately describing the resource status of each operating point, taking into account multiple dimensions such as resource availability, allocation, and utilization efficiency. Furthermore, based on the planned resource allocation paths, the resource allocation actions that can be taken for each operating point under different paths are analyzed to construct a resource scheduling action space. To enable the model to achieve optimal resource scheduling in the complex and ever-changing highway communication environment, a reward function is introduced. This reward function is designed around the goals of rationally utilizing highway communication resources and meeting the communication requirements of different operating points. By repeatedly training the resource scheduling state space and the resource scheduling action space, the model continuously tries different resource allocation strategies and adjusts its decisions in a timely manner based on the feedback from the reward function. As training progresses, the model gradually masters the ability to select the optimal resource allocation action under different operating point conditions, ultimately successfully constructing a resource scheduling model that can adapt to the complex and dynamic environment of highways.

[0025] Step S400: Synchronizing the high-speed real-time running data set to the resource scheduling model to perform resource scheduling and generate a communication resource scheduling allocation strategy.

[0026] Specifically, the high-speed real-time operation dataset is synchronized with the resource scheduling model to generate a communication resource scheduling and allocation strategy. First, the high-speed real-time operation dataset contains a large amount of complex information, covering multiple sources such as traffic flow, vehicle speeds, road conditions, and weather conditions, with varying data formats and characteristics. Data is preprocessed to unify the data format. For example, the format of traffic data collected by different sensors is standardized. Normalization and denoising techniques are used to scale data features to appropriate ranges, remove outliers and noise, and improve data quality. Next, the data is adapted according to the input requirements of the resource scheduling model. If the model receives data in vector form, the preprocessed data is converted to a vector of a specific dimension. If the model uses image input, the data is converted to the appropriate image format to ensure efficient reading and processing by the model. The preprocessed and adapted high-speed real-time operation dataset is then input into the resource scheduling model. This model, based on deep reinforcement learning, employs a deep neural network, with input data passing through an input layer, multiple hidden layers, and an output layer. In the hidden layer, data undergoes weighted calculations and nonlinear transformations by various neurons to uncover underlying patterns and associations, such as the nonlinear relationship between traffic flow and communication demand. Based on the model's established algorithm and incorporating parameters such as the dynamic coefficient of communication resource demand, the input data is processed to determine how communication resources should be allocated under different traffic scenarios. The output of the model's computational processing is a communication resource scheduling and allocation strategy. This strategy specifies key information such as the type, quantity, and priority of resources to be allocated to each communication node under current high-speed, real-time operating conditions. For example, for communication nodes on sections with heavy traffic and high accident risk, the strategy allocates more bandwidth and a higher priority; for nodes on sections with stable traffic flow, an appropriate amount of resources is allocated to maintain basic communication needs. The output strategy can be a straightforward resource allocation list detailing resource allocation details for each communication node, or a set of instructions containing allocation rules and parameters, which are used to execute resource allocation operations in actual scheduling.

[0027] Step S500: simulate and execute the communication resource scheduling and allocation strategy to make a scheduling judgment, perform feedback adjustment on the communication resource scheduling and allocation strategy according to the judgment result, and generate the communication resource scheduling and allocation optimization strategy to perform intelligent scheduling optimization on the communication resources of the smart highway.

[0028] Specifically, to ensure the effectiveness and rationality of the communication resource scheduling and allocation strategy, a simulated communication environment highly similar to the real environment is first constructed. Within this simulated environment, the generated communication resource scheduling and allocation strategy is rigorously simulated, encompassing various communication nodes and diverse traffic scenarios, to obtain detailed scheduling and allocation results. The scheduling and allocation results are then comprehensively analyzed across multiple communication nodes. Based on the actual requirements and performance standards of the smart highway communication system, an expected scheduling threshold is pre-set. This threshold serves as a key criterion for measuring whether the scheduling and allocation results meet the standards, encompassing key indicators such as communication bandwidth utilization, signal stability, and data transmission latency. The scheduling and allocation results are carefully compared with the expected scheduling threshold to determine whether they meet the scheduling requirements. If the scheduling and allocation results meet the expected scheduling threshold, indicating that the current strategy performs well in the simulated environment, a positive feedback instruction is generated. Based on this instruction, the communication resource scheduling and allocation strategy is continuously monitored, closely monitoring its performance over time as the simulated environment changes and traffic scenarios dynamically shift. A judgment result is then generated, confirming that the strategy is suitable for practical application as a basis for optimization. If the scheduling allocation result does not meet the expected scheduling threshold, it means that the current strategy has certain defects. At this time, a negative feedback instruction is generated, and the communication resource scheduling allocation strategy is comprehensively and deeply traced according to the instruction. By carefully checking each link in the simulation execution process, the unreasonable resource allocation under different traffic scenarios and communication nodes is analyzed, thereby generating multiple abnormal allocation information. Based on these abnormal allocation information, combined with the high-speed real-time operation data set and the relevant parameters of the resource scheduling model, the original strategy is adjusted with targeted feedback. By increasing or reducing the communication resource allocation in certain areas, adjusting the priority of resource allocation, etc., the communication resource scheduling allocation optimization strategy is regenerated. This optimization strategy fully takes into account the problems exposed in the simulation execution, can better adapt to the complex and changeable communication needs of smart highways, realize the intelligent scheduling optimization of communication resources, and ensure the efficient and stable operation of the smart highway communication system.

[0029] In one possible implementation, step S200 further includes:

[0030] Step S210: traverse the high-speed real-time running data set to extract scene features and determine multiple scene feature values.

