A rain day intelligent driving dynamic regulation method and system based on vehicle-network cooperation

Abstract

The application relates to the technical field of intelligent driving and provides a rain day intelligent driving dynamic regulation method and system based on vehicle network cooperation, which comprises the following steps: multi-source data fusion perception, obtaining minute-level rainfall forecast data of a meteorological bureau, rain day driving behavior data of other vehicles on the same road section and real-time environment data; dynamic risk assessment, constructing an LSTM+CNN hybrid neural network model, outputting a risk probability P and dividing the risk probability P into three risk levels of low, medium and high; calculating an adaptive following distance according to a dynamic following distance algorithm; cooperative control execution, dynamically adjusting ACC / AEB parameters according to the risk level, and displaying a real scene warning through AR-HUD superposition; adaptive system intervention strategy; triggering pre-braking and sound-light reminding, and realizing group cooperation. The application solves the problem of perception lag, provides a personalized decision scheme, solves the problems of single data and poor scene adaptability, and improves the accuracy of risk prediction.

Description

Technical Field

[0001] This invention relates to the field of intelligent driving technology, and in particular to a dynamic control method and system for intelligent driving in rainy weather based on vehicle-to-everything (V2X) collaboration. Background Technology

[0002] In recent years, with the rapid development of technology and the increasing popularity of intelligent driving, current in-depth research on intelligent driving focuses on vehicle safety and reliability. However, research on intelligent driving in adverse environments such as rainy days is difficult and relatively limited.

[0003] Existing intelligent driving assistance systems (such as patent CN110481554B) mainly rely on onboard sensors (such as rain sensors and wiper settings) for local environmental perception, which has the following limitations: perception lag, only able to passively identify the current rainfall status and unable to predict the rainfall area and intensity in advance; limited data, lacking real-time road condition sharing with other vehicles on the same road segment, making it difficult to judge the slipperiness of the road surface; coarse decision-making, based on preset thresholds (such as wiper settings ≥ 3) for graded control, without combining driving habits, weather forecasts, and other multi-dimensional parameters; poor scene adaptability, in extreme weather conditions such as heavy rain, onboard sensors are easily affected by mud and water mist, leading to misjudgments.

[0004] Therefore, there is an urgent need for a dynamic control scheme that integrates vehicle network data to improve the safety and forward-looking capabilities of intelligent driving in rainy weather. Summary of the Invention

[0005] This invention provides a dynamic control method and system for intelligent driving in rainy weather based on vehicle-network collaboration, aiming to solve the problems of perception lag, single data, rough decision-making and poor scene adaptability in traditional solutions, and improve the safety and adaptability of intelligent driving systems in complex weather conditions.

[0006] This invention is implemented as follows: a dynamic control method and system for intelligent driving in rainy weather based on vehicle-to-grid (V2G) cooperation. The dynamic control method for intelligent driving in rainy weather based on V2G cooperation includes:

[0007] Multi-source data fusion perception: obtain minute-level rainfall forecast data from the meteorological bureau through the vehicle-to-everything (V2X) API, and combine it with vehicle GPS to locate rainfall areas in advance; obtain rain-related driving behavior data of other vehicles on the same road segment through the V2X communication protocol, and share the information collected by itself with surrounding vehicles; obtain real-time environmental data through vehicle cameras, radar and infrared sensors.

[0008] Dynamic risk assessment is based on a hybrid LSTM+CNN neural network model built on the TensorFlow framework. The input includes rainfall forecast data, rainy driving behavior data, and real-time environmental data. The output is a risk probability P, which is divided into three risk levels: low, medium, and high. The adaptive following distance is calculated based on a dynamic following distance algorithm.

[0009] The system implements coordinated control, dynamically adjusts ACC / AEB parameters based on risk levels, and displays real-world warnings via AR-HUD overlay. It also adapts system intervention strategies based on drivers' rainy driving habits and obtains authorized pedestrian locations through vehicle-to-everything (V2X) communication to trigger pre-braking and audible / visual alerts, thus achieving group coordination.

[0010] Preferably, the rainfall forecast data includes the rainfall amount for the next 10 minutes, the latitude and longitude boundaries of the rainfall area, and the trend of road surface temperature changes; the accidental triggering data of other vehicles on the same road section includes tire slippage frequency, ABS intervention frequency, vehicle speed fluctuation data, and brake light triggering frequency; the real-time environmental data includes road surface water depth, water film thickness, road surface friction coefficient, visibility, and the status of traffic participants.

