Vehicle driving intelligent control method and device

Abstract

The embodiment of the invention provides a vehicle driving intelligent control method and device. Multi-dimensional data collection is achieved through an edge box, an intelligent camera and environment detection equipment. A regional linkage network is dynamically divided based on a vehicle driving track, and a traffic flow prediction and risk assessment model is trained at an edge end. The cloud platform integrates the data of the plurality of analysis and identification units, triggers an area linkage mechanism according to a road passing strategy and a risk assessment value, and generates accurate early warning information including real-time road conditions, suggested detouring routes and safe vehicle speeds. According to the method, the defects of the traditional technology in the aspects of scene perception, data collaboration, early warning decision and the like are effectively overcome, and the intelligent level and the service quality of the auxiliary driving system are remarkably improved.

Description

Technical Field

[0001] This application relates to the field of data processing, and in particular, to an intelligent vehicle driving control method and device. Background Art

[0002] The existing intelligent vehicle driving control methods have obvious deficiencies. Traditional systems often rely on a single in-vehicle device for road condition perception, lack the collaborative cooperation of roadside devices, and it is difficult to achieve comprehensive traffic scene perception.

[0003] In addition, there are bottlenecks in data processing and early warning in the existing technologies. Most systems fail to establish an effective regional linkage mechanism, lack the ability of intelligent prediction and risk assessment based on multi-source data, resulting in lagged or inaccurate early warning information.

[0004] The existing systems have technical shortcomings in traffic flow prediction and road condition analysis. They lack comprehensive consideration of environmental factors and it is difficult to provide timely and accurate driving suggestions for complex road conditions. Solving these problems is of great significance for improving the intelligence level and practicality of the assisted driving system. Summary of the Invention

[0005] In view of the problems in the existing technologies, this application provides an intelligent vehicle driving control method and device, which can effectively solve the deficiencies of traditional technologies in aspects such as scene perception, data collaboration, and early warning decision-making, and significantly improve the intelligence level and service quality of the assisted driving system.

[0006] To solve at least one of the above problems, this application provides the following technical solutions: In a first aspect, this application provides an intelligent vehicle driving control method, including: An analysis and recognition unit is set along the road. The analysis and recognition unit includes an edge box, an intelligent camera, and an environment detection device. The edge box establishes a communication connection with and registers to a cloud platform. The intelligent camera includes a license plate capture camera and a wide-angle camera. The environment detection device collects road condition environment parameters. An in-vehicle module is deployed on the vehicle. The in-vehicle module establishes a connection with and registers to the cloud platform through an Internet of Things communication module. The in-vehicle module calls a vehicle machine interface to obtain the vehicle driving trajectory, and divides multiple analysis and recognition units into a regional linkage network based on the driving trajectory; The license plate capture camera obtains the license plate number, license plate color and vehicle type information of the vehicle. The wide-angle camera obtains the real-time road scene. The environmental detection device obtains the road visibility and road surface state parameters. The edge box extracts the vehicle driving characteristics based on the real-time road scene. The edge box trains the traffic flow prediction model and the risk assessment model based on the vehicle driving characteristics, and sends the vehicle driving characteristics, the traffic flow prediction results, the risk assessment values and the environmental parameters collected by each of the analysis and recognition units to the cloud platform; The cloud platform receives the data uploaded by multiple analysis and recognition units, determines whether to trigger regional linkage based on the preset road traffic strategy and the risk assessment value. When regional linkage is triggered, it sends linkage instructions to the relevant analysis and recognition units according to the regional linkage network and establishes a communication link, and receives the traffic flow prediction results uploaded by the relevant analysis and recognition units. The cloud platform generates warning information including real-time road conditions, recommended detour routes and safe vehicle speeds based on the traffic flow prediction results, and sends it to the in-vehicle module corresponding to the license plate number for broadcasting.

[0007] Further, it also includes: deploying the analysis and recognition units at key positions on the road, initializing the parameters of the edge box. The edge box obtains a unique identification code through the Internet of Things communication module, and sends a registration request to the cloud platform based on the unique identification code. The cloud platform returns the registration authorization information, and the edge box establishes a secure communication link with the cloud platform based on the registration authorization information; The edge box establishes a communication connection with the intelligent camera and the environmental detection device, calibrates the angle and focal length of the intelligent camera. The license plate capture camera is set to point to the vehicle passing direction, and the wide-angle camera is set to the road panoramic monitoring angle. A sub-device management group is established with the edge box as the center, and the sampling period and trigger threshold of the environmental detection device are configured. The edge box obtains environmental parameters including visibility, temperature, humidity and light intensity from the environmental detection device at preset time intervals.

[0008] Further, it also includes: the in-vehicle module obtains the device unique identifier through the Internet of Things communication module, and sends a registration request including the vehicle VIN code to the cloud platform based on the device unique identifier. The cloud platform returns the device authorization code, and the in-vehicle module establishes a data transmission channel with the cloud platform based on the device authorization code. The in-vehicle module calls the vehicle CAN bus interface to collect the vehicle's real-time position coordinates, driving speed and direction angle data; The in-vehicle module uploads the position coordinates, driving speed, and direction angle data to the cloud platform at preset time intervals. The cloud platform maps the vehicle position coordinates to an electronic map based on the geographic information system, determines the forward prediction path according to the vehicle driving direction, extracts the analysis and recognition units within the coverage of the prediction path, constructs the extracted analysis and recognition units into a regional linkage network, and assigns network priorities to each analysis and recognition unit.

[0009] Further, it also includes: The license plate capture camera collects image data when the vehicle passes by. The edge box processes the image data based on a convolutional neural network model, extracts license plate area features and performs character segmentation, identifies the license plate number and color information, obtains the vehicle type classification result by analyzing the body contour features through a deep learning model, and stores the license plate number, color information, and vehicle type classification result in the cache database of the edge box; The wide-angle camera collects road traffic video streams at a preset frame rate. The environmental detection device obtains visibility values based on the principle of laser ranging, detects road surface state parameters through a surface temperature sensor and a humidity sensor. The edge box performs target detection and tracking processing on the video stream, extracts driving feature parameters such as vehicle traffic trajectories, vehicle speeds, and vehicle spacings. The edge box establishes a mapping relationship between the driving feature parameters and the license plate information.

[0010] Further, it also includes: The edge box divides the vehicle driving feature parameters into a training data set according to the time series, constructs the network structure of a traffic flow prediction model using a long short-term memory neural network, uses vehicle speed, vehicle density, and travel time as input features, iteratively trains the network structure to obtain a traffic flow prediction model, constructs a risk assessment model based on the decision tree algorithm, and calculates the collision risk assessment value using vehicle spacing, relative speed, and road surface state parameters as input variables; The edge box organizes the vehicle driving features, environmental parameters, and timestamp information into a data packet. The data packet contains the future traffic flow state prediction result output by the traffic flow prediction model and the risk assessment value calculated by the risk assessment model. The edge box sends the data packet to the cloud platform through a secure communication link according to a preset data upload period. The cloud platform performs time synchronization and data verification on the received data packet.

[0011] Further, it further includes: the cloud platform receives data packets uploaded by multiple analysis and recognition units, extracts index parameters such as the risk assessment value, traffic flow, and vehicle speed from the data packets, compares the index parameters with the thresholds in the road traffic strategy, and the road traffic strategy includes a traffic flow density threshold, an average vehicle speed threshold, and a per-unit-mile accident risk threshold. When the index parameters of any analysis and recognition unit exceed the corresponding threshold, area linkage judgment is triggered, and the cloud platform determines the affected range according to the position of the analysis and recognition unit that triggers the linkage; The cloud platform selects the analysis and recognition units located within the affected range from the area linkage network as linkage objects, generates a linkage session including the identification codes of the linkage objects, creates a data sharing channel based on the linkage session, sends a linkage instruction including the configuration information of the sharing channel to the linkage objects, and receives the confirmation response returned by the linkage objects through the data sharing channel to complete the establishment of the communication link between the linkage objects.

[0012] Further, it further includes: the cloud platform receives the traffic flow prediction results uploaded by the linkage objects, maps the traffic flow prediction results to road segments based on real-time position coordinates, calculates the congestion index of each road segment, the cloud platform calls a path planning engine to calculate the shortest path and the optimal detour path, inputs the lane-level road condition information, road segment congestion index, average passing speed, and passing time into a warning information generation model, and calculates the recommended safe vehicle speed according to the vehicle type and road traffic conditions to generate warning text content; The cloud platform extracts license plate information from the cache database of the linkage objects, establishes a corresponding relationship between the warning text content and the license plate information, generates a warning message including an addressing identifier, the cloud platform queries the communication address of the in-vehicle module, and sends the warning message to the corresponding in-vehicle module through a message queue. After receiving the warning message, the in-vehicle module calls a voice synthesis engine for broadcasting and displays the warning content on the in-vehicle screen.

[0013] In a second aspect, the present application provides a vehicle driving intelligent control device, including: A system construction module, configured to set up analysis and recognition units along the road. The analysis and recognition units include edge boxes, intelligent cameras, and environmental detection devices. The edge boxes establish a communication connection with the cloud platform and register. The intelligent cameras include license plate capture cameras and wide-angle cameras. The environmental detection devices collect road condition environment parameters; deploy an in-vehicle module at the vehicle end. The in-vehicle module establishes a connection with the cloud platform through an Internet of Things communication module and registers. The in-vehicle module calls a vehicle interface to obtain the vehicle driving trajectory, and divides the multiple analysis and recognition units into an area linkage network based on the driving trajectory; A traffic prediction module is used for the license plate capture camera to obtain the license plate number, license plate color and vehicle type information of the vehicle, the wide-angle camera to obtain the real-time road scene, the environment detection device to obtain the road visibility and road surface state parameters, the edge box to extract the vehicle driving characteristics based on the real-time road scene, the edge box to train a traffic flow prediction model and a risk assessment model based on the vehicle driving characteristics, and send the vehicle driving characteristics, the traffic flow prediction results, the risk assessment values and the environment parameters collected by each analysis and recognition unit to the cloud platform; An assisted driving module is used for the cloud platform to receive the data uploaded by multiple analysis and recognition units, judge whether to trigger regional linkage based on a preset road traffic strategy and the risk assessment value, when triggering regional linkage, send a linkage instruction to the relevant analysis and recognition units according to the regional linkage network and establish a communication link, receive the traffic flow prediction results uploaded by the relevant analysis and recognition units, and the cloud platform generates a warning message including real-time road conditions, recommended detour routes and safe vehicle speeds based on the traffic flow prediction results, and sends it to the in-vehicle module corresponding to the license plate number for broadcast.