[0031] Step S220: performing feature dimensionality reduction on the plurality of scene feature values ​​using principal component analysis to determine a plurality of scene features.

[0032] Step S230: performing scene analysis on the target highway based on the multiple scene features to determine a traffic scene information set.

[0033] Step S240: performing cluster analysis on the high-speed real-time operation data set according to the traffic scene information set to obtain a plurality of traffic operation cluster centers.

[0034] Step S250: Divide the high-speed real-time operation data set according to the multiple traffic operation cluster centers to determine multiple traffic operation clusters.

[0035] Step S260: Add the multiple traffic operation clusters to the analysis result.

[0036] Specifically, a traversal operation is performed on the real-time highway data set to extract scenario characteristics. Leveraging sensor and data processing technologies, the rich real-time information contained in the data set is then analyzed for targeted analysis. For meteorological data, meteorological characteristics such as temperature, humidity, and wind speed are directly extracted using data collected by meteorological monitoring equipment deployed along the highway. These devices accurately measure and transmit meteorological information in real time. For traffic data, geomagnetic sensors, cameras, and other traffic monitoring equipment collect data such as the number of vehicles passing through and the time interval between vehicles to calculate vehicle density. Average speed is also calculated based on vehicle trajectories and the time difference between passing vehicles. Temporal characteristics are extracted based on pre-defined time periods, such as weekdays and weekends, and combined with historical traffic flow data patterns to determine whether the current time period falls within peak or off-peak traffic. Furthermore, for traffic flow data, network monitoring equipment monitors the traffic volume of various types of traffic at highway communication base stations in real time, extracting scenario-related traffic characteristics. This data, acquired and processed from various data sources, is integrated to determine multiple scenario characteristics, providing comprehensive and critical foundational information for subsequent scenario analysis and data processing.

[0037] Covariance calculation is performed based on the multiple scene eigenvalues ​​obtained. Covariance can measure the linear correlation between different eigenvalues. By calculating the covariance between all eigenvalues, a feature data matrix is ​​constructed. This matrix fully presents the degree of correlation between the eigenvalues ​​of each scene. Next, the multiple scene eigenvalues ​​in the feature data matrix are sorted in descending order according to the variance of each eigenvalue. Eigenvalues ​​with larger variances mean that they contain rich information and have a greater impact on the variability of the data. After sorting, the first few principal components with larger variances are selected to determine the principal component basis vector matrix. These principal component basis vectors represent the most important direction of change in the original data and can maximize the retention of key information in the data. Subsequently, the multiple scene eigenvalues ​​are projected onto the principal component basis vector matrix. The projection operation can map the original high-dimensional data into the low-dimensional space generated by the principal component basis vectors, achieving data dimensionality reduction transformation and thus obtaining the feature dimensionality reduction matrix. Finally, to ensure that the reduced features accurately reflect the characteristics of the original data, dimensionality reduction verification is performed based on the feature dimensionality reduction matrix. The verification process is conducted through various methods, such as comparing the similarities between the data before and after dimensionality reduction and analyzing the effectiveness of the reduced data in subsequent analysis tasks. If the verification results meet expectations, multiple scene features are identified from the reduced data. These scene features include environmental characteristics such as temperature, humidity, and wind speed in meteorological conditions; traffic characteristics including vehicle density and average speed; and temporal characteristics such as peak and off-peak traffic hours. These scene features, identified through principal component analysis and dimensionality reduction, remove redundant information while highlighting key information, providing strong support for subsequent accurate scene analysis of the target highway.

[0038] A comprehensive and detailed scenario analysis is conducted on the target expressway to determine the traffic scenario information set. First, two key traffic characteristics, vehicle density and average speed, are analyzed in depth. By studying historical and real-time monitoring data, different threshold ranges for vehicle density and average speed are set to classify different traffic scenario types. When vehicle density is high and average speed is low, the road section is considered to be in a state of traffic congestion. For example, if a road section has a vehicle density consistently exceeding 200 vehicles per kilometer and an average speed below 30 kilometers per hour for a period of time, it is classified as a high-density, low-speed traffic congestion scenario. When vehicle density is moderate and average speed is high, it is considered a smooth flow scenario. When vehicle density is low and average speed is high, the road section is considered a sparse flow scenario. In addition to traffic characteristics, the impact of environmental and temporal characteristics on traffic scenarios is comprehensively considered. Under adverse weather conditions, such as heavy rain or strong winds, even if vehicle density and average speed are within the smooth or sparse flow ranges, environmental factors may lead to unstable traffic conditions. These situations are considered as special traffic scenarios. Furthermore, even if the vehicle density and average speed on certain sections of a highway temporarily meet the criteria for smooth flow during peak traffic hours, the traffic scenario will be adjusted accordingly to account for the potential congestion risks caused by this time period. Through comprehensive analysis and assessment of multiple scenario characteristics, a traffic scenario information set is determined. This information set clearly presents the traffic conditions on different sections of the target highway at different times, providing a key basis for subsequent cluster analysis of the highway's real-time operation dataset and the development of a categorized list of operational resources, thereby enabling more precise optimization of communication resource scheduling.

[0039] Cluster analysis is performed on high-speed real-time operation datasets. Using clustering algorithms, data points with similar characteristics are grouped together to obtain multiple traffic operation cluster centers. These cluster centers represent the typical characteristics of data in different traffic scenarios.

[0040] The high-speed real-time operation dataset is divided into multiple traffic operation clusters based on multiple traffic operation cluster centers. Each traffic operation cluster corresponds to a specific traffic scenario, and the data within the cluster have similar characteristics. For example, data belonging to the same congestion scenario are divided into the same cluster.

