Intelligent traffic operation system based on Internet of Vehicles

Abstract

The invention discloses an intelligent traffic operation system based on the Internet of Vehicles, and belongs to the technical field of the Internet of Vehicles. Comprising a data collection module used for collecting user data of a user when a vehicle runs, and the user data comprises historical data and real-time data; the user portrait construction module is connected with the data collection module and is used for generating a user portrait according to the historical data; a traffic demand prediction module which is connected with the user portrait construction module and is used for predicting the traffic demand of the user based on a deep learning model in combination with the user portrait and the real-time data; and the intelligent traffic recommendation module is connected with the traffic demand prediction module and is used for recommending a traffic plan matched with the traffic demand to the user. The technical scheme has the beneficial effects that accurate prediction of the traffic demand is realized, and a personalized traffic package recommendation function can be provided based on the traffic use habit and the traffic demand of the user.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle networking, and in particular to an intelligent traffic operation system. Background Art

[0002] With the rapid development of the automotive industry and Internet of Vehicles technology, more and more vehicles are connected to the Internet through Internet of Vehicles systems to provide a variety of in-vehicle services such as navigation, entertainment, and remote diagnosis. The widespread popularity of these Internet of Vehicles applications has led to an increasing demand for network traffic from users. However, the current traffic operation model is relatively rigid and it is difficult to effectively respond to the complex and changing needs in Internet of Vehicles scenarios.

[0003] In the prior art, the Internet of Vehicles traffic operation system uses a static package management function, and users can only choose fixed traffic packages, such as monthly or quarterly traffic packages. There is a lack of personalized recommendations, and there is insufficient understanding of users' traffic usage habits and preferences, and it is impossible to provide personalized traffic package recommendations based on users' specific needs. Summary of the invention

[0004] The purpose of the present invention is to provide an intelligent traffic operation system based on the Internet of Vehicles to solve the above technical problems.

[0005] An intelligent traffic operation system based on Internet of Vehicles, comprising:

[0006] A data collection module, used to collect user data of the user when the vehicle is running, the user data includes historical data and real-time data;

[0007] A user portrait building module, connected to the data collection module, for generating a user portrait based on the historical data;

[0008] A traffic demand prediction module, connected to the user portrait construction module, based on a deep learning model, combines the user portrait and the real-time data to predict the user's traffic demand;

[0009] The intelligent traffic recommendation module is connected to the traffic demand prediction module and is used to recommend traffic packages matching the traffic demand to the user.

[0010] Preferably, the data collection module comprises:

[0011] A data collection unit, which collects the historical data and the real-time data through a vehicle-mounted terminal;

[0012] The preprocessing unit is connected to the data collection unit and is used to clean and deduplicate the collected historical data and real-time data before outputting them.

[0013] Preferably, the user portrait building module includes:

[0014] A cluster construction unit, used for dividing the historical data into a plurality of clusters;

[0015] a feature extraction unit, connected to the cluster construction unit, for analyzing each cluster to extract the user's traffic usage pattern and preference;

[0016] A user portrait construction unit is connected to the feature extraction unit and is used to construct the user portrait according to the user's traffic usage pattern and preference.

[0017] Preferably, the flow demand is predicted using the following formula:

[0018] L 预测 =f(U 历史 ,V 实时 ,A 使用 );

[0019] Among them, L 预测 represents the flow demand, U 历史 Represents the user portrait, V 实时 represents the real-time status data of the vehicle in the real-time data, A 使用 It represents the user's real-time usage data in the real-time data.

[0020] Preferably, the intelligent traffic recommendation module includes:

[0021] A package pre-storage unit, used to pre-save multiple packages to be selected;

[0022] A package scoring unit, connected to the package pre-storage unit, for processing and obtaining a comprehensive score of each package to be selected according to the traffic demand;

[0023] The package recommendation unit is connected to the package scoring unit and is used to output the package to be selected with the highest comprehensive score as the traffic package recommended to the user.