[0011] Preferably, the risk probability P is divided into three risk levels: low, medium, and high. Specifically, when P < 30%, it is classified as low risk; when 30% ≤ P ≤ 70%, it is classified as medium risk; and when P > 70%, it is classified as high risk.

[0012] Preferably, the calculation formula for the dynamic following distance algorithm is:

[0013] ;

[0014] in, This is the standard following distance for the vehicle based on its current speed. This is the rainfall coefficient. The coefficient of slipperiness of the road surface. This is the curvature coefficient of the curve. , and These are the weights for rainfall coefficient, road surface slipperiness coefficient, and curve curvature coefficient, respectively.

[0015] Preferably, the step of dynamically adjusting the ACC / AEB parameters according to the risk level specifically includes:

[0016] When the risk level is low, slightly adjust the ACC following distance.

[0017] At medium-risk levels, significantly increase following distance and improve AEB sensitivity;

[0018] When the risk level is high, the ACC function will be disabled and a voice safety reminder will be triggered, while the vehicle speed will be forcibly limited to 80% of the road speed limit.

[0019] Preferably, the method of displaying real-world warnings via AR-HUD overlay; adapting system intervention strategies based on drivers' rainy driving habits; obtaining authorized pedestrian locations through vehicle-to-everything (V2X) communication to trigger pre-braking and audible / visual alerts, and achieving group coordination, specifically includes:

[0020] AR Real-Scene Intelligent Warning: Warning information is overlaid on the driver's screen via AR-HUD, including red grid markings for road surfaces with a water film thickness >5mm, yellow deviation arrows when the vehicle in front exhibits a skidding trajectory, and white virtual guide lines projected before entering a curve. Key warning information is accompanied by audible prompts. Driver Habit Adaptation: Driver's driving operation data in rainy weather is collected by onboard sensors. If the frequency of emergency braking is detected >3 times / 10 minutes, the AEB warning time is automatically advanced from the default 1.5 seconds to 2.0 seconds; if the frequency of emergency braking is detected ≤1 time / 10 minutes, the system's default parameters are maintained to reduce unnecessary intervention.

[0021] Pedestrian safety protection outside the vehicle: By receiving location data of pedestrians via a user-authorized mobile APP through the vehicle network, pedestrian targets can be identified 20 meters in advance in heavy rain, triggering the AEB system for pre-braking. At the same time, the vehicle's external speakers will play a voice reminder and the vehicle lights will flash to warn pedestrians that the vehicle is approaching.

[0022] A vehicle-to-everything (V2X) intelligent driving dynamic control system for rainy weather, comprising:

[0023] Data fusion module: used for the acquisition, preprocessing and fusion of multi-source data, including a meteorological data interface unit, a V2X communication unit and an on-board sensor fusion unit; the meteorological data interface unit connects to the meteorological bureau's API via a 4G / 5G network to parse and store minute-level rainfall forecast data; the V2X communication unit, based on the IEEE 802.11p standard, enables communication with surrounding vehicles and roadside units to complete the transmission and reception of road condition data; the on-board sensor fusion unit connects to cameras, millimeter-wave radar and infrared sensors to perform noise reduction, synchronization calibration and feature extraction on the acquired raw data, and output unified environmental perception results;

[0024] Risk assessment module: used to dynamically calculate driving risks in rainy weather and generate adaptive strategies, including a machine learning model unit, a dynamic algorithm unit, and a risk level classification unit; the machine learning model unit deploys a trained LSTM+CNN hybrid neural network, inputs preprocessed multi-source data, and outputs a risk probability P; the dynamic algorithm unit calls a dynamic following distance algorithm, combines the current vehicle speed and road condition parameters to calculate the adaptive following distance; the risk level classification unit determines the risk level based on the risk probability P and maps it to the corresponding control strategy;

[0025] Control Execution Module: This module executes the control strategies output by the risk assessment module, including an ACC / AEB adjustment unit, an AR interaction unit, a human-machine collaboration unit, and a pedestrian safety unit. The ACC / AEB adjustment unit dynamically adjusts the following distance, AEB sensitivity, and speed limit based on the risk level. The AR interaction unit controls the display content of the AR-HUD to visualize warning information. The human-machine collaboration unit generates voice reminders, which are played through the vehicle's audio system, while simultaneously collecting driver behavior data for habit adaptation. The pedestrian safety unit receives pedestrian location data and controls AEB pre-braking, external speakers, and vehicle lights to achieve pedestrian safety protection.