[0041] Multiple traffic operation clusters are added to the analysis results. These traffic operation clusters become an important basis for the subsequent development of a classification list of operation resources, laying the foundation for in-depth analysis of operation resource requirements in different scenarios and achieving precise communication resource scheduling optimization.

[0042] In one possible implementation, step S220 further includes:

[0043] Step S221: performing covariance calculation based on the multiple scene feature values ​​to construct a feature data matrix.

[0044] Step S222: Arrange the plurality of scene eigenvalues ​​in descending order according to the feature data matrix to determine a principal component basis vector matrix.

[0045] Step S223: Projecting the multiple scene feature values ​​onto the principal component basis vector matrix to obtain a feature dimension reduction matrix.

[0046] Step S224: performing dimensionality reduction verification according to the feature dimensionality reduction matrix, and obtaining the multiple scene features according to the verification result.

[0047] Specifically, covariance calculations are performed on multiple scenario eigenvalues. Scenario eigenvalues ​​encompass various data points, including weather conditions, traffic conditions, and time factors. The covariance between any two eigenvalues ​​is calculated to measure their linear correlation. For example, calculating the covariance between temperature and vehicle density can reflect the correlation between their changes. By performing covariance calculations on all pairwise eigenvalues, a characteristic data matrix is ​​constructed. This matrix comprehensively presents the interrelationships between the eigenvalues ​​of each scenario, laying the foundation for subsequent analysis.

[0048] Based on the feature data matrix, the eigenvalues ​​of multiple scenarios are sorted in descending order according to their variance. Eigenvalues ​​with larger variances contain richer information and have a more significant impact on data variance. The first few principal components with larger variances are selected, and the basis vectors corresponding to these principal components form the principal component basis vector matrix. This method can screen out the directions that are most influential in data variation.

[0049] Project multiple scene eigenvalues ​​onto a matrix of principal component basis vectors. Projection maps data from a high-dimensional space to a low-dimensional space composed of principal component basis vectors. This process reduces the dimensionality of the data and produces a feature-reduced matrix. For example, high-dimensional data containing multiple complex scene eigenvalues ​​can be transformed into data with significantly reduced dimensionality after projection, while preserving as much key information as possible.

[0050] The feature dimensionality reduction matrix is ​​validated using various methods, such as comparing the effectiveness of the pre- and post-dimensionality reduction data in subsequent scenario analysis tasks and calculating the similarity between the pre- and post-dimensionality reduction data. If the validation results indicate that the reduced data accurately reflects the characteristics of the original data and meets the analysis requirements, multiple scenario features are determined based on the validation results. These features remove redundant information while retaining the core information of the original scenario feature values, providing concise and effective data support for subsequent highway scenario analysis.

[0051] In one possible implementation, step S200 further includes:

[0052] Step S270: Retrieve the historical operation resource demand data set of the target expressway, traverse the historical operation resource demand data set to assign labels to the multiple traffic operation clusters, and determine multiple cluster labels; perform demand analysis on the multiple traffic operation clusters based on the multiple cluster labels, and determine multiple intra-cluster resource demand indicators; classify the multiple intra-cluster resource demand indicators according to demand quantity, and determine multiple demand priorities; arrange the multiple intra-cluster resource demand indicators according to the multiple demand priorities and construct the operation resource classification list.

[0053] Specifically, when constructing the operational resource classification list, the first step is to retrieve the historical operational resource demand dataset for the target expressway. This dataset records the past demands for various resources for the expressway under different operational conditions and is of great reference value. By traversing this dataset, labels are assigned to the multiple traffic operation clusters previously obtained based on the different traffic operation conditions reflected in the data. For example, if the data within a traffic operation cluster corresponds to traffic congestion and high business volume, it is marked as a high-demand congestion class, thereby determining multiple cluster labels. These labels become key identifiers for subsequent analysis of the resource requirements of each traffic operation cluster.

[0054] Based on the multiple cluster labels identified, demand analysis is performed for each traffic operation cluster. For each traffic operation cluster, the demand for various resources within the cluster, such as communication bandwidth, computing resources, and storage capacity, is analyzed based on the scenario characteristics represented by its label. This analysis then determines multiple resource demand indicators within the cluster. For example, for a traffic operation cluster labeled as high-demand congestion, the demand for communication bandwidth is high due to the frequent information exchange between vehicles and the vehicle control center during congestion. Furthermore, to process large amounts of traffic data in a timely manner, corresponding demand indicators for computing resources and storage capacity are also present.

[0055] Multiple resource demand indicators within a cluster are categorized based on their demand size. High-demand indicators are assigned high priority, moderate-demand indicators are assigned medium priority, and low-demand indicators are assigned low priority, thereby determining multiple demand priorities. For example, in scenarios with heavy traffic and high business volume, the urgent need for communication bandwidth is assigned high priority, while storage resource requirements that do not require real-time data are assigned medium or low priority.