[0024] Preferably, the comprehensive score is obtained by using the following formula:

[0025]

[0026] Among them, S i represents the comprehensive score of the i-th package to be selected, α and β are both pre-set weight coefficients, M i represents the matching degree of the i-th package to be selected, C i represents the cost performance of the i-th package to be selected;

[0027] L 满足 Indicates the traffic demand that the package to be selected can meet, L 预测indicating predicted traffic demand;

[0028] L i represents the traffic quota of the i-th package to be selected, P i Indicates the price of the package to be selected.

[0029] Preferably, it also includes a traffic allocation module connected to the intelligent traffic recommendation module, and the traffic allocation module includes:

[0030] A first weight adjustment unit, configured to obtain a flow distribution weight of each application program in the vehicle-mounted system according to real-time user usage data in the real-time data monitored in real time;

[0031] a traffic allocation unit, connected to the first weight adjustment unit, for dynamically allocating traffic usage of each application within the framework of the recommended traffic package according to the traffic allocation weight;

[0032] a second weight adjustment unit, connected to the first weight adjustment unit and the traffic allocation unit, respectively, and configured to readjust the traffic allocation weight of each application program by adopting a preset dynamic adjustment strategy, and the traffic allocation unit dynamically allocates the traffic usage of each application program according to the readjusted traffic allocation weight;

[0033] A resource management unit, configured to reserve traffic resources or cache application data in advance for high-priority applications within the framework of the recommended traffic package when the network signal is good;

[0034] The preset dynamic adjustment strategy includes:

[0035] When the network signal becomes weak, the traffic distribution weight of the application with high priority is increased.

[0036] Preferably, the flow distribution weight is obtained by using the following formula:

[0037]

[0038] Among them, ω m represents the traffic distribution weight of the mth application, P m represents the priority of the mth application, D m represents the real-time requirement of the mth application, Q m represents the network signal quality of the mth application;

[0039] m∈[1,j], where j is used to represent the total number of applications currently participating in traffic distribution.

[0040] Preferably, the real-time traffic allocated to each application by the traffic allocation module is obtained by using the following formula:

[0041] R n =ω n ×R total ;

[0042] Among them, R n represents the real-time traffic allocated to the nth application by the traffic allocation module, ω n represents the traffic distribution weight of the nth application, R total Indicates the network resources available within the framework of the recommended data plan.

[0043] Preferably, a feedback and optimization module is further included, connected to the traffic distribution module, and the feedback and optimization module includes:

[0044] A training unit, used to collect historical feedback data of users to form training data, and to build a reinforcement learning model based on the training data;

[0045] An optimization unit, connected to the training unit, is used to adjust the traffic distribution weight according to the reinforcement learning model and real-time feedback data from users.

[0046] The beneficial effects of the present invention are: achieving accurate prediction of traffic demand and being able to provide a personalized traffic package recommendation function based on the user's traffic usage habits and traffic demand. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 is a schematic diagram of an intelligent traffic operation system based on Internet of Vehicles of the present invention;

[0048] Figure 2 is a schematic diagram of a data collection module of the present invention;

[0049] Figure 3 It is a schematic diagram of a user portrait building module of the present invention;

[0050] Figure 4 is a schematic diagram of the intelligent traffic recommendation module of the present invention;

[0051] Figure 5 is a schematic diagram of a flow distribution module of the present invention;

[0052] Figure 6 It is a connection block diagram of the resource management unit and the intelligent traffic recommendation module of the present invention;

[0053] Figure 7 It is a schematic diagram of the feedback and optimization module of the present invention.

[0054] In the attached figure: 1. Data collection module; 11. Data acquisition unit; 12. Preprocessing unit; 2. User portrait construction module; 21. Cluster construction unit; 22. Feature extraction unit; 23. User portrait construction unit; 3. Traffic demand prediction module; 4. Intelligent traffic recommendation module; 41. Package pre-storage unit; 42. Package scoring unit; 43. Package recommendation unit; 5. Traffic allocation module; 51. First weight adjustment unit; 52. Traffic allocation unit; 53. Second weight adjustment unit; 54. Resource management unit; 6. Feedback and optimization module; 61. Training unit; 62. Optimization unit. DETAILED DESCRIPTION

[0055] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0057] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.