[0026] Preferably, the data fusion module further includes an edge computing node for real-time processing of V2X shared data, generating dynamic road condition heatmaps, intuitively displaying the distribution of road slippery risk, and supporting the push of regional risk warning information.

[0027] Preferably, the AR-HUD interface controlled by the AR interaction unit is divided into three areas: the upper area displays the water film thickness value and risk level text label, with high-risk information marked in red; the middle area is a real-scene overlay layer, displaying grid marks, offset arrows and virtual guide lines; the lower area displays a voice interaction prompt box and driving mode suggestions, with key operation suggestions highlighted in red.

[0028] Preferably, the system further includes a data storage unit for storing historical meteorological data, road condition data, driver behavior data, and system operation logs, with a storage capacity of ≥100GB and a data retention time of ≥6 months, providing data support for the iterative optimization of the neural network model.

[0029] Compared with the prior art, the beneficial effects of the present invention are:

[0030] This vehicle-network collaborative dynamic control method for intelligent driving in rainy weather uses meteorological data to provide early warnings of rainfall and combines historical road risk data for pre-adjustment, shortening the system's decision-making time and solving the problem of perception lag. Simultaneously, it adapts system intervention strategies to different drivers' driving habits, providing personalized decision-making solutions. Furthermore, through multi-dimensional control strategies, it can reduce the accident rate of intelligent driving in rainy weather, with particularly significant protective effects in complex scenarios such as heavy rain and curves, solving the problems of limited data and poor scenario adaptability. In addition, the closed-loop architecture of "perception-evaluation-control" improves the accuracy of road surface slippage judgment, and the risk prediction accuracy of the LSTM+CNN hybrid neural network is enhanced, providing a reliable basis for precise control. Attached Figure Description

[0031] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 This is an overall flowchart of the intelligent driving dynamic control method in rainy weather according to an embodiment of the present invention;

[0033] Figure 2 This is a connection block diagram of the rain-time intelligent driving dynamic control system according to an embodiment of the present invention. Detailed Implementation

[0034] To better understand the technical content of this invention, the technical solutions of this invention are further described and explained below with reference to specific embodiments, but are not limited thereto. The technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0035] Example 1

[0036] refer to Figure 1 A dynamic control method for intelligent driving in rainy weather based on vehicle-to-everything (V2X) collaboration includes:

[0037] Multi-source data fusion:

[0038] Meteorological data access: After the vehicle starts, the meteorological data interface unit of the data fusion module automatically calls the meteorological bureau's API via the 4G / 5G network, sets the data update frequency to 1 minute / time, and obtains minute-level rainfall forecast data within a 20-kilometer radius ahead, including rainfall, latitude and longitude boundaries of the rainfall area, and road surface temperature trends. When the vehicle's GPS (positioning accuracy ≤ 5 meters) detects that the estimated arrival time of the vehicle from the rainfall area is 3 minutes and the forecast rainfall is ≥ 5 mm / h, the system immediately triggers a pre-warning, adjusts the ACC following distance from the default 1.5 seconds to 2.0 seconds in advance, and activates the windshield preheating function to reduce the impact of raindrops on visibility.

[0039] V2X Data Sharing: During operation on urban expressways, the V2X communication unit establishes communication connections with vehicles within a 1-kilometer radius based on the IEEE 802.11p protocol, with communication latency controlled within 100ms. When it receives "ABS intervention" signals continuously uploaded by three vehicles 500 meters ahead (each vehicle intervenes ≥2 times within 1 minute), combined with the road surface reflectivity detected by its own millimeter-wave radar (1.5 times higher than the reflectivity threshold for dry roads), the edge computing node immediately generates an assessment result of "road surface slippery risk index = 0.8" and pushes a warning message of "high slippery road surface ahead, speed limit recommended 60km / h" to all related vehicles within a 1-kilometer radius behind via V2X communication.

[0040] Vehicle sensor fusion: The vehicle sensor fusion unit simultaneously collects data from cameras, millimeter-wave radar, and infrared sensors. It performs Gaussian filtering to denoise the images collected by the cameras, calibrates the distance measurement error of the radar (controlled within ±0.1 meters), and matches the heat source detection results of the infrared sensors with the visual recognition results of the cameras. Finally, it outputs a unified environmental perception report containing information such as raindrop density, distance to obstacles ahead, and pedestrian positions, providing real-time data support for risk assessment.