[0056] When constructing the operational resource classification list, the historical operational resource demand dataset for the target expressway is retrieved. This dataset is then traversed to assign labels to the multiple previously identified traffic operation clusters, resulting in multiple cluster labels. Based on these cluster labels, an in-depth demand analysis is conducted for each traffic operation cluster, determining resource demand indicators for different resource types within each cluster, such as bandwidth, communication terminals, and manual intervention resources. These indicators reflect the specific resource requirements of each operation cluster, such as the specific bandwidth demand of a cluster during peak traffic hours or the number of manual intervention resources required under special road conditions. Next, the resource demand indicators within the multiple clusters are categorized by demand, and multiple demand priorities are determined. Resource demands with high demand and critical to expressway operation are given higher priority; conversely, resource demands with relatively low demand and limited impact on overall operation are given lower priority. Finally, the cluster name, corresponding resource type, resource priority, and allocated quantity are integrated and sorted. Resources are sorted from high to low priority. Within the same priority level, resource types are further subdivided and sorted by resource type, such as bandwidth resources first, communication terminal resources second, and manual intervention resources last. Different operation clusters of the same resource type can be arranged alphabetically by cluster name or according to their geographic location on the highway. This approach creates a comprehensive and clear operational resource classification list, providing an important foundation for subsequent communication resource demand analysis and scheduling.

[0057] In one possible implementation, step S300 further includes:

[0058] Step S310: performing communication analysis on the target highway according to the historical operation resource demand data set to determine a plurality of communication nodes.

[0059] Step S320: performing communication demand analysis based on the operating resource classification list and the plurality of communication nodes, determining a plurality of basic communication demands, and extracting the plurality of demand priorities.

[0060] Step S330: performing real-time calculations based on the multiple traffic operation clusters and the multiple basic communication demands to obtain multiple real-time communication demand data.

[0061] Step S340: Performing a multiple regression analysis based on the plurality of real-time communication demand data to obtain the plurality of communication resource demand dynamic coefficients.

[0062] Specifically, a comprehensive and in-depth analysis of the historical operational resource demand dataset was conducted. First, all communication-related data records within the dataset were screened, including information on data transmission volume, connection duration, and stability at different times for each road section. Next, potential communication nodes were identified based on the distribution of data transmission volume. Sections with significantly higher-than-average data transmission volume are likely to contain critical communication nodes, as they host a significant amount of data exchange. Furthermore, attention was paid to communication connection stability indicators. Frequent interruptions or delays in a particular area indicate that this area may be a critical node with a high concentration of communication traffic, potentially experiencing communication quality issues due to excessive load. Furthermore, combined with existing information on the layout of highway infrastructure, such as the established locations of base stations and switches, the activity of these facilities as reflected in historical communication data was comprehensively assessed. Through comprehensive consideration and analysis of these various aspects of information, multiple communication nodes distributed along the target highway were identified and determined. These nodes are essential for subsequent accurate communication resource demand analysis and the construction of a scheduling model.

[0063] The operational resource classification list is combined with the identified multiple communication nodes to conduct a communication demand analysis. The operational resource classification list details the resource demand characteristics and priorities of different traffic operation clusters. Through this combined analysis, multiple basic communication requirements are determined for each communication node under different traffic operation scenarios. For example, for communication nodes corresponding to busy road sections, a higher basic communication requirement is determined, taking into account the frequent data exchange between vehicles and between vehicles and the control center. At the same time, multiple demand priorities corresponding to each traffic operation cluster are extracted from the operational resource classification list to provide a basis for subsequent resource allocation priority decisions.

[0064] Based on multiple traffic operation clusters, real-time calculation is carried out in combination with multiple communication infrastructure requirements. = · (t), operates on each traffic operation cluster. First, Indicates the multiple previously determined basic communication requirements. This is a fixed value based on the operation resource classification list and communication node analysis. It reflects the basic communication resource requirements under different traffic operation scenarios. (t) is a dynamic adjustment coefficient. By monitoring the status of each traffic cluster in real time, for example, by calculating the current traffic flow trend through time series prediction, if the traffic flow of a certain traffic cluster shows a rapid upward trend in a certain period of time, this will affect (t) value; At the same time, the current network load status will also be considered. If the network load in the area where the cluster is located is close to saturation, then (t) will also be adjusted accordingly; in addition, the performance of the communication node is also an important factor. If the performance of a communication node degrades, it will also cause (t) changes. Finally, and the calculated (t) multiplied, we get , i.e., multiple real-time communication demand data. This data can accurately reflect the actual communication resource requirements of each traffic operation cluster at the current moment, providing a key basis for subsequent analysis and scheduling, ensuring that highway communication resources can be reasonably allocated based on real-time traffic conditions.

[0065] To accurately quantify the dynamic relationship between communication resource demand and various influencing factors, a multivariate regression analysis was conducted on multiple real-time communication demand data sets. This multivariate regression analysis comprehensively considers multiple variables, such as changes in traffic clusters, basic communication demand, and demand priorities. By establishing a mathematical model, it determined multiple dynamic coefficients for communication resource demand. These coefficients comprehensively reflect the degree to which different factors influence communication resource demand, providing core parameters for constructing a resource scheduling model, enabling the model to more accurately schedule communication resources based on real-time traffic scenarios and demand changes.

[0066] In one possible implementation, step S300 further includes:

[0067] Step S350: traverse the plurality of communication nodes based on the plurality of communication resource demand dynamic coefficients, and plan a plurality of resource allocation paths.

[0068] Step S360: Perform resource status analysis according to the multiple communication resource demand dynamic coefficients to construct a resource scheduling state space.

[0069] Step S370: Perform resource action analysis according to the multiple resource allocation paths to construct a resource scheduling action space.

[0070] Step S380: introducing a reward function, interactively training the resource scheduling state space and the resource scheduling action space, and constructing the resource scheduling model.

[0071] Specifically, multiple communication nodes are traversed based on multiple dynamic coefficients of communication resource demand. These coefficients comprehensively reflect the impact of different traffic scenarios, resource demand priorities, and other factors on communication resource demand. During this traversal, multiple resource allocation paths are planned based on the location and function of each communication node, the traffic conditions in the area it serves, and the dynamic coefficients of communication resource demand. For example, for communication nodes located on busy roads, where communication resource needs are more urgent, allocation paths are planned to prioritize their resource supply, ensuring smooth communication in that area.