[0058] An intelligent traffic operation system based on Internet of Vehicles, such as Figure 1 As shown, including:

[0059] Data collection module 1, used to collect user data of the user when the vehicle is running, the user data includes historical data and real-time data;

[0060] User portrait building module 2, connected to data collection module 1, for generating user portraits based on historical data;

[0061] Traffic demand prediction module 3 is connected to user portrait construction module 2, and based on the deep learning model, combines user portrait and real-time data to predict the user's traffic demand;

[0062] The intelligent traffic recommendation module 4 is connected to the traffic demand prediction module 3 and is used to recommend traffic packages that match the traffic demand to the user.

[0063] Specifically, the present invention provides an intelligent traffic operation system based on the Internet of Vehicles, which collects user data by analyzing the data collection module 1, generates a user portrait based on historical data in the user data, and accurately predicts the user's traffic demand by combining the user portrait and real-time data, so that the intelligent traffic recommendation module 4 can recommend personalized traffic packages to users.

[0064] The present invention is applied to the Internet of Vehicles member mall. Based on the user's historical data, that is, the user's consumption behavior and usage habits, the system provides customized traffic package recommendations to improve user satisfaction. The user's historical data can be uploaded by the vehicle system. By analyzing the user's historical data, merchants can better understand user needs, thereby optimizing products and services and enhancing market competitiveness. Users can choose the most suitable package according to their actual needs, reduce traffic waste, and improve overall service benefits.

[0065] In a preferred embodiment, referring to Figure 2 , the data collection module 1 includes:

[0066] The data collection unit 11 collects historical data and real-time data through the vehicle terminal;

[0067] The pre-processing unit 12 is connected to the data collection unit 11 and is used for cleaning and deduplicating the collected historical data and real-time data before outputting them.

[0068] Specifically, historical data includes the user's historical traffic usage records, application usage frequency and duration, vehicle geographic location, common driving routes, and travel time. User real-time usage data refers to real-time network signals and traffic consumption. The data acquisition unit 11 can also reflect the user's latest usage in real time, so that the system can respond quickly to changes in user needs. By collecting multi-dimensional information such as the user's historical traffic usage records, application usage frequency and duration, and vehicle geographic location, a more detailed and accurate user portrait can be constructed. The preprocessing unit 12 is used to clean and deduplicate the data to ensure the accuracy of the input data and effectively remove redundant and erroneous data.

[0069] In a preferred embodiment, referring to Figure 3 , User portrait construction module 2 includes:

[0070] A cluster construction unit 21, used to divide the historical data into a plurality of clusters;

[0071] A feature extraction unit 22, connected to the cluster construction unit 21, for analyzing each cluster to extract the user's traffic usage pattern and preference;

[0072] The user portrait construction unit 23 is connected to the feature extraction unit 22 and is used to construct a user portrait according to the user's traffic usage pattern and preference.

[0073] Specifically, clustering algorithm is an unsupervised learning method that divides samples in a data set into several groups or clusters that are similar to each other, discovers structures and patterns in the data, and thus helps analyze and understand the data.

[0074] The user portrait construction module 2 processes the collected data, constructs a detailed user portrait, and extracts the user's traffic usage pattern and preferences.

[0075] For example, a user mainly uses navigation and communication apps during weekdays, but uses entertainment apps more frequently on weekends.

[0076] In a preferred embodiment, the flow demand is predicted using the following formula:

[0077] L 预测 =f(U 历史 ,V 实时 ,A 使用 );

[0078] Among them, L 预测 represents the flow demand, U 历史 Represents user portrait, V 实时 Indicates the real-time status data of the vehicle in the real-time data, A 使用 Indicates the real-time usage data of users in real-time data.