[0014] In a third aspect, the present application provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the program, the steps of the vehicle driving intelligent control method are implemented.

[0015] In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the vehicle driving intelligent control method are implemented.

[0016] In a fifth aspect, the present application provides a computer program product, including a computer program / instructions, and when the computer program / instructions are executed by a processor, the steps of the vehicle driving intelligent control method are implemented.

[0017] As can be seen from the above technical solutions, the present application provides a vehicle driving intelligent control method and device, which realizes multi-dimensional data collection through an edge box, an intelligent camera and an environment detection device. Dynamically divides the regional linkage network based on the vehicle driving trajectory, and trains traffic flow prediction and risk assessment models at the edge. The cloud platform integrates the data of multiple analysis and recognition units, triggers the regional linkage mechanism according to the road traffic strategy and the risk assessment value, and generates accurate warning information including real-time road conditions, recommended detour routes and safe vehicle speeds. This method effectively solves the deficiencies of traditional technologies in aspects such as scene perception, data collaboration and warning decision-making, and significantly improves the intelligent level and service quality of the assisted driving system. Description of the Drawings

[0018] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0019] Figure 1 It is a schematic flowchart of the vehicle driving intelligent control method in the embodiments of the present application; Figure 2 It is a structural diagram of the vehicle driving intelligent control device in the embodiments of the present application; Figure 3 It is a schematic structural diagram of the electronic device in the embodiments of the present application.

[0020] Reference numerals: Electronic device 9600, central processing unit 9100, memory 9140, communication module 9110, input unit 9120, audio processor 9130, display 9160, power supply 9170, buffer memory 9141, application / function storage unit 9142, data storage unit 9143, driver program storage unit 9144, antenna 9111, speaker 9131, microphone 9132. Detailed implementation manners

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present application.

[0022] In the technical solutions of the present application, the acquisition, storage, use, processing, etc. of data all comply with the relevant regulations of national laws and regulations.

[0023] Considering the problems existing in the prior art, the present application provides a vehicle driving intelligent control method and device, which realizes multi-dimensional data collection through an edge box, an intelligent camera, and an environmental detection device. Based on the vehicle driving trajectory, a regional linkage network is dynamically divided, and traffic flow prediction and risk assessment models are trained at the edge. The cloud platform integrates the data of multiple analysis and recognition units, triggers the regional linkage mechanism according to the road traffic strategy and the risk assessment value, and generates accurate warning information including real-time road conditions, recommended detour routes, and safe vehicle speeds. This method effectively solves the deficiencies of traditional technologies in aspects such as scene perception, data collaboration, and warning decision-making, and significantly improves the intelligence level and service quality of the assisted driving system.

[0024] To effectively address the deficiencies of traditional technologies in aspects such as scene perception, data collaboration, and early warning decision-making, and significantly improve the intelligence level and service quality of the assisted driving system, this application provides an embodiment of a vehicle driving intelligent control method. Refer to Figure 1 , the vehicle driving intelligent control method specifically includes the following content: Step S101: Set up analysis and recognition units along the road. The analysis and recognition units include edge boxes, intelligent cameras, and environmental detection devices. The edge boxes establish a communication connection with the cloud platform and register. The intelligent cameras include license plate capture cameras and wide-angle cameras. The environmental detection devices collect road condition environment parameters; deploy in-vehicle modules on the vehicle side. The in-vehicle modules establish a connection with the cloud platform through the Internet of Things communication module and register. The in-vehicle modules call the vehicle machine interface to obtain the vehicle driving trajectory, and divide the multiple analysis and recognition units into a regional linkage network based on the driving trajectory; Optionally, in this embodiment, for major traffic arteries such as highways and urban expressways, analysis and recognition units are deployed at key road nodes. By analyzing the road geometric features and historical traffic data at locations such as accident-prone sections, viaduct areas, and tunnel entrances and exits, representative monitoring points are selected. A standardized pole structure is built at each monitoring point, and adjustable cantilever brackets are used to install the equipment to ensure that the installation height and angle of the equipment meet the monitoring requirements.

[0025] The edge box in this embodiment adopts an industrial-grade embedded computing platform, equipped with a high-performance ARM processor and an AI acceleration chip, supporting edge computing and deep learning algorithms. The edge box accesses the Internet of Things through a 4G / 5G dual-mode communication module, generates a registration key based on the unique device serial number, and establishes a secure communication link with the cloud platform using the TLS encryption protocol. During the registration process, the edge box reports information such as the device location coordinates and coverage range to the cloud platform, and the cloud platform assigns a unique identifier to it and returns an authorization certificate.

[0026] The intelligent camera in this embodiment includes two independent imaging units. The license plate capture camera uses a high-definition CCD sensor, equipped with an automatic aperture lens and a fill light, and can clearly capture license plate images under different lighting conditions. The wide-angle camera uses a fish-eye lens, covering a 180-degree field of view, realizing panoramic road monitoring. The two cameras are calibrated in terms of angle and calibrated for spatial mapping to establish an image coordinate system, realizing multi-target tracking and spatial positioning.

[0027] The environmental detection equipment of this embodiment integrates multiple sensors, including a laser visibility meter, a surface temperature sensor, a humidity sensor, and a light sensor. The laser visibility meter measures the atmospheric transmittance based on the scattering principle and calculates the visibility value. The surface sensor array is embedded in the road surface to monitor the road surface temperature distribution and water accumulation in real time. All sensor data are connected to the edge box through the industrial bus to achieve unified management and data fusion.

[0028] This embodiment deploys a standardized vehicle-mounted module on the vehicle side, adopts a plug-and-play design, and is connected to the vehicle CAN bus through the OBD interface. The vehicle-mounted module has a built-in high-precision positioning chip and an inertial measurement unit, integrating multi-mode positioning signals such as GPS and Beidou to achieve sub-meter positioning accuracy. A data channel with the cloud platform is established through the Internet of Things communication module, and the lightweight MQTT protocol is used to transmit vehicle status data.

[0029] The vehicle-mounted module of this embodiment calls the vehicle system through a standardized interface to obtain vehicle condition data such as engine speed, vehicle speed, steering wheel angle, etc. Combined with high-precision positioning information, a vehicle motion state model is constructed to generate an accurate driving trajectory. The vehicle-mounted module uploads the trajectory data to the cloud platform at fixed time intervals. The cloud platform maps the trajectory points to the road network based on the electronic map and predicts the vehicle's future driving path.

[0030] This embodiment innovatively proposes a method for dynamically constructing a regional linkage network. The cloud platform extracts analysis and identification units along the route based on the predicted vehicle path, and constructs a directed graph structure based on geographic spatial relationships. The nodes in the graph represent analysis and identification units, and the edges represent the associations between units. By assigning network priorities and establishing a cascade trigger mechanism, when a node detects an anomaly, it can quickly notify downstream nodes to respond in a linkage manner.

[0031] This embodiment adopts a distributed edge computing architecture to delegate computing tasks such as data collection and feature extraction to edge nodes to reduce the load on the cloud platform. Both the edge box and the vehicle-mounted module have local computing capabilities, which can realize data preprocessing and preliminary analysis. At the same time, through unified cloud platform scheduling, coordination between devices is achieved to build an intelligent perception network covering the entire road area.

[0032] This embodiment solves the problems of scattered equipment, data islands, and delayed response in traditional road monitoring systems. Through edge computing and regional linkage mechanisms, real-time perception and rapid warning of road conditions are achieved. In practical applications, this solution can significantly improve traffic safety in scenarios such as severe weather and frequent accidents. Especially in closed road environments such as highways, the system can provide drivers with accurate road condition information and safety warnings, effectively reducing the risk of traffic accidents.

[0033] Step S102: The license plate capture camera obtains the license plate number, license plate color, and vehicle type information of the vehicle. The wide-angle camera obtains the real-time road scene. The environmental detection device obtains the road visibility and road surface state parameters. The edge box extracts the vehicle driving characteristics based on the real-time road scene. The edge box trains a traffic flow prediction model and a risk assessment model based on the vehicle driving characteristics, and sends the vehicle driving characteristics, traffic flow prediction results, risk assessment values, and environmental parameters collected by each analysis and recognition unit to the cloud platform; Optionally, in this embodiment, a multi-level processing strategy is adopted in the license plate capture link. The license plate capture camera is triggered by vehicle front-end detection and uses adaptive exposure technology to ensure image quality. The edge box runs an improved YOLOv5 object detection algorithm to locate the license plate area and combines morphological processing to achieve license plate character segmentation. Character recognition is performed through a deep convolutional neural network, and the model is trained using a synthetic dataset with noise to improve the recognition accuracy in complex environments. At the same time, vehicle type classification is performed based on the body contour features, supporting accurate recognition of various vehicle types such as sedans, buses, and trucks.

[0034] The wide-angle camera in this embodiment obtains the real-time road scene at a sampling rate of 25 frames per second. The edge box performs real-time processing on the video stream and uses an improved DeepSORT algorithm to achieve multi-object tracking. The algorithm fuses appearance features and motion features in the feature extraction stage to improve the stability of vehicle tracking. Vehicle motion parameters, including instantaneous speed, acceleration, lane change frequency, etc., are calculated through the tracking trajectory. At the same time, the relative distance and relative speed between vehicles are calculated based on the spatial mapping relationship to construct vehicle interaction behavior characteristics.

[0035] The environmental detection device in this embodiment adopts a multi-sensor collaborative perception scheme. The laser visibility meter calculates the visibility value by measuring the atmospheric attenuation coefficient, and the sampling period is 1 minute. The surface sensor array monitors the road surface temperature distribution in real time and combines humidity data to judge the road surface state, such as dry, wet, waterlogged, icy, etc. The light sensor monitors the environmental brightness change for adjusting the camera parameters and the light supplement strategy. All environmental parameters are filtered and smoothed to ensure data quality.

[0036] This embodiment innovatively designs a traffic flow prediction model. The model adopts a long short-term memory network (LSTM) structure, and the input features include time series data such as historical traffic flow, average vehicle speed, and vehicle type composition. The attention mechanism is used to capture the traffic flow change patterns at different time scales to improve the prediction accuracy. The model training adopts a sliding window strategy, and the window size is dynamically adjusted according to the prediction time span. The prediction results include the traffic flow states at multiple future time points, providing a decision-making basis for traffic control.