[0041] Dynamic risk assessment:

[0042] A hybrid LSTM+CNN neural network model is constructed based on the TensorFlow framework. The input includes rainfall forecast data, rainy driving behavior data, and real-time environmental data. The output is a risk probability P, which is divided into three risk levels: low, medium, and high. When P < 30%, it is classified as low risk; when 30% ≤ P ≤ 70%, it is classified as medium risk; and when P > 70%, it is classified as high risk.

[0043] Model Training: Over 100,000 historical rainy-day driving data points were collected from different regions, covering varying rainfall intensities (light, moderate, heavy rain), road types (urban roads, highways, mountain roads), and driver operation data for different driving styles. The data includes environmental data (rainfall, road surface temperature, light intensity), road condition data (road curvature, slope angle, water depth), vehicle status data (vehicle speed, braking frequency), and accident data (accident type, environmental and vehicle status at the time of occurrence). A hybrid LSTM+CNN neural network was constructed using the TensorFlow framework. The CNN layer was used to extract spatial features from the environmental and road condition data, while the LSTM layer was used to capture temporal series features of driver behavior data. Supervised training was performed using accident probability as a label, with 100 training iterations and an initial learning rate of 0.001. An adaptive learning rate adjustment strategy was used to optimize the model's convergence speed. The final model achieved a risk prediction accuracy of 93.5% on the test set.

[0044] Dynamic following distance calculation:

[0045] ;

[0046] in, This is the standard following distance for the vehicle based on its current speed. This is the rainfall coefficient. The coefficient of slipperiness of the road surface. This is the curvature coefficient of the curve. , and These are the weights for rainfall coefficient, road surface slip coefficient, and curve curvature coefficient, respectively. When a vehicle travels on a mountain curve with curvature C=0.05 (safe curvature threshold is 0.03), the real-time rainfall is 8mm / h (rainstorm threshold is 10mm / h), the road surface adhesion coefficient obtained via V2X sharing is 0.4 (standard value for dry road surface is 0.8), the current vehicle speed is 80km / h, and the corresponding standard following distance D0=50 meters. This is calculated based on a dynamic following distance algorithm. =8 / 10=0.8, =(0.8-0.4) / 0.8=0.5, =0.05 / 0.03≈1.667, substitute into the coefficient weight =0.8、 =0.5、 =0.6, A weight of 0.8 indicates that rainfall has a moderate to strong impact on driving safety (e.g., during heavy rain). Approaching 1, the distance increases by an additional 80%. A weight of 0.5 indicates that road surface slippage has a weaker impact on driving safety than rainfall. The weight of 0.6 indicates that the impact of curve curvature on driving safety lies between that of rainfall and road surface slipperiness.

[0047] Collaborative control execution:

[0048] The ACC / AEB parameters are dynamically adjusted according to the risk level, and the real-world warning is displayed by overlaying it with AR-HUD; the system intervention strategy is adapted based on the driver's driving habits in rainy weather; the location of authorized pedestrians is obtained through vehicle-to-everything (V2X) communication to trigger pre-braking and audible and visual alerts, thereby achieving group collaboration.

[0049] Taking a scenario of a sudden rainstorm hitting an urban expressway as an example: Early warning stage: The vehicle's navigation system displays that a rainstorm will occur 3 kilometers ahead in the next 5 minutes. The system retrieves historical data for this section from the cloud (3 rear-end collisions caused by flooding during the same period last year), and adjusts the ACC following distance to 2.5 seconds in advance, activating the rearview mirror heating and windshield defroster. Real-time control stage: After entering the rain area, the vehicle network receives data from the vehicle ahead showing "tire slippage count = 3 times / minute." The system marks this section as high-risk, forcibly limiting the vehicle speed to 60 km / h, and increasing the AEB sensitivity to the highest level. If the driver brakes suddenly at this time, the system applies an additional 10% braking force to prevent skidding. Group collaboration stage: The current vehicle pushes a "rainstorm ahead + highly slippery road surface" warning to vehicles behind via V2X. Upon receiving the signal, the five vehicles behind automatically adjust their following strategies, forming a regional collaborative driving network to reduce the risk of chain-reaction rear-end collisions.