[0079] Specifically, the traffic demand prediction module 3 uses a time series model, such as LSTM (Long Short-Term Memory Network) or regression analysis, to predict the traffic demand in the future, and provides accurate traffic demand prediction based on the vehicle's geographic location, driving time, and traffic usage.

[0080] In a preferred embodiment, referring to Figure 4 , the intelligent traffic recommendation module 4 includes:

[0081] A package pre-storage unit 41 is used to pre-storage multiple packages to be selected;

[0082] The package scoring unit 42 is connected to the package pre-storage unit 41 and is used to process and obtain a comprehensive score of each package to be selected according to the traffic demand;

[0083] The package recommendation unit 43 is connected to the package scoring unit 42 and is used to output the package to be selected with the highest comprehensive score as the traffic package recommended to the user.

[0084] Specifically, the pre-storage method of the packages to be selected in the package pre-storage unit 41 is to update them regularly with the service provider to obtain the latest list of traffic packages, or the user can perform an initial screening first to retain the traffic packages that the user is interested in.

[0085] The intelligent traffic recommendation module 4 intelligently recommends personalized traffic packages based on the prediction results. The traffic packages include but are not limited to navigation-specific traffic packages, entertainment traffic packages, and long-distance self-driving-specific packages. The system cooperates with third-party streaming services, navigation applications, e-commerce platforms, etc., and pushes preferential activities based on user needs, such as limited-time streaming packages or long-distance navigation traffic packages.

[0086] In a preferred embodiment, the following formula is used to obtain a comprehensive score:

[0087]

[0088] Among them, S i represents the comprehensive score of the i-th package to be selected, α and β are both pre-set weight coefficients and satisfy α+β=1, M i represents the matching degree of the i-th package to be selected, C i Indicates the cost-effectiveness of the i-th package to be selected;

[0089] L 满足 Indicates the traffic demand that can be met by the package to be selected, L 预测 represents the predicted traffic demand;

[0090] L i represents the traffic quota of the i-th package to be selected, P i Indicates the price of the package to be selected.

[0091] Specifically, the traffic demand that a package can meet refers to the actual traffic demand that a package can effectively cover after predicting the user's traffic demand, which is the estimated value of the user's traffic usage in the future. The specific formula is: traffic demand that a package can meet = min (predicted traffic demand, package traffic quota). If the user's demand is less than or equal to the package quota, the package fully meets the demand; otherwise, it can only partially meet the demand.

[0092] The traffic quota of a package refers to the fixed traffic included in the traffic package itself, such as 15GB, 20GB, etc. per month, which is a clear value.

[0093] The highest-rated packages are recommended to users based on the comprehensive scores. Based on the user's real-time scenarios (such as long-distance driving, city commuting, and night driving), combined with geographic location and time information, traffic packages that are more suitable for the current scenarios are recommended.

[0094] For example: recommending a package that includes more navigation and entertainment traffic when driving long distances.

[0095] In a preferred embodiment, referring to Figure 5 , further comprising a traffic allocation module 5 connected to the intelligent traffic recommendation module 4, the traffic allocation module 5 comprising:

[0096] A first weight adjustment unit 51 is used to obtain the flow distribution weight of each application in the vehicle system according to the user real-time usage data in the real-time data monitored in real time;

[0097] The traffic allocation unit 52 is connected to the first weight adjustment unit 51 and is used to dynamically allocate the traffic usage of each application within the framework of the recommended traffic package according to the traffic allocation weight.

[0098] Specifically, the applications that currently occupy the most processing resources of the in-vehicle entertainment system are applications with high real-time demand, such as applications that provide users with video, audio and other streaming media services in the vehicle. Especially when the traffic is limited, the system can automatically optimize the traffic allocation according to the user's needs and network conditions, thereby achieving higher resource utilization and reducing the waste of network resources.