[0037] In this embodiment, a risk assessment model based on multi-dimensional features is constructed. The gradient boosting decision tree (GBDT) algorithm is adopted, and the input features include behavioral features such as vehicle distance, relative speed, acceleration, and lane change, as well as environmental features such as visibility and road surface conditions. The model analyzes historical accident data to establish the mapping relationship between features and risk levels. The evaluation result is output in the form of probability, intuitively reflecting the safety risk level of the current road section.

[0038] In this embodiment, an efficient data transmission mechanism is designed. The edge box packages vehicle features, prediction results, and risk assessment values into structured data, and uses the protobuf encoding format to reduce transmission overhead. The data packet is uploaded to the cloud platform at a fixed period, and the upload period can be dynamically adjusted according to the network condition. For emergencies, a real-time push mechanism is adopted to ensure that key information is delivered in time. The cloud platform synchronizes and performs correlation analysis on the data through timestamps.

[0039] In this embodiment, a computing architecture that combines edge intelligence and cloud collaboration is implemented. Tasks with high real-time requirements such as feature extraction and target tracking are deployed on the edge box to reduce network transmission latency. Model training and optimization are carried out in the cloud, and the updated model parameters are regularly sent to the edge nodes. This hierarchical architecture not only ensures the real-time response ability of the system but also supports the continuous optimization and evolution of the model.

[0040] In this embodiment, through multi-source data fusion and intelligent analysis, problems existing in traditional traffic monitoring systems such as insufficient perception ability and low prediction accuracy are solved. In practical applications, this solution can accurately identify vehicle features, predict the change trend of traffic flow, and timely discover potential safety hazards. Especially in scenarios such as bad weather and complex road conditions, the system can comprehensively analyze multi-dimensional data, provide reliable decision-making support for traffic management and vehicle warning, and effectively improve road traffic efficiency and safety.

[0041] Step S103: The cloud platform receives the data uploaded by multiple analysis and recognition units, determines whether to trigger area linkage based on a preset road traffic strategy and the risk assessment value. When area linkage is triggered, a linkage instruction is sent to the relevant analysis and recognition units according to the area linkage network and a communication link is established. The cloud platform receives the traffic flow prediction results uploaded by the relevant analysis and recognition units, generates warning information including real-time road conditions, recommended detour routes, and safe vehicle speeds based on the traffic flow prediction results, and sends it to the in-vehicle module corresponding to the license plate number for broadcasting.

[0042] Optionally, this embodiment innovatively realizes the intelligent data acquisition and processing process at the edge. The license plate capture camera realizes license plate positioning through a two-stage detection mechanism based on a method combining image processing and deep learning. In the first stage, an improved YOLOv5 model is used for rough positioning, and in the second stage, a high-precision license plate character segmentation network is adopted to achieve accurate recognition at the character level. It can still maintain a high recognition accuracy under complex lighting and occlusion conditions.

[0043] The wide-angle camera in this embodiment adopts a multi-scale object detection algorithm to perform real-time detection and tracking of key objects such as vehicle contours, pedestrians, and road markings. Through spatio-temporal feature analysis, driving features such as the instantaneous speed, acceleration, and lane departure of the vehicle are calculated. The edge box adopts a stream processing architecture to establish a feature extraction pipeline to achieve real-time processing of the video stream, with the delay controlled at the millisecond level.

[0044] This embodiment constructs a traffic flow prediction model based on the long short-term memory network (LSTM). The input features of the model include time series data such as historical traffic flow, average vehicle speed, and vehicle density. At the same time, environmental factors such as weather conditions and road surface conditions are introduced. The attention mechanism is used to capture the correlation between different features to improve the prediction accuracy. The prediction results include the traffic flow states at multiple future time points, providing a basis for predictive decision-making in traffic management.

[0045] This embodiment designs a risk assessment model based on random forest. The input variables of the model include multi-dimensional features such as vehicle spacing, relative speed, weather visibility, and road surface friction coefficient. Through feature importance analysis, key factors that have a significant impact on traffic safety are identified. The risk assessment value output by the model reflects the probability of a traffic accident occurring in the current section and is used to trigger the warning mechanism. The model is continuously optimized through online learning in practical applications to adapt to the safety features of different scenarios.

[0046] This embodiment optimizes the data transmission strategy. The edge box adopts a hierarchical caching mechanism to store the original data, feature data, and model prediction results respectively. Data upload adopts a batch processing mode, and the batch size is dynamically adjusted according to data priority and network conditions. Through data compression and incremental update mechanisms, the occupancy of the transmission bandwidth is reduced.

[0047] This embodiment realizes an intelligent regional linkage mechanism based on the road traffic strategy. After receiving the data uploaded by each analysis and recognition unit, the cloud platform first performs data quality assessment and outlier detection. When the risk assessment value of a certain section exceeds the threshold, the regional linkage judgment is triggered. The linkage range is determined based on the traffic flow propagation model, considering factors such as traffic flow direction and traffic saturation.

[0048] In this embodiment, an early warning information generation engine is constructed, and the congestion index of each road section is calculated based on the traffic flow prediction results. Through the path planning algorithm, considering factors such as distance, time, and safety, personalized detour suggestions are generated for different types of vehicles. The recommended safe speed is calculated based on the current road conditions, weather conditions, and vehicle type characteristics to ensure driving safety.

[0049] In this embodiment, an accurate information push mechanism is implemented. The early warning information is sent to the in-vehicle module through the message queue system, supporting message priority management and reliable transmission. After receiving the early warning information, the in-vehicle module selects an appropriate broadcast time according to the vehicle driving state, and converts the text information into clear voice prompts through voice synthesis technology.

[0050] Through the above technological innovations, this embodiment effectively solves the problems existing in traditional traffic early warning systems, such as perception blind spots, prediction lags, and untimely linkages. In practical applications, this solution can accurately identify traffic risks, predict traffic flow changes in advance, and achieve rapid response through the regional linkage mechanism. Especially in scenarios such as bad weather and accident-prone areas, the system can provide accurate early warning information and safety suggestions for drivers, significantly improving the safety and efficiency of road traffic.

[0051] As can be seen from the above description, the intelligent vehicle driving control method provided by the embodiment of the present application can realize multi-dimensional data collection through the edge box, intelligent camera, and environmental detection device. Dynamically divide the regional linkage network based on the vehicle driving trajectory, and train the traffic flow prediction and risk assessment models at the edge. The cloud platform integrates the data of multiple analysis and recognition units, triggers the regional linkage mechanism according to the road traffic strategy and risk assessment value, and generates accurate early warning information including real-time road conditions, recommended detour routes, and safe speeds. This method effectively solves the deficiencies of traditional technologies in scenario perception, data collaboration, and early warning decision-making, and significantly improves the intelligent level and service quality of the assisted driving system.

[0052] In an embodiment of the intelligent vehicle driving control method of the present application, the following specific content may also be included: Step S201: Select key positions on the road to deploy the analysis and recognition unit, perform parameter initialization configuration on the edge box. The edge box obtains a unique identification code through the Internet of Things communication module, sends a registration request to the cloud platform based on the unique identification code, and the cloud platform returns registration authorization information. The edge box establishes a secure communication link with the cloud platform based on the registration authorization information; Step S202: The edge box establishes communication connections with the intelligent camera and the environment detection device, calibrates the angle and focal length of the intelligent camera, sets the license plate capture camera to point to the vehicle passing direction, sets the wide-angle camera to the road panoramic monitoring angle, establishes a sub-device management group centered on the edge box, configures the sampling period and trigger threshold of the environment detection device, and the edge box obtains environmental parameters including visibility, temperature, humidity, and light intensity from the environment detection device at preset time intervals.

[0053] Optionally, in this embodiment, key monitoring positions are identified by analyzing historical traffic accident data and road network structure characteristics. Special sections such as accident-prone sections, sharp curves, steep slopes, tunnel entrances and exits, and overpasses are mainly considered, and the optimal deployment positions of the analysis and identification units are determined in combination with the traffic flow distribution characteristics. A standardized pole structure is established at each monitoring point, and the equipment is installed using an adjustable bracket to ensure that the monitoring perspective completely covers the target area.

[0054] In this embodiment, comprehensive parameter initialization configuration is performed on the edge box. First, set the device basic parameters, including processor frequency, memory allocation, storage strategy, etc. The network configuration adopts a dual-link redundancy design, supporting both wired network and 4G / 5G wireless network at the same time to ensure communication reliability. The edge box obtains a unique identification code based on the hardware characteristics through the Internet of Things communication module, and this identification code generates a globally unique device ID in combination with the geographical location information and device type.

[0055] In this embodiment, a secure device registration mechanism is designed. The edge box generates a registration request message based on the unique identification code and signs the request content using an asymmetric encryption algorithm. After receiving the registration request, the cloud platform verifies the legitimacy of the device identity and allocates resource quotas. The registration authorization information includes content such as access tokens, encryption keys, and service configurations. The edge box establishes a TLS encryption channel based on the authorization information to achieve secure communication with the cloud platform.

[0056] In this embodiment, precise calibration of the intelligent camera is achieved. The license plate capture camera uses a two-axis pan-tilt bracket and automatically adjusts the shooting angle and focal length through an image quality evaluation algorithm. During the calibration process, a standard test license plate is used to collect image samples under different lighting conditions to optimize the image processing parameters. The wide-angle camera uses a special calibration board to establish an image coordinate system through corner detection and spatial mapping.

[0057] In this embodiment, a device management system based on the edge box is constructed. The edge box serves as a local control center and accesses the intelligent camera and the environment detection device through an industrial Ethernet. The main-slave communication architecture is adopted between devices, and the edge box regularly sends heartbeat packets to the sub-devices to detect the online status. When a device anomaly is detected, it automatically switches to the backup communication link to ensure the continuity of data collection.

[0058] This embodiment optimizes the configuration strategy of environmental detection devices. The sampling period of the visibility sensor is dynamically adjusted according to the speed of weather change, and the sampling frequency is increased under harsh weather conditions such as foggy days. The temperature and humidity sensor adopts a threshold trigger mechanism, and reports data immediately when the parameter change exceeds the preset threshold. The parameters of the light sensor and the camera are linked to adjust the image gain and exposure time in real time.

[0059] This embodiment designs an intelligent data acquisition scheduling mechanism. The edge box sets different sampling strategies for different types of environmental parameters according to business requirements and device characteristics. For example, visibility data is sampled once a minute, temperature and humidity data is sampled once every 5 minutes, and the sampling frequency of light intensity is dynamically adjusted according to sunrise and sunset times. The collected environmental parameters are filtered and outliers are detected to ensure data quality.