[0050] Example 2

[0051] refer to Figure 2 A vehicle-to-everything (V2X) intelligent driving dynamic control system for rainy weather, comprising:

[0052] Data fusion module: used for the acquisition, preprocessing and fusion of multi-source data, including a meteorological data interface unit, a V2X communication unit and an on-board sensor fusion unit; the meteorological data interface unit connects to the meteorological bureau's API via a 4G / 5G network to parse and store minute-level rainfall forecast data; the V2X communication unit, based on the IEEE 802.11p standard, enables communication with surrounding vehicles and roadside units to complete the transmission and reception of road condition data; the on-board sensor fusion unit connects to cameras, millimeter-wave radar and infrared sensors to perform noise reduction, synchronization calibration and feature extraction on the acquired raw data, and output unified environmental perception results;

[0053] Risk assessment module: used to dynamically calculate driving risks in rainy weather and generate adaptive strategies, including a machine learning model unit, a dynamic algorithm unit, and a risk level classification unit; the machine learning model unit deploys a trained LSTM+CNN hybrid neural network, inputs preprocessed multi-source data, and outputs a risk probability P; the dynamic algorithm unit calls a dynamic following distance algorithm, combines the current vehicle speed and road condition parameters to calculate the adaptive following distance; the risk level classification unit determines the risk level based on the risk probability P and maps it to the corresponding control strategy;

[0054] Control Execution Module: This module executes the control strategies output by the risk assessment module, including an ACC / AEB adjustment unit, an AR interaction unit, a human-machine collaboration unit, and a pedestrian safety unit. The ACC / AEB adjustment unit dynamically adjusts the following distance, AEB sensitivity, and speed limit based on the risk level. The AR interaction unit controls the display content of the AR-HUD to visualize warning information. The human-machine collaboration unit generates voice reminders, which are played through the vehicle's audio system, while simultaneously collecting driver behavior data for habit adaptation. The pedestrian safety unit receives pedestrian location data and controls AEB pre-braking, external speakers, and vehicle lights to achieve pedestrian safety protection.

[0055] The interaction flow of the system modules is as follows: The data fusion module collects real-time meteorological, vehicle-road-cloud and vehicle sensor data, and inputs it into the risk assessment module after preprocessing; The risk assessment module calculates the risk probability through a machine learning model, calls a dynamic algorithm to generate a following strategy, and outputs it to the control execution module; The control execution module adjusts the ACC / AEB parameters according to the strategy and provides feedback to the driver through AR-HUD and voice, forming a closed-loop control.

[0056] The above-described embodiments are only some embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the technical scope disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A dynamic control method for intelligent driving in rainy weather based on vehicle-to-everything (V2X) communication, characterized in that, The method includes: Multi-source data fusion perception: obtain minute-level rainfall forecast data from the meteorological bureau through the vehicle-to-everything (V2X) API, and combine it with vehicle GPS to locate rainfall areas in advance; obtain rain-related driving behavior data of other vehicles on the same road segment through the V2X communication protocol, and share the information collected by itself with surrounding vehicles; obtain real-time environmental data through vehicle cameras, radar and infrared sensors. Dynamic risk assessment is based on a hybrid LSTM+CNN neural network model built on the TensorFlow framework. The input includes rainfall forecast data, rainy driving behavior data, and real-time environmental data. The output is a risk probability P, which is divided into three risk levels: low, medium, and high. The adaptive following distance is calculated based on a dynamic following distance algorithm. Collaborative control execution dynamically adjusts ACC / AEB parameters based on risk level, including following distance, AEB sensitivity, and vehicle speed limit, and displays real-world warnings via AR-HUD overlay; it adapts system intervention strategies based on driver's rain-time driving habits; and it obtains authorized pedestrian locations through vehicle-to-everything (V2X) communication to trigger pre-braking and audible / visual alerts, achieving group collaboration, specifically including: AR Real-Scene Intelligent Warning: Warning information is overlaid on the driver's screen via AR-HUD, including red grid markings for road areas with a water film thickness >5mm, yellow offset arrows when the vehicle in front exhibits a skidding trajectory, and white virtual guide lines projected before entering a curve. Key warning information is accompanied by audible prompts. Driver Habit Adaptation: Driver's driving operation data in rainy weather is collected by onboard sensors. If the frequency of emergency braking is detected >3 times / 10 minutes, the AEB is automatically advanced from the default 1.5 seconds to 2.0 seconds; if the frequency of emergency braking is detected ≤1 time / 10 minutes, the system's default parameters are maintained to reduce unnecessary intervention. Pedestrian safety protection outside the vehicle: By receiving location data of pedestrians via a user-authorized mobile APP through the vehicle network, pedestrian targets can be identified 20 meters in advance in heavy rain, triggering the AEB system for pre-braking. At the same time, the vehicle's external speakers will play a voice reminder and the vehicle lights will flash to warn pedestrians that the vehicle is approaching.