[0099] In the system application for remote data transmission, when the present invention is not used, the priority of the system application is low. The present invention can identify potential problems through vehicle status data, adjust the priority and traffic allocation weight of the system application, and then dynamically adjust the data transmission rate to ensure that data is transmitted only when necessary, which helps to reduce network traffic consumption and thus reduce operating costs. Through remote monitoring, maintenance personnel can predict potential faults in advance, arrange maintenance work reasonably, and improve vehicle management efficiency.

[0100] In a preferred embodiment, the flow distribution weight is obtained by using the following formula:

[0101]

[0102] Among them, ω m represents the traffic distribution weight of the mth application, P m represents the priority of the mth application (3 represents a high priority application, 2 represents a medium priority application, and 1 represents a low priority application), D m represents the real-time requirement of the mth application (D m According to the current activity of the application, the value range is 0-1), Q m represents the network signal quality (Q m According to the signal strength, the value range is 0-1);

[0103] m∈[1,j], where j is used to represent the total number of applications currently participating in traffic distribution.

[0104] Specifically, the application currently participating in traffic distribution may be all the application programs in the vehicle-mounted system, or may be the application program currently using traffic.

[0105] In a preferred embodiment, the real-time traffic allocated to each application by the traffic allocation module 5 is obtained by using the following formula:

[0106] R n =ω n ×R total ;

[0107] Among them, R n represents the real-time traffic allocated to the nth application by the traffic allocation module 5, ω n represents the traffic distribution weight of the nth application, R total Represents the network resources available within the framework of the recommended data plan.

[0108] In a preferred embodiment, the flow distribution module 5 further includes:

[0109] The second weight adjustment unit 53 is respectively connected to the first weight adjustment unit 51 and the traffic allocation unit 52, and is used to adopt a preset dynamic adjustment strategy to readjust the traffic allocation weight of each application. The traffic allocation unit 52 dynamically allocates the traffic usage of each application according to the readjusted traffic allocation weight.

[0110] Specifically, if the real-time demand of a certain application suddenly increases (such as a navigation application needs more data on a complex road section), the second weight adjustment unit 53 will adjust the weight immediately to increase the flow allocation to the application. Through dynamic adjustment, limited network bandwidth can be reasonably utilized, so that high-priority applications can obtain sufficient resources, while unimportant applications can reduce their flow allocation accordingly.

[0111] In a preferred embodiment, the preset dynamic adjustment strategy includes:

[0112] When the network signal becomes weak, the traffic distribution weight of applications with high priority is increased.

[0113] Specifically, existing technologies lack the ability to monitor and analyze user traffic usage in real time and cannot make dynamic adjustments, which limits the user experience in the Internet of Vehicles, especially in long-term driving or complex scenarios, where insufficient or wasted traffic often occurs. And it is impossible to flexibly adjust according to the specific status of the vehicle (such as driving area, signal strength, application requirements, etc.), resulting in a poor user experience. The dynamic adjustment mechanism is that when the network signal becomes weak, the system automatically increases the weight of high-priority applications to ensure the normal operation of key functions and reduce service interruptions caused by poor network.

[0114] In a preferred embodiment, referring to Figure 6 , the traffic distribution module 5 also includes:

[0115] The resource management unit 54 is used to reserve traffic resources or cache application data in advance for high-priority applications within the framework of the recommended traffic package when the network signal is good.

[0116] Specifically, when network conditions are good, the system will reserve a certain amount of traffic resources for high-priority applications, or cache necessary data in advance, to improve service stability during network fluctuations, avoid delays or freezes caused by network fluctuations, and thus provide a smoother service experience.

[0117] The system monitors the user's traffic usage in real time, predicts the traffic consumption trend, and adjusts the traffic allocation in time to avoid excess or insufficient traffic, making resource allocation more flexible and efficient. The traffic allocation module 5 can adjust the allocated traffic in real time. In order to ensure that traffic resources are allocated to appropriate applications at the appropriate time, the system monitors the vehicle's driving conditions, network signals and traffic consumption in real time, and uses a dynamic adjustment mechanism to adjust traffic allocation. The specific dynamic adjustment rules are as follows:

[0118] (1) Real-time monitoring

[0119] Monitor vehicle status, obtain information such as vehicle speed, location, driving route, etc., and determine the current driving scenario (such as high-speed driving, congested road section, parking status);

[0120] Monitor network signals, detect network signal strength, bandwidth and other parameters in real time, and evaluate the current network quality;

[0121] Monitor application traffic, record the real-time traffic consumption of each application, and understand the application usage frequency and traffic demand.