[0072] Resource status analysis is performed based on multiple dynamic coefficients of communication resource demand. The status of communication resources is assessed across multiple dimensions, including availability, allocation, and utilization efficiency. By analyzing the real-time status of communication resources under different traffic scenarios and demand priorities, a resource scheduling state space is constructed. This state space comprehensively describes the various possible states of communication resources at different times and in different scenarios, providing a clear context for subsequent decision-making. For example, during peak traffic hours, the resources of certain communication nodes may be nearing saturation, and this state is incorporated into the resource scheduling state space.

[0073] For each resource allocation path, different types of resource actions are considered. For bandwidth resources, these may include increasing bandwidth, decreasing bandwidth, and reallocating bandwidth. For communication terminal resources, this involves increasing or decreasing the deployment of terminal devices or adjusting terminal operating parameters. For manual intervention resources, these include actions such as adding personnel and reallocating personnel responsibilities. These actions are determined based on a deep understanding of the characteristics of different resource types and a comprehensive consideration of the highway operation scenario. Secondly, the feasibility and potential impact of each action are evaluated based on historical data and current traffic operation cluster information. By analyzing the effects of different actions taken under similar resource allocation paths in historical data, as well as the traffic conditions and resource demand indicators presented by the current traffic operation cluster, it is determined whether executing a particular resource action can meet the needs of the traffic scenario. For example, if historical data shows that increasing bandwidth resources on a certain road section in a certain traffic congestion scenario can effectively alleviate data transmission pressure, and the road section is currently congested, the action of increasing bandwidth is included in the list of possible actions for this resource allocation path. Furthermore, resource actions are quantitatively analyzed, and for each resource action, its possible operation range and adjustment range are determined. Taking bandwidth resources as an example, based on factors such as the current total available bandwidth and the priority of traffic clusters, we determine the range within which bandwidth can be increased or decreased, as well as the impact of different adjustments on system performance and traffic communication. This allows for a more precise description of resource actions, providing a more specific basis for subsequent resource scheduling. Finally, the various resource actions under the different resource allocation paths identified above are integrated to form the resource scheduling action space.

[0074] A reward function is introduced. When interactively training the resource scheduling state space and the resource scheduling action space, the model selects a resource scheduling action from the action space based on the current resource scheduling state and applies it in a real-world simulation environment. Subsequently, based on the results of executing this action, the reward function evaluates how much it improves or deteriorates the resource scheduling state. For example, if a resource scheduling action significantly improves communication efficiency on a congested road section, reduces data transmission latency, and fully utilizes existing communication resources, the reward function will assign a higher reward. Conversely, if the action leads to uneven resource allocation or poor communication performance, such as overallocation of bandwidth in some areas and insufficient bandwidth in others, the reward function will assign a lower reward or even a penalty. During training, the model continuously extracts different states from the resource scheduling state space, tries various actions in the resource scheduling action space, and adjusts its decisions based on the feedback from the reward function. This training is an iterative process. Through numerous simulations, the model gradually learns which resource scheduling actions, under different resource scheduling states, yield higher rewards, contributing to optimal resource allocation and meeting communication needs. As training progresses, the model continuously optimizes its decision-making mechanism based on resource scheduling status and reward function feedback. It learns to select actions that best improve resource utilization efficiency and meet communication resource scheduling requirements, tailored to different traffic scenarios, resource status, and demands. Ultimately, through continuous learning and optimization, a resource scheduling model is constructed. This model can flexibly and efficiently make resource scheduling decisions within complex highway communication environments based on the real-time status of various traffic clusters and the dynamic coefficients of communication resource demand. This ensures optimal allocation of communication resources and provides reliable resource support for intelligent highway operations.

[0075] In one possible implementation, step S500 further includes:

[0076] Step S510: constructing simulated communication environment information, simulating and executing the communication resource scheduling and allocation strategy based on the simulated communication environment information, and generating a scheduling and allocation result.

[0077] Step S520: traverse the plurality of communication nodes to perform scheduling analysis, set an expected scheduling threshold, and determine whether the scheduling allocation result meets the expected scheduling threshold.

[0078] Step S530: If the scheduling allocation result meets the expected scheduling threshold, a positive feedback instruction is generated, and the communication resource scheduling allocation strategy is continuously monitored through the positive feedback instruction to generate the determination result.

[0079] Step S540: If the scheduling allocation result does not meet the expected scheduling threshold, a negative feedback instruction is generated, and the communication resource scheduling allocation strategy is traversed through the negative feedback instruction to perform abnormality tracing, generate multiple abnormal allocation information, and add the multiple abnormal allocation information to the judgment result.

[0080] Specifically, a simulated communication environment is constructed, encompassing factors such as fluctuations in highway traffic flow, changes in communication demand across different road sections, and the dynamic adjustment of communication equipment performance parameters, to recreate the actual communication environment as realistically as possible. Based on this simulated communication environment, a communication resource scheduling and allocation strategy is simulated. Different types and quantities of communication resources are allocated to each communication node according to the strategy, and the transmission and usage of resources in the virtual environment are simulated to ultimately generate scheduling and allocation results. For example, the simulation simulates the usage of bandwidth resources by communication nodes on different road sections during peak traffic hours after being allocated according to a predetermined strategy.