2. The dynamic control method for intelligent driving in rainy weather based on vehicle-network collaboration as described in claim 1, characterized in that, The rainfall forecast data includes the rainfall amount for the next 10 minutes, the latitude and longitude boundaries of the rainfall area, and the trend of road surface temperature changes; the rainy driving behavior data of other vehicles on the same road segment includes tire slippage frequency, ABS intervention frequency, vehicle speed fluctuation data, and brake light triggering frequency; the real-time environmental data includes road surface water depth, water film thickness, road surface friction coefficient, visibility, and the status of traffic participants.

3. The dynamic control method for intelligent driving in rainy weather based on vehicle-network collaboration as described in claim 1, characterized in that, The risk probability P is divided into three risk levels: low, medium, and high. Specifically, when P < 30%, it is classified as low risk; when 30% ≤ P ≤ 70%, it is classified as medium risk; and when P > 70%, it is classified as high risk.

4. The dynamic control method for intelligent driving in rainy weather based on vehicle-network collaboration as described in claim 1, characterized in that, The formula for calculating the dynamic following distance algorithm is: D=D0×(1+k1×R+k2×S+k3×C); Where D0 is the standard following distance of the vehicle based on the current vehicle speed, R is the rainfall coefficient, S is the road surface slipperiness coefficient, C is the curvature coefficient, and k1, k2 and k3 are the weights of the rainfall coefficient, the road surface slipperiness coefficient and the curvature coefficient, respectively.

5. A vehicle-to-everything (V2X) intelligent driving dynamic control system for rainy weather, employing the intelligent driving dynamic control method for rainy weather as described in any one of claims 1-4, characterized in that, include: Data fusion module: used for the acquisition, preprocessing and fusion of multi-source data, including meteorological data interface unit, V2X communication unit and vehicle sensor fusion unit; The meteorological data interface unit connects to the meteorological bureau's API via a 4G / 5G network to parse and store minute-level rainfall forecast data; the V2X communication unit, based on the IEEE 802.11p standard, enables communication with surrounding vehicles and roadside units to transmit and receive road condition data; the vehicle-mounted sensor fusion unit connects to cameras, millimeter-wave radar, and infrared sensors to denoise, synchronize, and extract features from the collected raw data, outputting unified environmental perception results. Risk assessment module: used to dynamically calculate driving risks in rainy weather and generate appropriate strategies, including machine learning model unit, dynamic algorithm unit and risk level classification unit; The machine learning model unit deploys a trained LSTM+CNN hybrid neural network model, inputs preprocessed multi-source data, and outputs a risk probability P; the dynamic algorithm unit calls a dynamic following distance algorithm to calculate the appropriate following distance based on the current vehicle speed and road condition parameters; the risk level classification unit determines the risk level based on the risk probability P and maps it to the corresponding control strategy. Control Execution Module: This module executes the control strategies output by the risk assessment module, including an ACC / AEB adjustment unit, an AR interaction unit, a human-machine collaboration unit, and a pedestrian safety unit. The AR interaction unit controls the display content of the AR-HUD to visualize warning information. The human-machine collaboration unit generates voice reminders, which are played through the vehicle's audio system, while simultaneously collecting driver behavior data for habit adaptation. The pedestrian safety unit receives pedestrian location data and controls AEB pre-braking, external speakers, and vehicle lights to achieve pedestrian safety protection.

6. The vehicle-network cooperative intelligent driving dynamic control system for rainy weather as described in claim 5, characterized in that, The data fusion module also includes edge computing nodes, which are used to process V2X shared data in real time, generate dynamic road condition heat maps, intuitively display the distribution of road slippery risk, and support the push of regional risk warning information.

7. The vehicle-network cooperative intelligent driving dynamic control system for rainy weather as described in claim 5, characterized in that, The AR-HUD interface controlled by the AR interaction unit is divided into three areas: the upper area displays the water film thickness value and risk level text label, with high-risk information marked in red; the middle area is a real-scene overlay layer, displaying grid marks, offset arrows, and virtual guide lines; the lower area displays voice interaction prompts and driving mode suggestions, with key operation suggestions highlighted in red.

8. The vehicle-network cooperative intelligent driving dynamic control system for rainy weather as described in claim 5, characterized in that, The system also includes a data storage unit for storing historical meteorological data, road condition data, driver behavior data, and system operation logs. The storage capacity is ≥100GB, and the data retention time is ≥6 months, providing data support for the iterative optimization of the neural network model.

Images