[0122] (2) Application priority setting

[0123] Priority classification is to classify in-vehicle applications into three priorities: high, medium, and low according to their importance;

[0124] High-priority applications, such as navigation, emergency communications, security monitoring and other key functions;

[0125] Medium priority applications, such as vehicle remote diagnosis, system updates and other auxiliary functions;

[0126] Low priority applications such as entertainment, social media, and other non-critical functions.

[0127] (3) Dynamic Scheduling Algorithm

[0128] By using application priority, real-time requirements and signal quality, effective bandwidth management is achieved by calculating traffic allocation weights and corresponding traffic allocation.

[0129] In a preferred embodiment, it also includes a feedback and optimization module 6, connected to the flow distribution module 5, referring to Figure 7 , the feedback and optimization module 6 includes:

[0130] A training unit 61, used to collect historical feedback data of users to form training data, and to build a reinforcement learning model according to the training data;

[0131] The optimization unit 62 is connected to the training unit 61 and is used to adjust the traffic distribution weight according to the reinforcement learning model and the real-time feedback data of the user.

[0132] Specifically, existing user traffic usage feedback has not been effectively applied to the system optimization process, resulting in difficulty in improving user experience. The feedback and optimization module 6 of the present invention can continuously optimize the recommendation strategy according to the user's actual usage and improve user satisfaction.

[0133] Historical feedback data refers to feedback data from the same user at different usage periods, or feedback data from different users. A reinforcement learning model is constructed by using a large amount of user feedback data as a training data set. The user's real-time feedback data is input into the reinforcement learning model to adjust the traffic allocation weight for the next use.

[0134] The real-time feedback data input by users into the reinforcement learning model is diverse, such as user voice input "The traffic of a certain application is not enough", "A certain application has not been used recently", or regular questionnaires on application traffic usage are issued to users. The content of the questionnaire includes: the length of time the application is used over a period of time, the smoothness of application use, whether there are any freezes and other problems.

[0135] The optimization unit 62 may adjust the traffic distribution weight directly by adjusting the traffic distribution weight calculated next time, or may directly adjust the factors affecting the calculation of the traffic distribution weight, such as the priority of the application program.

[0136] The present invention relies on in-vehicle intelligent terminals (such as in-vehicle navigation, in-vehicle entertainment equipment, etc.) and user mobile phone applications for operation, and the background is based on cloud computing and big data analysis.

[0137] The on-board intelligent terminal is used to collect real-time vehicle data, on-board application traffic usage, and geographic location information.

[0138] Cloud servers are used for data processing, model training, traffic demand prediction, and personalized recommendations.

[0139] On mobile devices (such as mobile phone APP), users can manage in-vehicle data packages, check usage and order data packages through the APP.

[0140] The front end of the system is responsible for interacting with users, while the back end is responsible for data processing, analysis, and optimization:

[0141] The front-end interface of the system includes traffic monitoring module, personalized recommendation module and dynamic prompt module.

[0142] The traffic monitoring module displays the current vehicle traffic usage in real time, including the traffic consumed by various applications such as navigation, entertainment, and remote diagnosis;

[0143] The personalized recommendation module displays the traffic packages recommended by the backend to users, and users can order them with one click;

[0144] Dynamic reminder module: when the user's traffic is close to insufficient, the system front end will prompt the user to purchase additional traffic packages through pop-up windows or notifications.