[0060] This embodiment realizes the remote management function of device configuration. The cloud platform can dynamically adjust the configuration parameters of the edge box and sub-devices according to actual needs. The configuration update adopts an incremental method, and only the changed configuration items are sent down, reducing network transmission overhead. After receiving the configuration update, the edge box ensures atomic update of the configuration through a transaction mechanism to avoid configuration inconsistency.

[0061] Through the above technological innovations, this embodiment solves the problems existing in the deployment and management of traditional road monitoring devices. In practical applications, this solution realizes the plug-and-play and intelligent management of devices, significantly improving the reliability and maintenance efficiency of the system. Especially under complex road conditions and harsh weather conditions, the system can operate stably and provide reliable data support for traffic monitoring and early warning.

[0062] In an embodiment of the intelligent vehicle driving control method of the present application, the following content may also be specifically included: Step S301: The in-vehicle module obtains the device unique identifier through the Internet of Things communication module, sends a registration request including the vehicle VIN code to the cloud platform based on the device unique identifier, the cloud platform returns the device authorization code, and the in-vehicle module establishes a data transmission channel with the cloud platform based on the device authorization code. The in-vehicle module calls the vehicle CAN bus interface to collect vehicle real-time position coordinates, driving speed, and direction angle data; Step S302: The in-vehicle module uploads the position coordinates, driving speed, and direction angle data to the cloud platform at a preset time interval. The cloud platform maps the vehicle position coordinates to an electronic map based on the geographic information system, determines the forward prediction path according to the vehicle driving direction, extracts the analysis and recognition units within the coverage of the prediction path, constructs a regional linkage network with the extracted analysis and recognition units, and assigns network priorities to each analysis and recognition unit.

[0063] Optionally, this embodiment realizes the intelligent deployment and automatic registration process of the in-vehicle module. The in-vehicle module adopts a plug-and-play design, connects to the vehicle CAN bus through the OBD interface, automatically identifies the vehicle model, and loads the corresponding protocol stack. The Internet of Things communication module generates a unique device identifier based on the built-in security chip, which is bound to the hardware characteristics to ensure the uniqueness and immutability of the device identity.

[0064] This embodiment designs a secure device registration mechanism. The in-vehicle module obtains the VIN code by scanning the vehicle electronic control unit, and combines the VIN code with the unique device identifier to generate a registration certificate. The registration request is signed using an asymmetric encryption algorithm to ensure the authenticity of the request. After verifying the registration information, the cloud platform generates a device authorization code containing permission information and an encryption key to achieve secure access to the device.

[0065] This embodiment constructs a reliable data transmission channel. The in-vehicle module establishes a TLS encrypted communication based on the device authorization code, supports data compression and resume from breakpoint. The channel adopts a two-way authentication mechanism to ensure the security of data transmission through session key negotiation. At the same time, the network switching function is realized, and it automatically switches to the backup network when the 4G signal is unstable to ensure the continuity of communication.

[0066] This embodiment optimizes the vehicle data acquisition strategy. The in-vehicle module obtains the vehicle status information in real time through the CAN bus interface, and adopts a multi-thread parallel processing mechanism to ensure the real-time nature of data acquisition. The position coordinates are obtained through the GPS and Beidou dual-mode positioning module, and data fusion is performed in combination with the inertial measurement unit to improve the positioning accuracy. The vehicle speed and direction angle data are directly obtained from the vehicle sensors to ensure the data accuracy.

[0067] This embodiment realizes an intelligent data upload mechanism. The in-vehicle module dynamically adjusts the data upload frequency according to the vehicle motion state, increases the sampling rate when the vehicle turns or accelerates / decelerates, and appropriately reduces the sampling rate when the vehicle is stationary or moving at a constant speed. When packing data, an incremental coding method is adopted to only transmit the changed data items, reducing the bandwidth occupancy. At the same time, a local caching mechanism is realized to temporarily store data when the network is interrupted and automatically retransmit it when the network resumes.

[0068] This embodiment innovatively designs a position mapping algorithm. The cloud platform uses a high-precision electronic map and quickly locates the road section where the vehicle is located through a multi-level index structure. The position mapping takes into account the road geometric features and traffic rules, and can accurately identify the lane where the vehicle is located. Based on the historical trajectory data and the current direction angle, the Kalman filtering algorithm is used to predict the future driving path of the vehicle, and the prediction range is dynamically adjusted according to the vehicle speed.

[0069] In this embodiment, an adaptive regional linkage network is constructed. The cloud platform extracts relevant analysis and recognition units based on the predicted path and constructs a directed acyclic graph structure using graph theory algorithms. Nodes in the graph represent analysis and recognition units, and edges represent the association relationships between units. The network priority assignment takes into account multiple factors, including the distance from the predicted path, road grade, historical accident frequency, etc. The priority value is obtained through weighted calculation and is used for subsequent linkage trigger judgment.

[0070] This embodiment realizes the dynamic update of the network topology. As the vehicle position changes, the cloud platform adjusts the coverage area of the regional linkage network in real time. The sliding window mechanism is adopted to add the analysis and recognition units that are about to enter the predicted path to the network in advance, and at the same time remove the units that have moved far away. This dynamic adjustment mechanism ensures that the network structure is always synchronized with the vehicle driving path.

[0071] Through the above technological innovations, this embodiment effectively solves the problems existing in the management of traditional in-vehicle devices and path prediction. In practical applications, this solution realizes the intelligent access of in-vehicle devices and the reliable transmission of data, providing an accurate location basis for subsequent traffic warnings. Especially in a complex road network environment, the system can accurately predict the vehicle driving path and establish an efficient regional linkage mechanism, providing strong support for proactive safety warnings.

[0072] In an embodiment of the intelligent vehicle driving control method of this application, the following content may also be specifically included: Step S401: The license plate capture camera collects image data when the vehicle passes by. The edge box processes the image data based on a convolutional neural network model, extracts license plate area features and performs character segmentation, identifies the license plate number and color information, and obtains the vehicle type classification result by analyzing the body contour features through a deep learning model. The license plate number, color information, and vehicle type classification result are stored in the cache database of the edge box. Step S402: The wide-angle camera collects the road traffic video stream at a preset frame rate. The environmental detection device obtains the visibility value based on the principle of laser ranging, and detects the road surface state parameters through the surface temperature sensor and humidity sensor. The edge box performs target detection and tracking processing on the video stream, extracts driving feature parameters such as vehicle passing trajectory, vehicle speed, and vehicle spacing. The edge box establishes a mapping relationship between the driving feature parameters and the license plate information.

[0073] Optionally, this embodiment adopts a multi-level processing strategy in the license plate recognition link. First, the license plate capture camera adopts adaptive exposure technology to dynamically adjust the exposure parameters according to the environmental light conditions to ensure image quality. When a vehicle is detected entering the field of view, high-speed capture is triggered to collect multiple frames of image data, and the most clear image is selected for processing through a motion compensation algorithm.

[0074] In this embodiment, an improved license plate detection network is designed. The convolutional neural network model based on the YOLOv5 framework adopts a multi-scale feature fusion strategy to enhance the detection ability for license plates at different distances. A spatial pyramid pooling module is added to the network structure to enhance the adaptability to license plate deformation and inclination. The model training uses a license plate image dataset containing various lighting and weather conditions to improve the generalization ability of the model.

[0075] In this embodiment, an accurate character segmentation algorithm is implemented. An improved FCN semantic segmentation network is used to achieve pixel segmentation at the character level. By introducing an attention mechanism, the ability to distinguish similar characters is enhanced. In the character recognition stage, a deep residual network is combined with the CTC loss function to improve the recognition accuracy. At the same time, the license plate background color is determined through color histogram analysis to achieve accurate classification of license plate types.

[0076] In this embodiment, a vehicle type classification method is innovatively designed. The body contour features are extracted through a deep learning model, including key features such as the aspect ratio of the body, the shape of the front of the vehicle, and the window layout. A multi-task learning framework is adopted to simultaneously predict the vehicle type, brand, and model to improve the classification accuracy. The model adopts a transfer learning strategy to use a pre-trained model to accelerate the training convergence speed.

[0077] In this embodiment, an efficient cache database is constructed. A key-value storage structure is adopted, with the license plate number as the primary key to store vehicle-related feature information. The database supports concurrent access and transaction processing to ensure data consistency. The memory data is managed through the LRU cache strategy, and the oldest unaccessed records are automatically cleared when the capacity limit is exceeded.

[0078] In this embodiment, the video stream processing flow is optimized. The wide-angle camera uses a sampling rate of 25 frames per second, and video compression is achieved through a hardware encoder. The edge box adopts a multi-threaded parallel processing architecture to process multiple video streams simultaneously. The target detection uses an improved DeepSORT algorithm to achieve stable tracking of vehicles. The mapping relationship between pixel coordinates and actual distances is established through projective transformation to calculate the actual position and speed of the vehicle.

[0079] In this embodiment, the intelligent acquisition of environmental parameters is realized. The laser ranging device calculates the visibility value by measuring the attenuation degree of the laser signal, and the sampling period is dynamically adjusted according to the weather change speed. The surface sensor array measures the road surface temperature distribution through the thermocouple principle, and combines the humidity data to judge the road surface state, such as dry, wet, waterlogged, frozen and other conditions.

[0080] In this embodiment, a feature mapping mechanism is innovatively established. Through a spatio-temporal correlation algorithm, a corresponding relationship is established between the vehicle tracking trajectory and the license plate information. The algorithm takes into account the continuity of vehicle movement and physical constraints, effectively solving the association problem in case of occlusion and dense traffic. For vehicles that temporarily lose license plate information, the tracking state is maintained through trajectory prediction.

[0081] This embodiment designs a method for extracting feature parameters. Based on the tracking trajectory, kinematic features of the vehicle are calculated, including instantaneous speed, acceleration, steering angular velocity, etc. The vehicle spacing and relative speed are calculated through the positional relationship of adjacent vehicles to evaluate potential collision risks. All feature parameters are smoothed and outliers are filtered to ensure data quality.

[0082] Through the above technological innovations, this embodiment solves the limitations of traditional traffic monitoring systems in vehicle identification and behavior analysis. In practical applications, this solution can accurately identify vehicle features, real-time track the vehicle movement state, and provide reliable data support for traffic flow prediction and risk assessment. Especially under complex road conditions and adverse weather conditions, the system can still maintain stable identification and tracking performance, significantly improving the accuracy and reliability of traffic monitoring.