[0081] After generating the dispatch allocation results, the system traverses multiple communication nodes for dispatch analysis. For each communication node, a desired dispatch threshold is set based on the traffic and communication characteristics of the area in which it is located. This threshold can be an ideal range for metrics such as bandwidth utilization, data transmission delay, and communication connection success rate. The various metrics in the dispatch allocation results are compared with the desired dispatch threshold to determine whether the dispatch allocation results meet expectations. For example, if a communication node expects bandwidth utilization to be between 70% and 90% and data transmission delay to be no more than 50 milliseconds, the actual bandwidth utilization and data transmission delay of the node after resource allocation are checked.

[0082] If the scheduling results meet the expected scheduling threshold, the current communication resource scheduling strategy performs well in the simulation environment. At this point, a positive feedback instruction is generated to continuously monitor the communication resource scheduling strategy. In subsequent simulations or actual applications, the strategy's performance is continuously monitored, and its effects in different scenarios are recorded to confirm its stability and reliability. The monitoring results will generate a judgment result, proving that the strategy is effective in the current evaluation.

[0083] When the scheduling allocation result fails to meet the expected scheduling threshold, the exception handling process is immediately initiated, generating a negative feedback instruction. Based on the negative feedback instruction, the policy begins its initial configuration parameters and conducts a comprehensive review of the resource allocation rules, execution logic, and application in the simulated communication environment. During this review, the system focuses on identifying factors that may have contributed to the unsatisfactory allocation results. For example, the system examines whether the resource allocation rules are reasonable and whether underestimation of the complexity of certain traffic scenarios may have resulted in the allocation scheme failing to meet demand in specific situations. For example, during a sudden traffic surge, if the bandwidth allocated to a communication node on a certain road section fails to increase according to actual demand, resulting in significant data transmission delays at that node, this is a possible anomaly. Furthermore, the system monitors the data flow and resource allocation process during policy execution to identify resource allocation conflicts or improper priority settings. If multiple communication nodes compete for a certain resource type and the policy fails to coordinate effectively, resource shortages may occur at some nodes, resulting in failure to meet the expected scheduling threshold. This comprehensive anomaly tracing process identifies multiple abnormal allocation information points. This information details the location of the anomaly, the relevant parameter settings, and the possible causes. For example, it can record instances where a resource allocation rule fails in a specific traffic scenario, or where a communication node doesn't receive sufficient resources due to incorrect priority settings. Finally, these multiple abnormal allocation information are added to the judgment results, providing detailed and critical evidence for subsequent optimization and improvement of communication resource scheduling and allocation strategies, ensuring more accurate and efficient communication resource allocation in real-world applications.

[0084] Example 2, based on the same inventive concept as the method for optimizing the scheduling of communication resources for smart high-speed communication in the above embodiment, Figure 2 As shown, the present application provides a communication resource scheduling optimization system for intelligent high-speed communication. The system and method embodiments in the present application are based on the same inventive concept. The system includes:

[0085] The high-speed real-time operation data set acquisition module 10 is used to start the multi-source data acquisition module to perform real-time sensing on the target highway and obtain the high-speed real-time operation data set.

[0086] The operation resource classification list formulation module 20 is used to perform multi-scenario analysis on the target highway based on the high-speed real-time operation data set, determine the traffic scenario information set, perform cluster analysis on the high-speed real-time operation data set according to the traffic scenario information set, and formulate an operation resource classification list based on the analysis results.

[0087] The resource scheduling model construction module 30 is used to traverse the operating resource classification list to dynamically calculate the communication resource demand, determine multiple communication resource demand dynamic coefficients, perform deep reinforcement learning according to the multiple communication resource demand dynamic coefficients, and construct a resource scheduling model.

[0088] The communication resource scheduling allocation strategy generating module 40 is configured to synchronize the high-speed real-time running data set to the resource scheduling model for resource scheduling and generate a communication resource scheduling allocation strategy.

[0089] The intelligent scheduling optimization module 50 is used to simulate the execution of the communication resource scheduling allocation strategy to make a scheduling judgment, make feedback adjustments to the communication resource scheduling allocation strategy based on the judgment result, and generate the communication resource scheduling allocation optimization strategy to perform intelligent scheduling optimization on the communication resources of the smart highway.

[0090] Furthermore, the operating resource classification list formulation module 20 further includes:

[0091] The scene feature value determination unit is used to traverse the high-speed real-time running data set to extract scene features and determine multiple scene feature values.

[0092] The scene feature determination unit is used to perform feature dimensionality reduction on the multiple scene feature values ​​by using principal component analysis to determine multiple scene features.

[0093] The traffic scene information set determination unit is used to perform scene analysis on the target expressway based on the multiple scene features to determine the traffic scene information set.

[0094] The traffic operation cluster center acquisition unit is used to perform cluster analysis on the high-speed real-time operation data set according to the traffic scene information set to obtain multiple traffic operation cluster centers.

[0095] The traffic operation cluster determining unit is configured to divide the high-speed real-time operation data set according to the multiple traffic operation cluster centers to determine multiple traffic operation clusters.

[0096] An analysis result adding unit is used to add the multiple traffic operation clusters to the analysis result.

[0097] Furthermore, the scene feature determination unit further includes:

[0098] The characteristic data matrix construction unit is used to perform covariance calculation based on the multiple scene characteristic values ​​to construct a characteristic data matrix.

[0099] The principal component basis vector matrix determining unit is configured to determine the principal component basis vector matrix by arranging the characteristic data matrix in descending order according to the plurality of scene characteristic values.