[0145] The present invention is applied to long-distance driving. Taking the intelligent traffic management and optimization of long-distance driving as an example, the car owner user A plans a long-distance self-driving tour. The driving route spans multiple provinces and is expected to last for two days. During the driving process, the navigation is frequently used on the front end of the vehicle system of user A, and the vehicle entertainment system is frequently used during rest. At present, there is insufficient traffic, especially in areas with poor network signals, which may cause network disconnection. The present invention is used in the above-mentioned long-distance self-driving tour, and the specific implementation process is as follows,

[0146] Step 1: Data collection and analysis:

[0147] The system obtains the long-distance route of user A through data collection module 1, including the highway sections, service areas and network coverage;

[0148] At the same time, the usage frequency and content preferences (such as watching movies, listening to music, etc.) of user A's in-vehicle entertainment system are monitored to generate a user portrait.

[0149] Step 2: Predict traffic demand:

[0150] Traffic demand prediction module 3 combines user portraits and uses deep learning models to make real-time predictions on traffic demand, predicting that user A will need to use approximately 500MB of navigation traffic, 2GB of streaming data, and 200MB of other application data during the journey.

[0151] Step 3: Recommended data packages:

[0152] The intelligent traffic recommendation module 4 scores multiple pre-saved packages to be selected based on the traffic demand of user A on the way: using about 500MB of navigation traffic, 2GB of streaming data and 200MB of other application data, and outputs the package to be selected with the highest comprehensive score as the traffic package recommended to the user.

[0153] Step 4: Intelligent traffic allocation:

[0154] On roads with stronger signals, more traffic resources are allocated to user A to ensure smooth use of the navigation and entertainment systems;

[0155] Before entering an area with weak signals, the resource management unit 54 caches necessary navigation data in advance and preloads streaming media videos to prevent network fluctuations from affecting the user A's experience;

[0156] The system also recommends a long-distance self-driving data package to user A, which is suitable for navigation, entertainment and other applications when driving across provinces, providing more data reserves.

[0157] Specifically, the present invention introduces an intelligent dynamic adjustment mechanism in traffic allocation, dynamically adjusts traffic allocation according to the user's real-time needs and network environment, and avoids traffic waste or shortage. It provides a personalized traffic package recommendation function. Based on the user's traffic usage habits and application scenarios, the system can push traffic packages or preferential activities that meet the user's needs in real time, which is more flexible and efficient than the single, static package selection mode in the prior art.

[0158] The system also uses a feedback learning mechanism, which can continuously optimize the recommendation strategy based on the user's actual usage, thereby improving the accuracy of subsequent recommendations and user satisfaction.

[0159] In summary, this application provides an intelligent traffic operation system based on the Internet of Vehicles, which realizes accurate prediction and real-time regulation of traffic demand through machine learning algorithms and data analysis, and improves the utilization rate of traffic resources and user experience. It provides personalized traffic package recommendations based on the user's traffic usage habits and application scenarios, realizes accurate prediction and real-time regulation of traffic demand, and can continuously optimize the recommendation strategy according to the user's actual usage, thereby improving user satisfaction.

[0160] The above description is only a preferred embodiment of the present invention, and does not limit the implementation mode and protection scope of the present invention. For those skilled in the art, it should be aware that all solutions obtained by equivalent substitutions and obvious changes made using the description and illustrations of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent traffic operation system based on Internet of Vehicles, characterized in that: include: A data collection module, used to collect user data of the user when the vehicle is running, the user data includes historical data and real-time data; A user portrait building module, connected to the data collection module, for generating a user portrait based on the historical data; A traffic demand prediction module, connected to the user portrait construction module, based on a deep learning model, combines the user portrait and the real-time data to predict the user's traffic demand; The intelligent traffic recommendation module is connected to the traffic demand prediction module and is used to recommend traffic packages matching the traffic demand to the user.

2. The intelligent traffic operation system according to claim 1, characterized in that: The data collection module includes: A data collection unit, which collects the historical data and the real-time data through a vehicle-mounted terminal; The preprocessing unit is connected to the data collection unit and is used to clean and deduplicate the collected historical data and real-time data before outputting them.