[0083] In an embodiment of the intelligent vehicle driving control method of this application, the following content may also be specifically included: Step S501: The edge box divides the vehicle driving feature parameters into a training data set according to the time series, constructs the network structure of the traffic flow prediction model using a long short-term memory neural network, uses vehicle speed, vehicle density, and passing time as input features, iteratively trains the network structure to obtain a traffic flow prediction model, constructs a risk assessment model based on the decision tree algorithm, and calculates the collision risk assessment value using vehicle spacing, relative speed, and road surface state parameters as input variables; Step S502: The edge box organizes the vehicle driving features, environmental parameters, and timestamp information into a data packet. The data packet contains the predicted result of the future traffic flow state output by the traffic flow prediction model and the risk assessment value calculated by the risk assessment model. The edge box sends the data packet to the cloud platform through a secure communication link according to a preset data upload period, and the cloud platform performs time synchronization and data verification on the received data packet.

[0084] Optionally, this embodiment innovatively designs a data preprocessing process. The edge box adopts a sliding window mechanism to organize the continuously collected vehicle driving features into training samples in chronological order. The window size is dynamically adjusted according to the prediction time span. For example, a 5-minute window is used for short-term prediction and a 15-minute window is used for medium-term prediction. In the data preprocessing stage, the feature scale is unified through the maximum-minimum normalization method, and median filtering is used to remove outliers.

[0085] In this embodiment, a traffic flow prediction model is constructed based on the Long Short-Term Memory network (LSTM). The network structure includes an input layer, an LSTM layer, a fully connected layer, and an output layer. The input features include three key indicators: vehicle speed, vehicle density, and travel time, which can comprehensively reflect the road traffic status. The LSTM layer captures the long-term dependencies of time series data through a gating mechanism. The forget gate can filter out irrelevant information, and the input gate and output gate control the update and output of information.

[0086] This embodiment optimizes the model training strategy. The batch gradient descent method is used for parameter optimization, and the learning rate adopts an adaptive adjustment strategy. A larger learning rate is used at the beginning of training for rapid convergence, and the learning rate is gradually reduced in the later stage for fine-tuning. To prevent overfitting, a Dropout layer is introduced to randomly discard some neurons, and L2 regularization is used to constrain the model parameters. The mean squared error is used as the loss function, and the gradient is calculated through the backpropagation algorithm to update the parameters.

[0087] This embodiment implements a risk assessment model based on decision trees. The CART algorithm is used to construct a binary decision tree, and the input variables include vehicle spacing, relative speed, and road surface state parameters. The splitting criterion of the decision tree adopts the Gini index, and the optimal splitting feature and threshold are selected recursively. The predicted value of the leaf node represents the collision risk probability, and the value range is from 0 to 1. The model is trained using historical accident data with labels, and the optimal tree depth is determined through cross-validation.

[0088] This embodiment innovatively designs a risk assessment method. The risk assessment value R is calculated by the following formula: R = w1 D + w2 V + w3 S where D is the normalized vehicle spacing, V is the relative speed, S is the road surface state coefficient, and w1, w2, and w3 are weight coefficients. The weight coefficients are determined through historical data analysis and reflect the influence degree of each factor on safety risks. When the R value exceeds the preset threshold, a risk warning is triggered.

[0089] This embodiment optimizes the data packet organizational structure. A hierarchical design is adopted. The bottom layer is the basic data field, including device ID, timestamp, location coordinates, etc.; the middle layer is the feature data field, containing vehicle driving features and environmental parameters; the top layer is the analysis result field, containing traffic flow prediction results and risk assessment values. The data packet is encoded in a compact binary format to reduce transmission overhead.

[0090] This embodiment implements an intelligent data upload strategy. Under normal circumstances, data packets are uploaded at fixed intervals, and the intervals can be dynamically adjusted according to network conditions. When abnormal situations are detected, such as a sudden increase in the risk assessment value or a significant change in traffic flow shown by the prediction results, a real-time upload mechanism is triggered. The data transmission uses the TLS encryption protocol to ensure data security.

[0091] This embodiment designs a data synchronization and verification mechanism. After receiving the data packets, the cloud platform first checks the continuity of the timestamps and reorders the out-of-order packets. The data integrity is verified through CRC check codes, and packets that fail the verification are required to be retransmitted. At the same time, a clock synchronization mechanism is implemented to ensure the time consistency between the edge box and the cloud platform.

[0092] Through the above technological innovations, this embodiment solves the problems existing in traditional traffic prediction systems, such as low prediction accuracy and inaccurate risk assessment. In practical applications, this solution can accurately predict the change trend of traffic flow and timely identify potential safety risks. Especially in complex road conditions, the system can comprehensively analyze multi-dimensional data, provide a reliable basis for traffic management decisions, and effectively improve the safety and efficiency of road traffic.

[0093] In an embodiment of the intelligent vehicle driving control method of this application, the following content may also be specifically included: Step S601: The cloud platform receives the data packets uploaded by multiple analysis and identification units, extracts the index parameters such as the risk assessment value, traffic flow, and vehicle speed from the data packets, compares the index parameters with the thresholds in the road traffic strategy, and the road traffic strategy includes the traffic flow density threshold, average vehicle speed threshold, and accident risk threshold per unit mileage. When the index parameters of any analysis and identification unit exceed the corresponding threshold, area linkage judgment is triggered, and the cloud platform determines the affected range according to the position of the analysis and identification unit that triggers the linkage; Step S602: The cloud platform selects the analysis and identification units located within the affected range from the area linkage network as linkage objects, generates a linkage session including the linkage object identification codes, creates a data sharing channel based on the linkage session, sends a linkage instruction including the shared channel configuration information to the linkage objects, and receives the confirmation response returned by the linkage objects through the data sharing channel to complete the establishment of the communication link between the linkage objects.

[0094] Optionally, this embodiment implements an intelligent data processing flow. The cloud platform uses a distributed message queue to receive the data packets uploaded by multiple analysis and identification units, and realizes reliable data transmission and load balancing through a message middleware. After the data packets arrive, they are first parsed and verified, and key index parameters including the risk assessment value, traffic flow, and average vehicle speed are extracted. These parameters are processed through data cleaning and standardization to ensure data consistency.

[0095] This embodiment designs an adaptive threshold judgment mechanism. The threshold in the road traffic strategy is not fixed, but dynamically adjusted according to road grades, time period characteristics, and historical data. The traffic flow density threshold D is calculated by the following formula: D = Dbase (1 + k1 T + k2 W) where Dbase is the reference density threshold, T is the time period coefficient, W is the weather influence coefficient, and k1 and k2 are adjustment factors. This dynamic threshold mechanism can better adapt to the requirements of different scenarios.

[0096] This embodiment innovatively implements a regional linkage judgment algorithm. When the index parameter of an analysis and recognition unit exceeds the threshold, the system calculates the influence range based on the traffic flow propagation model. The propagation model takes into account factors such as the road network topology, traffic flow direction, and propagation speed. The road network connectivity is analyzed through graph theory algorithms to identify the associated road segments that may be affected. The calculation of the affected range uses a spatio-temporal diffusion model, and the size of the range is positively correlated with the degree of over-limit and the duration.

[0097] This embodiment constructs an efficient linkage object selection mechanism. The cloud platform extracts the analysis and recognition units within the affected range from the pre-established regional linkage network and filters them based on priority and location relationships. The selection process considers multiple dimensions such as device performance, network status, and historical reliability to ensure the selection of the most suitable linkage objects. At the same time, a backup mechanism is implemented, and standby devices are configured for key nodes to improve the system reliability.

[0098] This embodiment optimizes the linkage session management process. Each linkage session is assigned a globally unique session ID, which contains information such as trigger conditions, participating objects, and validity periods. The session management adopts a distributed architecture and supports parallel processing of multiple sessions. The session life cycle is managed through a state machine mechanism to ensure that the creation, maintenance, and release processes of the session are controllable and traceable.

[0099] This embodiment designs a reliable data sharing channel. A full-duplex communication link is established based on the WebSocket protocol, supporting real-time data push and instruction issuance. The channel configuration adopts a hierarchical design, including basic configuration and extended configuration. The basic configuration ensures the basic connectivity of the channel, and the extended configuration is dynamically loaded according to business requirements. The channel creation adopts a handshake mechanism to ensure that the participating parties reach an agreement on communication parameters.

[0100] This embodiment realizes an intelligent instruction distribution strategy. The linkage instructions are processed hierarchically according to priorities, and the emergency instructions are distributed first. The instruction format adopts a unified protocol specification, including fields such as instruction type, parameter configuration, execution time, etc. The reliable transmission of instructions is realized through the message queue mechanism, and the confirmation and retransmission mechanisms of instructions are supported. At the same time, the real-time monitoring of the instruction execution status is realized to discover and handle abnormal situations in a timely manner.

[0101] This embodiment constructs a complete confirmation response mechanism. After receiving the instruction, the linkage object first verifies the legality and integrity of the instruction. Then, it evaluates the executability of the instruction according to the local resource status and generates a response message containing the execution plan. The response message is returned to the cloud platform through the data sharing channel, and the cloud platform confirms the establishment of the linkage relationship accordingly. For the devices that fail to respond in time, a timeout processing mechanism is triggered.

[0102] Through the above technological innovations, this embodiment solves the problems of untimely regional linkage and low collaborative efficiency in traditional traffic management systems. In practical applications, this solution can quickly respond to traffic anomalies and establish an efficient device collaboration mechanism. Especially in emergency situations such as traffic accidents and bad weather, the system can quickly organize relevant devices for linkage response, providing strong support for traffic guidance and safety warning, and significantly improving the intelligent management level of road traffic.

[0103] In an embodiment of the intelligent vehicle driving control method of this application, the following content may also be specifically included: Step S701: The cloud platform receives the traffic flow prediction result uploaded by the linkage object, maps the traffic flow prediction result to the road section based on the real-time position coordinates, calculates the congestion index of each section, the cloud platform calls the path planning engine to calculate the shortest path and the optimal detour path, inputs the lane-level road condition information, section congestion index, average traffic speed, and travel time into the warning information generation model, calculates the recommended safe speed according to the vehicle type and road traffic conditions, and generates the warning text content; Step S702: The cloud platform extracts the license plate information from the cache database of the linkage object, establishes a corresponding relationship between the warning text content and the license plate information, generates a warning message containing the addressing identifier, the cloud platform queries the communication address of the in-vehicle module, and sends the warning message to the corresponding in-vehicle module through the message queue. After receiving the warning message, the in-vehicle module calls the speech synthesis engine for broadcasting and displays the warning content on the car machine screen.