[0100] The feature dimension reduction matrix acquisition unit is used to project the multiple scene feature values ​​onto the principal component basis vector matrix to obtain a feature dimension reduction matrix.

[0101] A dimensionality reduction verification unit is used to perform dimensionality reduction verification based on the feature dimensionality reduction matrix and obtain the multiple scene features based on the verification result.

[0102] Furthermore, the operating resource classification list formulation module 20 further includes:

[0103] The cluster label determination unit is used to retrieve a historical operation resource demand dataset of a target highway, traverse the historical operation resource demand dataset, assign labels to the multiple traffic operation clusters, and determine multiple cluster labels.

[0104] The intra-cluster resource demand index determining unit is configured to perform demand analysis on the plurality of traffic operation clusters based on the plurality of cluster labels to determine a plurality of intra-cluster resource demand indexes.

[0105] The demand priority determination unit is configured to classify the multiple intra-cluster resource demand indicators according to demand quantities and determine multiple demand priorities.

[0106] The operating resource classification list construction unit is configured to construct the operating resource classification list by arranging the plurality of demand priorities in combination with the plurality of intra-cluster resource demand indicators.

[0107] Furthermore, the resource scheduling model building module 30 also includes:

[0108] The communication node determination unit is used to perform communication analysis on the target expressway according to the historical operation resource demand data set to determine multiple communication nodes.

[0109] The demand priority extraction unit is used to perform communication demand analysis based on the operating resource classification list in combination with the multiple communication nodes, determine multiple basic communication demands, and extract the multiple demand priorities.

[0110] The real-time communication demand data acquisition unit is used to perform real-time calculation based on the multiple traffic operation clusters according to the multiple communication basic demand quantities to obtain multiple real-time communication demand data.

[0111] The multiple regression analysis unit is used to perform multiple regression analysis based on the multiple real-time communication demand data to obtain the multiple communication resource demand dynamic coefficients.

[0112] Furthermore, the resource scheduling model building module 30 also includes:

[0113] The resource allocation path planning unit is configured to traverse the plurality of communication nodes based on the plurality of communication resource demand dynamic coefficients and plan a plurality of resource allocation paths.

[0114] The resource scheduling state space construction unit is used to perform resource state analysis according to the multiple communication resource demand dynamic coefficients and construct a resource scheduling state space.

[0115] The resource scheduling action space construction unit is used to perform resource action analysis according to the multiple resource allocation paths and construct a resource scheduling action space.

[0116] An interactive training unit is used to introduce a reward function, interactively train the resource scheduling state space and the resource scheduling action space, and construct the resource scheduling model.

[0117] Furthermore, the intelligent scheduling optimization module 50 also includes:

[0118] The scheduling allocation result generating unit is used to construct simulated communication environment information, simulate and execute the communication resource scheduling allocation strategy based on the simulated communication environment information, and generate a scheduling allocation result.

[0119] The expected scheduling threshold judgment unit is used to traverse the multiple communication nodes to perform scheduling analysis, set an expected scheduling threshold, and judge whether the scheduling allocation result meets the expected scheduling threshold.

[0120] The continuous monitoring unit is used to generate a positive feedback instruction if the scheduling allocation result meets the expected scheduling threshold, and continuously monitor the communication resource scheduling allocation strategy through the positive feedback instruction to generate the judgment result.

[0121] An abnormal allocation information generation unit is used to generate a negative feedback instruction if the scheduling allocation result does not meet the expected scheduling threshold, traverse the communication resource scheduling allocation strategy through the negative feedback instruction to trace the abnormality, generate multiple abnormal allocation information, and add the multiple abnormal allocation information to the judgment result.

[0122] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0123] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application.

[0124] This specification and drawings are merely illustrative of the present application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Obviously, those skilled in the art may make various modifications and variations to this application without departing from the scope of this application. Thus, this application is intended to include such modifications and variations as fall within the scope of this application and its equivalents.

Claims

1. A method for optimizing communication resource scheduling for intelligent high-speed communication, characterized in that: The method comprises: Start the multi-source data acquisition module to perform real-time sensing on the target highway and obtain a high-speed real-time operation data set; performing a multi-scenario analysis on the target highway based on the high-speed real-time operation dataset, determining a traffic scenario information set, performing a cluster analysis on the high-speed real-time operation dataset according to the traffic scenario information set, and formulating an operation resource classification list based on the analysis results; Traversing the operational resource classification list to dynamically calculate communication resource requirements, determining a plurality of communication resource requirement dynamic coefficients, performing deep reinforcement learning according to the plurality of communication resource requirement dynamic coefficients, and constructing a resource scheduling model; Synchronizing the high-speed real-time running data set to the resource scheduling model for resource scheduling, and generating a communication resource scheduling allocation strategy; The communication resource scheduling and allocation strategy is simulated and executed to make a scheduling judgment, and the communication resource scheduling and allocation strategy is feedback-adjusted according to the judgment result to generate the communication resource scheduling and allocation optimization strategy to perform intelligent scheduling optimization on the communication resources of the smart highway.

2. A method for optimizing communication resource scheduling for intelligent high-speed communication according to claim 1, characterized in that: Performing a multi-scenario analysis on a target highway based on the high-speed real-time operation dataset to determine a traffic scenario information set, and performing a cluster analysis on the high-speed real-time operation dataset according to the traffic scenario information set, the method comprising: Traversing the high-speed real-time running data set to extract scene features and determine multiple scene feature values; Performing feature dimensionality reduction on the plurality of scene feature values ​​using principal component analysis to determine a plurality of scene features; Performing scene analysis on the target highway based on the multiple scene features to determine a traffic scene information set; performing cluster analysis on the high-speed real-time operation data set according to the traffic scene information set to obtain a plurality of traffic operation cluster centers; Dividing the high-speed real-time operation data set according to the multiple traffic operation cluster centers to determine multiple traffic operation clusters; The plurality of traffic operation clusters are added to the analysis result.