3. The intelligent traffic operation system according to claim 1, characterized in that: The user portrait building module includes: A cluster construction unit, used for dividing the historical data into a plurality of clusters; a feature extraction unit, connected to the cluster construction unit, for analyzing each cluster to extract the user's traffic usage pattern and preference; A user portrait construction unit is connected to the feature extraction unit and is used to construct the user portrait according to the user's traffic usage pattern and preference.

4. The intelligent traffic operation system according to claim 1, characterized in that: The flow demand is predicted using the following formula: L 预测 =f(U 历史 ,V 实时 ,A 使用 ); Among them, L 预测 represents the flow demand, U 历史 Represents the user portrait, V 实时 represents the real-time status data of the vehicle in the real-time data, A 使用 It represents the user's real-time usage data in the real-time data.

5. The intelligent traffic operation system according to claim 1, characterized in that: The intelligent traffic recommendation module includes: A package pre-storage unit, used to pre-save multiple packages to be selected; A package scoring unit, connected to the package pre-storage unit, for processing and obtaining a comprehensive score of each package to be selected according to the traffic demand; The package recommendation unit is connected to the package scoring unit and is used to output the package to be selected with the highest comprehensive score as the traffic package recommended to the user.

6. The intelligent traffic operation system according to claim 5, characterized in that: The comprehensive score is obtained by using the following formula: Among them, S i represents the comprehensive score of the i-th package to be selected, α and β are both pre-set weight coefficients, M i represents the matching degree of the i-th package to be selected, C i represents the cost performance of the i-th package to be selected; L 满足 Indicates the traffic demand that the package to be selected can meet, L 预测 indicating predicted traffic demand; L i represents the traffic quota of the i-th package to be selected, P i Indicates the price of the package to be selected.

7. The intelligent traffic operation system according to claim 1, characterized in that: It also includes a traffic allocation module connected to the intelligent traffic recommendation module, and the traffic allocation module includes: A first weight adjustment unit, configured to obtain a flow distribution weight of each application program in the vehicle-mounted system according to real-time user usage data in the real-time data monitored in real time; a traffic allocation unit, connected to the first weight adjustment unit, for dynamically allocating traffic usage of each application within the framework of the recommended traffic package according to the traffic allocation weight; a second weight adjustment unit, connected to the first weight adjustment unit and the traffic allocation unit, respectively, and configured to readjust the traffic allocation weight of each application program by adopting a preset dynamic adjustment strategy, and the traffic allocation unit dynamically allocates the traffic usage of each application program according to the readjusted traffic allocation weight; A resource management unit, configured to reserve traffic resources or cache application data in advance for high-priority applications within the framework of the recommended traffic package when the network signal is good; The preset dynamic adjustment strategy includes: When the network signal becomes weak, the traffic distribution weight of the application with high priority is increased.

8. The intelligent traffic operation system according to claim 7, characterized in that: The flow distribution weight is obtained by using the following formula: Among them, ω m represents the traffic distribution weight of the mth application, P m represents the priority of the mth application, D m represents the real-time requirement of the mth application, Q m Indicates the network signal quality of the mth application; m∈[1,j], where j is used to represent the total number of applications currently participating in traffic distribution.

9. The intelligent traffic operation system according to claim 7, characterized in that: The real-time traffic allocated to each application by the traffic allocation module is obtained by using the following formula: R n =ω n ×R total ; Among them, R n represents the real-time traffic allocated to the nth application by the traffic allocation module, ω n represents the traffic distribution weight of the nth application, R total Indicates the network resources available within the framework of the recommended data plan.

10. The intelligent traffic operation system according to claim 7, characterized in that: It also includes a feedback and optimization module connected to the traffic distribution module, and the feedback and optimization module includes: A training unit, used to collect historical feedback data of users to form training data, and to build a reinforcement learning model based on the training data; An optimization unit, connected to the training unit, is used to adjust the traffic distribution weight according to the reinforcement learning model and real-time feedback data from users.

Images