[0104] Optionally, this embodiment implements an intelligent traffic flow analysis mechanism. After receiving the prediction results uploaded by the associated objects, the cloud platform first conducts spatio-temporal consistency checks to ensure the validity of the data. The prediction results are mapped to specific road segments through a high-precision electronic map to establish the association between the prediction data and the road network topology. The congestion index CI is calculated by the following formula: CI = (V0 - V) / V0 (ρ / ρmax) where V0 is the free flow speed, V is the current average speed, ρ is the current traffic density, and ρmax is the maximum traffic density. This index comprehensively reflects the road traffic state.

[0105] This embodiment optimizes the path planning algorithm. Based on the improved algorithm, shortest path search is implemented. The congestion index is used as the road segment weight factor to dynamically adjust the path evaluation criteria. When calculating the optimal detour path, factors such as distance increment, time increment, and road condition reliability are comprehensively considered. Through a multi-objective optimization method, the optimal solution is selected from multiple candidate paths to ensure the practicality of the detour recommendation.

[0106] This embodiment innovatively designs a warning information generation model. The model adopts a deep neural network structure, and the input features include multi-dimensional data such as lane-level road conditions, congestion index, and average speed. Key information is identified through an attention mechanism to generate warning texts that conform to human cognitive habits. A large amount of real-scene data is used for model training to ensure the accuracy and comprehensibility of the generated content.

[0107] This embodiment implements an intelligent safe vehicle speed calculation method. The benchmark speed is set according to the vehicle type and dynamically adjusted in combination with environmental factors such as road curvature, slope, and weather conditions. In special road segments such as construction areas and school zones, additional speed constraints are introduced. The system also considers safety factors such as the following distance from the vehicle ahead and braking distance to ensure that the recommended speed meets safety requirements.

[0108] This embodiment constructs an efficient vehicle information management mechanism. The cloud platform maintains a distributed cache database to store the mapping relationship between license plate information and in-vehicle modules. A multi-level cache strategy is adopted, and hot data is kept in memory to improve query efficiency. The database supports real-time updates to ensure the timeliness of vehicle information. Through a data synchronization mechanism, the consistency of data among all nodes is maintained.

[0109] This embodiment optimizes the organizational structure of warning messages. Warning messages contain multiple levels: the metadata layer contains basic information such as message ID, timestamp, and priority; the content layer contains specific content such as road conditions, detour recommendations, and recommended speeds; the control layer contains display parameters such as broadcast strategies and display styles. A compact serialization format is used to reduce transmission overhead.

[0110] This embodiment designs a reliable message delivery mechanism. The cloud platform realizes the asynchronous transmission of warning messages through a message queue system, supporting message priority management and failure retry. Adopting a combination of push and pull methods, emergency messages are actively pushed, and ordinary messages allow the in-vehicle module to pull regularly. The communication status is monitored through a heartbeat mechanism to ensure message delivery.

[0111] This embodiment implements an intelligent broadcast control strategy. After receiving a warning message, the in-vehicle module first analyzes the message priority and timeliness. For emergency warnings, the current broadcast is immediately interrupted for an insert broadcast, while ordinary warnings wait for an appropriate time to be broadcast. Deep learning models are used for speech synthesis to generate natural and fluent voice prompts. At the same time, key information is prominently displayed on the car machine screen.

[0112] This embodiment optimizes the information display effect through a dynamic display strategy. The screen display adopts a partition layout, and important information is displayed in prominent positions. Key prompts are enhanced through animation effects, such as using dynamic arrows to guide the detour route. The display content automatically adjusts the information density according to the vehicle speed to ensure that the driver can quickly obtain key information.

[0113] Through the above technological innovations, this embodiment solves problems such as weak information pertinence and low transmission efficiency in traditional warning systems. In practical applications, this solution can provide personalized warning information for different vehicles, helping drivers understand road conditions in a timely manner and make reasonable decisions. Especially in complex road conditions and emergencies, the system timely transmits warning information through multiple channels, effectively improving the safety and efficiency of road traffic and significantly enhancing the driving experience of drivers.

[0114] In order to effectively solve the deficiencies of traditional technologies in aspects such as scene perception, data collaboration, and warning decision-making, and significantly improve the intelligence level and service quality of the assisted driving system, this application provides an embodiment of a vehicle driving intelligent control device for implementing all or part of the content of the vehicle driving intelligent control method. Refer to Figure 2 , the vehicle driving intelligent control device specifically includes the following content: The system construction module 10 is used to set up analysis and recognition units along the road. The analysis and recognition units include edge boxes, intelligent cameras, and environmental detection devices. The edge boxes establish a communication connection with the cloud platform and register. The intelligent cameras include license plate capture cameras and wide-angle cameras. The environmental detection devices collect road condition environment parameters. An in-vehicle module is deployed on the vehicle side. The in-vehicle module establishes a connection with the cloud platform through an Internet of Things communication module and registers. The in-vehicle module calls the car machine interface to obtain the vehicle driving trajectory, and divides multiple analysis and recognition units into a regional linkage network based on the driving trajectory; The traffic prediction module 20 is used for the license plate capture camera to obtain the license plate number, license plate color and vehicle type information of the vehicle, the wide-angle camera to obtain the real-time road picture, the environment detection device to obtain the road visibility and road surface state parameters, the edge box to extract the vehicle driving characteristics based on the real-time road picture, the edge box to train the traffic flow prediction model and the risk assessment model based on the vehicle driving characteristics, and send the vehicle driving characteristics, the traffic flow prediction result, the risk assessment value and the environment parameters collected by each analysis and recognition unit to the cloud platform; The assisted driving module 30 is used for the cloud platform to receive the data uploaded by multiple analysis and recognition units, judge whether to trigger area linkage based on the preset road traffic strategy and the risk assessment value, when triggering area linkage, send linkage instructions to the relevant analysis and recognition units according to the area linkage network and establish a communication link, receive the traffic flow prediction results uploaded by the relevant analysis and recognition units, and the cloud platform generates warning information including real-time road conditions, recommended detour routes and safe vehicle speeds based on the traffic flow prediction results, and sends it to the in-vehicle module corresponding to the license plate number for broadcast.

[0115] As can be seen from the above description, the vehicle driving intelligent control device provided by the embodiment of the present application can realize multi-dimensional data collection through the edge box, intelligent camera and environment detection device. Dynamically divide the area linkage network based on the vehicle driving trajectory, and train the traffic flow prediction and risk assessment models at the edge. The cloud platform integrates the data of multiple analysis and recognition units, triggers the area linkage mechanism according to the road traffic strategy and the risk assessment value, and generates accurate warning information including real-time road conditions, recommended detour routes and safe vehicle speeds. This method effectively solves the deficiencies of traditional technologies in aspects such as scene perception, data collaboration and warning decision-making, and significantly improves the intelligent level and service quality of the assisted driving system.

[0116] From the hardware level, in order to effectively solve the deficiencies of traditional technologies in aspects such as scene perception, data collaboration and warning decision-making, and significantly improve the intelligent level and service quality of the assisted driving system, the present application provides an embodiment of an electronic device for implementing all or part of the content in the vehicle driving intelligent control method. The electronic device specifically includes the following content: A processor, a memory, a communications interface, and a bus; wherein, the processor, the memory, and the communications interface complete communication with each other through the bus; the communications interface is used to implement information transmission between the vehicle driving intelligent control device and related devices such as a core business system, a user terminal, and a related database, etc.; the logic controller may be a desktop computer, a tablet computer, a mobile terminal, etc., and this embodiment is not limited thereto. In this embodiment, the logic controller may be implemented with reference to the embodiments of the vehicle driving intelligent control method and the embodiments of the vehicle driving intelligent control device in the embodiments, and the content is incorporated herein, and the repeated parts will not be described again.

[0117] It can be understood that the user terminal may include a smart phone, a tablet electronic device, a network set-top box, a portable computer, a desktop computer, a personal digital assistant (PDA), an in-vehicle device, a smart wearable device, etc. Among them, the smart wearable device may include smart glasses, a smart watch, a smart bracelet, etc.

[0118] In practical applications, part of the vehicle driving intelligent control method may be executed on the electronic device side as described above, or all operations may be completed in the client device. Specifically, it can be selected according to the processing capacity of the client device and the limitations of the user usage scenario, etc. This application does not make any limitations in this regard. If all operations are completed in the client device, the client device may further include a processor.

[0119] The above-mentioned client device may have a communication module (i.e., a communication unit), and may be communicatively connected to a remote server to implement data transmission with the server. The server may include a server on the task scheduling center side, and may also include a server on an intermediate platform in other implementation scenarios, such as a server on a third-party server platform communicatively linked to the task scheduling center server. The server may include a single computer device, or may include a server cluster composed of multiple servers, or a server structure of a distributed device.

[0120] Figure 3 This is a schematic block diagram of the system composition of the electronic device 9600 according to an embodiment of the present application. As Figure 3 shown, the electronic device 9600 may include a central processing unit 9100 and a memory 9140; the memory 9140 is coupled to the central processing unit 9100. It should be noted that this Figure 3 is exemplary; other types of structures may also be used to supplement or replace this structure to implement telecommunication functions or other functions.

[0121] In one embodiment, the function of the intelligent vehicle driving control method can be integrated into the central processing unit 9100. Among them, the central processing unit 9100 can be configured to perform the following controls: Step S101: Set up analysis and recognition units along the road. The analysis and recognition units include edge boxes, intelligent cameras, and environmental detection devices. The edge boxes establish communication connections with the cloud platform and register. The intelligent cameras include license plate capture cameras and wide-angle cameras. The environmental detection devices collect road condition environment parameters. Deploy in-vehicle modules at the vehicle end. The in-vehicle modules establish connections with the cloud platform through the Internet of Things communication modules and register. The in-vehicle modules call the vehicle machine interface to obtain the vehicle driving trajectory, and divide the multiple analysis and recognition units into a regional linkage network based on the driving trajectory. Step S102: The license plate capture camera obtains the license plate number, license plate color, and vehicle type information of the vehicle. The wide-angle camera obtains the real-time road picture. The environmental detection device obtains the road visibility and road surface state parameters. The edge box extracts the vehicle driving characteristics based on the real-time road picture. The edge box trains a traffic flow prediction model and a risk assessment model based on the vehicle driving characteristics, and sends the vehicle driving characteristics, traffic flow prediction results, risk assessment values, and environmental parameters collected by each analysis and recognition unit to the cloud platform. Step S103: The cloud platform receives the data uploaded by the multiple analysis and recognition units, determines whether to trigger regional linkage based on the preset road traffic strategy and the risk assessment value. When triggering regional linkage, it issues linkage instructions to the relevant analysis and recognition units according to the regional linkage network and establishes a communication link, and receives the traffic flow prediction results uploaded by the relevant analysis and recognition units. The cloud platform generates a warning message including real-time road conditions, recommended detour routes, and safe vehicle speeds based on the traffic flow prediction results, and sends it to the in-vehicle module corresponding to the license plate number for broadcast.