3. A method for optimizing communication resource scheduling for intelligent high-speed communication according to claim 2, characterized in that: Performing feature dimensionality reduction on the plurality of scene feature values ​​using principal component analysis to determine a plurality of scene features, the method comprising: Performing covariance calculation based on the multiple scene feature values ​​to construct a feature data matrix; Arrange the plurality of scene eigenvalues ​​in descending order according to the feature data matrix to determine a principal component basis vector matrix; Projecting the multiple scene feature values ​​onto the principal component basis vector matrix to obtain a feature dimension reduction matrix; Dimensionality reduction verification is performed according to the feature dimensionality reduction matrix, and the multiple scene features are obtained according to the verification result.

4. A method for optimizing communication resource scheduling for intelligent high-speed communication according to claim 2, characterized in that: The method of formulating the operation resource classification list according to the analysis results includes: Retrieving a historical operation resource demand dataset of a target expressway, traversing the historical operation resource demand dataset to assign labels to the plurality of traffic operation clusters, and determining a plurality of cluster labels; performing demand analysis on the multiple traffic operation clusters based on the multiple cluster labels to determine multiple intra-cluster resource demand indicators; Classifying the multiple intra-cluster resource demand indicators according to demand quantities and determining multiple demand priorities; The operating resource classification list is constructed by arranging the multiple demand priorities in combination with the multiple intra-cluster resource demand indicators.

5. A method for optimizing communication resource scheduling for intelligent high-speed communication according to claim 4, characterized in that: Traversing the operational resource classification list to dynamically calculate communication resource requirements and determine a plurality of communication resource requirement dynamic coefficients, the method comprising: performing communication analysis on the target highway according to the historical operation resource demand data set to determine a plurality of communication nodes; Performing communication demand analysis based on the operating resource classification list and the plurality of communication nodes, determining a plurality of basic communication demands, and extracting the plurality of demand priorities; Performing real-time calculations based on the multiple traffic operation clusters and the multiple communication basic demand quantities to obtain multiple real-time communication demand data; A multiple regression analysis is performed based on the plurality of real-time communication demand data to obtain the plurality of communication resource demand dynamic coefficients.

6. A method for optimizing communication resource scheduling for intelligent high-speed communication according to claim 5, characterized in that: Deep reinforcement learning is performed according to the multiple communication resource demand dynamic coefficients to build a resource scheduling model, and the method includes: Traversing the plurality of communication nodes based on the plurality of communication resource demand dynamic coefficients to plan a plurality of resource allocation paths; Perform resource status analysis according to the multiple communication resource demand dynamic coefficients to construct a resource scheduling state space; Perform resource action analysis according to the multiple resource allocation paths to construct a resource scheduling action space; A reward function is introduced, and interactive training is performed on the resource scheduling state space and the resource scheduling action space to construct the resource scheduling model.

7. A method for optimizing communication resource scheduling for intelligent high-speed communication according to claim 5, characterized in that: The method of simulating and executing the communication resource scheduling and allocation strategy to make a scheduling decision includes: Constructing simulated communication environment information, simulating and executing the communication resource scheduling and allocation strategy based on the simulated communication environment information, and generating a scheduling and allocation result; Traversing the plurality of communication nodes to perform scheduling analysis, setting an expected scheduling threshold, and determining whether the scheduling allocation result meets the expected scheduling threshold; If the scheduling allocation result meets the expected scheduling threshold, a positive feedback instruction is generated, and the communication resource scheduling allocation strategy is continuously monitored through the positive feedback instruction to generate the determination result; If the scheduling allocation result does not meet the expected scheduling threshold, a negative feedback instruction is generated, and the communication resource scheduling allocation strategy is traversed through the negative feedback instruction to perform abnormality tracing, generate multiple abnormal allocation information, and add the multiple abnormal allocation information to the judgment result.

8. A communication resource scheduling optimization system for intelligent high-speed communication, characterized in that: The system is used to implement a method for optimizing communication resource scheduling for smart highways according to any one of claims 1 to 7, and the system includes: A high-speed real-time operation data set acquisition module is used to start the multi-source data acquisition module to perform real-time sensing on the target highway and obtain a high-speed real-time operation data set; an operation resource classification list formulation module, configured to perform a multi-scenario analysis on the target expressway based on the high-speed real-time operation data set, determine a traffic scenario information set, perform a cluster analysis on the high-speed real-time operation data set according to the traffic scenario information set, and formulate an operation resource classification list based on the analysis results; a resource scheduling model construction module, configured to traverse the operational resource classification list to dynamically calculate communication resource requirements, determine a plurality of communication resource requirement dynamic coefficients, perform deep reinforcement learning based on the plurality of communication resource requirement dynamic coefficients, and construct a resource scheduling model; a communication resource scheduling allocation strategy generation module, configured to synchronize the high-speed real-time running data set to the resource scheduling model for resource scheduling and generate a communication resource scheduling allocation strategy; The intelligent scheduling optimization module is used to simulate the execution of the communication resource scheduling and allocation strategy to make scheduling judgments, make feedback adjustments to the communication resource scheduling and allocation strategy based on the judgment results, and generate the communication resource scheduling and allocation optimization strategy to perform intelligent scheduling optimization on the communication resources of the smart highway.

Images