[0122] As can be seen from the above description, the electronic device provided in the embodiment of the present application realizes multi-dimensional data collection through edge boxes, intelligent cameras, and environmental detection devices. Dynamically divides the regional linkage network based on the vehicle driving trajectory, and trains the traffic flow prediction and risk assessment models at the edge. The cloud platform integrates the data of multiple analysis and recognition units, triggers the regional linkage mechanism according to the road traffic strategy and the risk assessment value, and generates accurate warning messages including real-time road conditions, recommended detour routes, and safe vehicle speeds. This method effectively solves the deficiencies of traditional technologies in aspects such as scenario perception, data collaboration, and warning decision-making, and significantly improves the intelligent level and service quality of the assisted driving system.

[0123] In another embodiment, the intelligent vehicle driving control device may be separately configured from the central processor 9100. For example, the intelligent vehicle driving control device may be configured as a chip connected to the central processor 9100, and the functions of the intelligent vehicle driving control method are implemented through the control of the central processor.

[0124] As Figure 3 shown, the electronic device 9600 may further include: a communication module 9110, an input unit 9120, an audio processor 9130, a display 9160, and a power supply 9170. It should be noted that the electronic device 9600 does not necessarily have to include Figure 3 all the components shown in Figure 3 ; in addition, the electronic device 9600 may further include

[0125] As Figure 3 shown, the central processor 9100, sometimes also referred to as a controller or an operation control, may include a microprocessor or other processor devices and / or logic devices. The central processor 9100 receives inputs and controls the operations of the various components of the electronic device 9600.

[0126] Among them, the memory 9140, for example, may be one or more of a buffer, a flash memory, a hard drive, a removable medium, a volatile memory, a non-volatile memory, or other suitable devices. It can store the above-mentioned information related to failures, and can also store programs for executing relevant information. And the central processor 9100 can execute the programs stored in the memory 9140 to implement information storage or processing, etc.

[0127] The input unit 9120 provides inputs to the central processor 9100. The input unit 9120 is, for example, a key or a touch input device. The power supply 9170 is used to supply power to the electronic device 9600. The display 9160 is used to display display objects such as images and texts. The display may be, for example, an LCD display, but is not limited thereto.

[0128] The memory 9140 may be a solid-state memory. For example, a read-only memory (ROM), a random access memory (RAM), a SIM card, etc. It may also be such a memory that stores information even when powered off, can be selectively erased and has more data. Examples of such a memory are sometimes referred to as EPROMs, etc. The memory 9140 may also be some other type of device. The memory 9140 includes a buffer memory 9141 (sometimes referred to as a buffer). The memory 9140 may include an application / function storage unit 9142, and the application / function storage unit 9142 is used to store application programs and function programs or the processes for operating the electronic device 9600 through the central processor 9100.

[0129] The memory 9140 may further include a data storage unit 9143 for storing data such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage unit 9144 of the memory 9140 may include various drivers for the communication functions of the electronic device and / or for performing other functions of the electronic device (such as a messaging application, an address book application, etc.).

[0130] The communication module 9110 is a transmitter / receiver that transmits and receives signals via the antenna 9111. The communication module 9110 (transmitter / receiver) is coupled to the central processor 9100 to provide input signals and receive output signals, which may be the same as in the case of a conventional mobile communication terminal.

[0131] Based on different communication technologies, multiple communication modules 9110 may be provided in the same electronic device, such as a cellular network module, a Bluetooth module, and / or a wireless local area network module, etc. The communication module 9110 (transmitter / receiver) is also coupled to the speaker 9131 and the microphone 9132 via the audio processor 9130 to provide an audio output via the speaker 9131 and receive an audio input from the microphone 9132, thereby implementing normal telecommunication functions. The audio processor 9130 may include any suitable buffers, decoders, amplifiers, etc. Additionally, the audio processor 9130 is also coupled to the central processor 9100, so that recording can be performed on the local device through the microphone 9132, and the sound stored on the local device can be played through the speaker 9131.

[0132] Embodiments of the present application also provide a computer-readable storage medium capable of implementing all steps of the vehicle driving intelligent control method with the execution subject being a server or a client in the above embodiments. A computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, all steps of the vehicle driving intelligent control method with the execution subject being a server or a client in the above embodiments are implemented. For example, when the processor executes the computer program, the following steps are implemented: Step S101: Set up an analysis and recognition unit along the road. The analysis and recognition unit includes an edge box, an intelligent camera, and an environment detection device. The edge box establishes a communication connection with the cloud platform and registers. The intelligent camera includes a license plate capture camera and a wide-angle camera. The environment detection device collects road condition environment parameters; deploy an in-vehicle module at the vehicle end. The in-vehicle module establishes a connection with the cloud platform through the Internet of Things communication module and registers. The in-vehicle module calls the vehicle machine interface to obtain the vehicle driving trajectory, and divides multiple analysis and recognition units into a regional linkage network based on the driving trajectory; Step S102: The license plate capture camera obtains the license plate number, license plate color, and vehicle type information of the vehicle. The wide-angle camera obtains the real-time road scene. The environment detection device obtains the road visibility and road surface state parameters. The edge box extracts the vehicle driving characteristics based on the real-time road scene. The edge box trains the traffic flow prediction model and the risk assessment model based on the vehicle driving characteristics, and sends the vehicle driving characteristics, the traffic flow prediction results, the risk assessment values, and the environment parameters collected by each analysis and recognition unit to the cloud platform; Step S103: The cloud platform receives the data uploaded by multiple analysis and recognition units, determines whether to trigger area linkage based on the preset road traffic strategy and the risk assessment value. When area linkage is triggered, it sends a linkage instruction to the relevant analysis and recognition units according to the area linkage network and establishes a communication link, and receives the traffic flow prediction results uploaded by the relevant analysis and recognition units. The cloud platform generates a warning message including real-time road conditions, recommended detour routes, and safe vehicle speeds based on the traffic flow prediction results, and sends it to the in-vehicle module corresponding to the license plate number for broadcasting.

[0133] As can be seen from the above description, the computer-readable storage medium provided in the embodiment of the present application realizes multi-dimensional data collection through the edge box, intelligent cameras, and environment detection devices. Dynamically divides the area linkage network based on the vehicle driving trajectory, and trains the traffic flow prediction and risk assessment models at the edge. The cloud platform integrates the data of multiple analysis and recognition units, triggers the area linkage mechanism according to the road traffic strategy and the risk assessment value, and generates accurate warning messages including real-time road conditions, recommended detour routes, and safe vehicle speeds. This method effectively solves the deficiencies of traditional technologies in aspects such as scene perception, data collaboration, and warning decision-making, and significantly improves the intelligence level and service quality of the assisted driving system.

[0134] The embodiment of the present application also provides a computer program product that can implement all the steps of the vehicle driving intelligent control method with the execution subject being a server or a client in the above embodiment. When the computer program / instructions are executed by a processor, the steps of the vehicle driving intelligent control method are implemented. For example, the computer program / instructions implement the following steps: Step S101: Set up analysis and recognition units along the road. The analysis and recognition units include an edge box, intelligent cameras, and an environment detection device. The edge box establishes a communication connection with the cloud platform and registers. The intelligent cameras include a license plate capture camera and a wide-angle camera. The environment detection device collects road condition environment parameters; deploy an in-vehicle module at the vehicle end. The in-vehicle module establishes a connection with the cloud platform through the Internet of Things communication module and registers. The in-vehicle module calls the vehicle computer interface to obtain the vehicle driving trajectory, and divides multiple analysis and recognition units into an area linkage network based on the driving trajectory; Step S102: The license plate capture camera obtains the license plate number, license plate color, and vehicle type information of the vehicle. The wide-angle camera obtains the real-time road scene. The environmental detection device obtains the road visibility and road surface state parameters. The edge box extracts the vehicle driving characteristics based on the real-time road scene. The edge box trains a traffic flow prediction model and a risk assessment model based on the vehicle driving characteristics, and sends the vehicle driving characteristics, traffic flow prediction results, risk assessment values, and environmental parameters collected by each analysis and recognition unit to the cloud platform; Step S103: The cloud platform receives the data uploaded by multiple analysis and recognition units, determines whether to trigger area linkage based on a preset road traffic strategy and the risk assessment value. When area linkage is triggered, a linkage instruction is sent to the relevant analysis and recognition units according to the area linkage network, and a communication link is established. The cloud platform receives the traffic flow prediction results uploaded by the relevant analysis and recognition units. The cloud platform generates a warning message including real-time road conditions, recommended detour routes, and safe vehicle speeds based on the traffic flow prediction results, and sends it to the in-vehicle module corresponding to the license plate number for broadcasting.

[0135] As can be seen from the above description, the computer program product provided by the embodiment of the present application realizes multi-dimensional data collection through an edge box, an intelligent camera, and an environmental detection device. Dynamically divides the area linkage network based on the vehicle driving trajectory, and trains traffic flow prediction and risk assessment models at the edge. The cloud platform integrates the data of multiple analysis and recognition units, triggers the area linkage mechanism according to the road traffic strategy and the risk assessment value, and generates accurate warning messages including real-time road conditions, recommended detour routes, and safe vehicle speeds. This method effectively solves the deficiencies of traditional technologies in aspects such as scene perception, data collaboration, and warning decision-making, and significantly improves the intelligence level and service quality of the assisted driving system.

[0136] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a device, or a computer program product. Therefore, the present invention can be in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can be in the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0137] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatus (devices), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices to generate a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices generate means for implementing the functions specified in one or more flows and / or blocks in the flow Figure 1 one or more flows and / or blocks Figure 1 or means for implementing the functions specified in one or more blocks or multiple blocks.

[0138] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in one or more flows and / or blocks in the flow Figure 1 one or more flows and / or blocks Figure 1 or means for implementing the functions specified in one or more blocks or multiple blocks.

[0139] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more flows and / or blocks in the flow Figure 1 one or more flows and / or blocks Figure 1 or means for implementing the functions specified in one or more blocks or multiple blocks.

[0140] Specific embodiments are applied in the present invention to elaborate on the principles and implementation manners of the present invention. The descriptions of the above embodiments are only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, based on the idea of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A vehicle driving intelligent control method, characterized in that: The method comprises: An analysis and identification unit is set up along the road, and the analysis and identification unit includes an edge box, a smart camera and an environmental detection device. The edge box establishes a communication connection with the cloud platform and registers. The smart camera includes a license plate capture camera and a wide-angle camera. The environmental detection device collects road environment parameters; an on-board module is deployed on the vehicle side. The on-board module establishes a connection with the cloud platform and registers through the Internet of Things communication module. The on-board module calls the vehicle-machine interface to obtain the vehicle's driving trajectory, and divides multiple analysis and identification units into a regional linkage network based on the driving trajectory; The license plate capture camera obtains the vehicle's license plate number, license plate color and vehicle model information, the wide-angle camera obtains real-time road images, the environmental detection equipment obtains road visibility and road surface state parameters, the edge box extracts vehicle driving characteristics based on the real-time road images, the edge box trains a traffic flow prediction model and a risk assessment model based on the vehicle driving characteristics, and sends the vehicle driving characteristics, traffic flow prediction results, risk assessment values ​​and environmental parameters collected by each analysis and recognition unit to the cloud platform; The cloud platform receives data uploaded by multiple analysis and identification units, determines whether to trigger regional linkage based on a preset road traffic strategy and the risk assessment value, and when regional linkage is triggered, sends linkage instructions to the relevant analysis and identification units according to the regional linkage network and establishes a communication link, receives the traffic flow prediction results uploaded by the relevant analysis and identification units, and the cloud platform generates warning information including real-time road conditions, recommended detour routes and safe vehicle speeds based on the traffic flow prediction results, and sends it to the vehicle-mounted module corresponding to the license plate number for broadcast.

2. The vehicle driving intelligent control method according to claim 1, characterized in that: The analysis and identification unit is set up along the road, and the analysis and identification unit includes an edge box, a smart camera and an environmental detection device. The edge box establishes a communication connection with the cloud platform and registers. The smart camera includes a license plate capture camera and a wide-angle camera. The environmental detection device collects road environment parameters, including: Select a key position of the road to deploy the analysis and identification unit, initialize the parameters of the edge box, obtain a unique identification code through the Internet of Things communication module, send a registration request to the cloud platform based on the unique identification code, the cloud platform returns registration authorization information, and the edge box establishes a secure communication link with the cloud platform based on the registration authorization information; The edge box establishes a communication connection with the smart camera and the environmental detection device, calibrates the angle and focal length of the smart camera, sets the license plate capture camera to point in the direction of vehicle traffic, sets the wide-angle camera to a panoramic road monitoring angle, establishes a sub-device management group with the edge box as the center, configures the sampling period and trigger threshold of the environmental detection device, and obtains environmental parameters including visibility, temperature, humidity and light intensity from the environmental detection device at preset time intervals.

3. The vehicle driving intelligent control method according to claim 1, characterized in that: The vehicle-mounted module is deployed on the vehicle side, the vehicle-mounted module establishes a connection with the cloud platform through the Internet of Things communication module and registers, the vehicle-mounted module calls the vehicle-machine interface to obtain the vehicle driving trajectory, and divides the multiple analysis and identification units into a regional linkage network based on the driving trajectory, including: The vehicle-mounted module obtains the device unique identifier through the Internet of Things communication module, and sends a registration request containing the vehicle VIN code to the cloud platform based on the device unique identifier. The cloud platform returns a device authorization code, and the vehicle-mounted module establishes a data transmission channel with the cloud platform based on the device authorization code. The vehicle-mounted module calls the vehicle CAN bus interface to collect the vehicle's real-time position coordinates, driving speed and direction angle data; The on-board module uploads the position coordinates, driving speed and direction angle data to the cloud platform at preset time intervals. The cloud platform maps the vehicle position coordinates to an electronic map based on a geographic information system, determines a forward predicted path according to the vehicle's driving direction, extracts the analysis and identification units within the coverage of the predicted path, constructs the extracted analysis and identification units into a regional linkage network, and assigns a network priority to each of the analysis and identification units.

4. The vehicle driving intelligent control method according to claim 1, characterized in that: The license plate capture camera obtains the vehicle's license plate number, license plate color and vehicle model information, the wide-angle camera obtains real-time road images, the environmental detection equipment obtains road visibility and road surface state parameters, and the edge box extracts vehicle driving characteristics based on the real-time road images, including: The license plate capture camera collects image data when a vehicle passes by, and the edge box processes the image data based on a convolutional neural network model, extracts license plate area features and performs character segmentation, identifies license plate numbers and color information, analyzes vehicle body contour features through a deep learning model to obtain vehicle model classification results, and stores the license plate number, color information and vehicle model classification results in a cache database of the edge box; The wide-angle camera collects road traffic video streams at a preset frame rate, the environmental detection equipment obtains visibility values ​​based on the laser ranging principle, and detects road surface state parameters through surface temperature sensors and humidity sensors. The edge box performs target detection and tracking processing on the video stream, extracts driving characteristic parameters such as vehicle traffic trajectory, vehicle speed, and vehicle spacing, and the edge box establishes a mapping relationship between the driving characteristic parameters and license plate information.

5. The vehicle driving intelligent control method according to claim 1, characterized in that: The edge box trains a traffic flow prediction model and a risk assessment model based on the vehicle driving characteristics, and sends the vehicle driving characteristics, the traffic flow prediction results, the risk assessment value and the environmental parameters collected by each of the analysis and recognition units to the cloud platform, including: The edge box divides the vehicle driving characteristic parameters into training data sets according to time series, uses a long short-term memory neural network to build a network structure of a traffic flow prediction model, takes vehicle speed, vehicle density, and travel time as input features, iteratively trains the network structure to obtain a traffic flow prediction model, builds a risk assessment model based on a decision tree algorithm, and takes vehicle spacing, relative speed, and road state parameters as input variables to calculate a collision risk assessment value; The edge box organizes the vehicle driving characteristics, environmental parameters and timestamp information into a data packet, which contains the future traffic flow state prediction results output by the traffic flow prediction model and the risk assessment value calculated by the risk assessment model. The edge box sends the data packet to the cloud platform through a secure communication link according to a preset data upload cycle, and the cloud platform performs time synchronization and data verification on the received data packet.

6. The vehicle driving intelligent control method according to claim 1, characterized in that: The cloud platform receives data uploaded by the plurality of analysis and identification units, determines whether to trigger regional linkage based on a preset road traffic strategy and the risk assessment value, and when triggering regional linkage, issues linkage instructions to the relevant analysis and identification units according to the regional linkage network and establishes a communication link, including: The cloud platform receives data packets uploaded by a plurality of the analysis and identification units, extracts the risk assessment value, vehicle flow, vehicle speed and other index parameters from the data packets, and compares the index parameters with the thresholds in the road traffic strategy, wherein the road traffic strategy includes a vehicle flow density threshold, an average vehicle speed threshold, and a unit mileage accident risk threshold. When the index parameter of any of the analysis and identification units exceeds the corresponding threshold, a regional linkage judgment is triggered, and the cloud platform determines the affected range according to the position of the analysis and identification unit that triggers the linkage; The cloud platform selects the analysis and identification unit located within the affected range from the regional linkage network as a linkage object, generates a linkage session including a linkage object identification code, creates a data sharing channel based on the linkage session, sends a linkage instruction including sharing channel configuration information to the linkage object, receives a confirmation response returned by the linkage object through the data sharing channel, and completes the establishment of a communication link between the linkage objects.

7. The vehicle driving intelligent control method according to claim 1, characterized in that: The receiving of the traffic flow prediction result uploaded by the relevant analysis and identification unit, the cloud platform generating warning information including real-time road conditions, recommended detour routes and safe vehicle speed based on the traffic flow prediction result, and sending the warning information to the vehicle-mounted module corresponding to the license plate number for broadcasting, includes: The cloud platform receives the traffic flow prediction result uploaded by the linkage object, maps the traffic flow prediction result to the road section based on the real-time location coordinates, calculates the congestion index of each road section, and calls the path planning engine to calculate the shortest path and the optimal detour path, inputs the lane-level traffic condition information, the road section congestion index, the average travel speed, and the travel time into the warning information generation model, calculates the recommended safe speed according to the vehicle type and road traffic conditions, and generates the warning text content; The cloud platform extracts the license plate information from the cache database of the linkage object, establishes a corresponding relationship between the warning text content and the license plate information, and generates a warning message containing an addressing identifier. The cloud platform queries the communication address of the on-board module and sends the warning message to the corresponding on-board module through a message queue. After receiving the warning message, the on-board module calls the speech synthesis engine for broadcasting and displays the warning content on the vehicle screen.

8. A vehicle driving intelligent control device, characterized in that: The device comprises: A system building module is used to set up analysis and identification units along the road. The analysis and identification units include edge boxes, smart cameras and environmental detection equipment. The edge boxes establish a communication connection with the cloud platform and register. The smart cameras include license plate capture cameras and wide-angle cameras. The environmental detection equipment collects road environment parameters. An on-board module is deployed on the vehicle side. The on-board module establishes a connection with the cloud platform and registers through the Internet of Things communication module. The on-board module calls the vehicle-machine interface to obtain the vehicle's driving trajectory, and divides multiple analysis and identification units into regional linkage networks based on the driving trajectory. Traffic prediction module, used for the license plate capture camera to obtain the vehicle's license plate number, license plate color and vehicle model information, the wide-angle camera to obtain real-time road images, the environmental detection equipment to obtain road visibility and road surface state parameters, the edge box to extract vehicle driving characteristics based on the real-time road images, the edge box to train traffic flow prediction models and risk assessment models based on the vehicle driving characteristics, and send the vehicle driving characteristics, traffic flow prediction results, risk assessment values ​​and environmental parameters collected by each analysis and recognition unit to the cloud platform; The auxiliary driving module is used for the cloud platform to receive the data uploaded by the multiple analysis and identification units, judge whether to trigger regional linkage based on the preset road traffic strategy and the risk assessment value, and when regional linkage is triggered, send linkage instructions to the relevant analysis and identification units according to the regional linkage network and establish a communication link, receive the traffic flow prediction results uploaded by the relevant analysis and identification units, and the cloud platform generates warning information including real-time road conditions, recommended detour routes and safe vehicle speeds based on the traffic flow prediction results, and sends it to the vehicle-mounted module corresponding to the license plate number for broadcast.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the vehicle driving intelligent control method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the vehicle driving intelligent control method described in any one of claims 1 to 7 are implemented.

Images