Vehicle road safety interaction system based on Internet of Vehicles

Abstract

The invention discloses a vehicle road safety interaction system based on the Internet of Vehicles, and relates to the technical field of intelligent driving and safety assistance, and the system comprises an emergency response module which is in communication connection with a vehicle road data collection module, a vehicle road data encryption module and a vehicle road data analysis module; the vehicle road data acquisition module acquires vehicle driving data and road environment data through a vehicle-mounted sensor, observes the attention concentration degree of a driver through a biological recognition technology, and verifies the identities of a user and equipment; the vehicle road data encryption module visualizes original data, tracks and records user operation in the whole process by using log records and watermarks, and encrypts sensitive data by adopting a hybrid encryption algorithm; the vehicle road data analysis module judges whether to give a safety prompt to the driver according to the visibility and the driver state, and authenticates the user identity according to the Euclidean distance calculation result; and the emergency response module evaluates potential risks during vehicle driving, isolates a leakage source and sends out an early warning notification.

Description

Technical Field

[0001] The present invention relates to the field of intelligent driving and safety assistance technology, and in particular to a vehicle road safety interaction system based on an Internet of Vehicles. Background Art

[0002] Vehicle safety has always been a focus of attention for consumers and manufacturers. In the context of the rapid development of intelligent transportation, the vehicle road safety interaction system based on the Internet of Vehicles is gradually becoming an important technical means to improve road safety, optimize traffic flow, and enhance driving experience. The system integrates real-time communication between vehicles, data exchange with roadside infrastructure, and intelligent analysis of the cloud platform to achieve real-time monitoring of vehicle driving status, rapid sharing of road condition information, and automatic response in emergency situations, significantly improving the safety and efficiency of road traffic.

[0003] However, with the widespread application of Internet of Vehicles technology, the problem of user information leakage has become increasingly prominent and has become an important challenge that needs to be solved in this field. In the vehicle road safety interaction system, vehicles not only need to transmit their own driving status, location information and other key data in real time, but also need to frequently exchange data with other vehicles, roadside infrastructure and cloud platforms. In this process, security vulnerabilities in any link may lead to the leakage of user information, including sensitive data such as vehicle driving trajectory, driver identity, and vehicle identification information. The leakage of user information not only infringes on personal privacy, but also may cause a series of security issues. For example, attackers can use leaked vehicle information to implement criminal acts such as location tracking and vehicle theft; at the same time, by analyzing a large amount of vehicle driving data, attackers may also gain insight into the weak links of urban traffic and attack key infrastructure. Therefore, effective measures must be taken to prevent and solve the problem of user information leakage in the vehicle road safety interaction system based on the Internet of Vehicles. This includes strengthening data encryption technology, improving identity authentication and access control mechanisms, regularly scanning and evaluating security vulnerabilities, and strengthening user education and awareness. Only by comprehensively strengthening data security management can we ensure that Internet of Vehicles technology can fully protect the privacy rights of users while improving road safety. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides a vehicle road safety interaction system based on the Internet of Vehicles, which solves the problems in the above-mentioned background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a vehicle road safety interaction system based on the Internet of Vehicles, comprising an emergency response module, wherein the emergency module is communicatively connected to a vehicle road data acquisition module, a vehicle road data encryption module and a vehicle road data analysis module;

[0006] The vehicle road data acquisition module includes an on-board OBD system, coolant and oil temperature sensors, an on-board GPS, a weather collection system, a bioelectric sensor and face recognition technology to monitor the vehicle engine health status and vehicle real-time location information, observe the driver's concentration, verify the user's identity and verify the device information through registration;

[0007] The vehicle road data encryption module visualizes the collected raw data, uses log records and watermarks to track and record user operations throughout the process, and encrypts sensitive data using asymmetric encryption and symmetric encryption algorithms;

[0008] The vehicle road data analysis module is used to determine whether to provide a safety reminder to the driver based on the visibility prediction result and the final classification decision of the driver's state, and to obtain a user identity authentication result based on the calculated Euclidean distance;

[0009] The emergency response module evaluates potential risks during vehicle driving, and when abnormal user information access, transmission and storage behaviors are identified, verification and risk assessment are immediately performed, the source of leakage is isolated, and an early warning notification is issued.

[0010] Furthermore, the vehicle road data acquisition module obtains the vehicle engine health status through the vehicle OBD system, the coolant temperature sensor and the oil temperature sensor, including:

[0011] Utilizing the vehicle-mounted OBD system, the engine speed sensor detects the change in magnetic field generated when the ring gear on the flywheel passes the sensor, and the engine speed is calculated based on the principle of electromagnetic induction. The coolant temperature sensor and the oil temperature sensor monitor the temperature of the coolant and oil in the engine. When the coolant and oil flow through the corresponding sensors, the thermistor inside the corresponding sensor changes its resistance value as the coolant and oil temperature changes. This temperature change is converted into a voltage signal and sent to the ECU. The ECU adjusts the engine operating parameters based on the oil temperature information to keep the engine running at a normal temperature.

[0012] Furthermore, the vehicle road data acquisition module calculates the longitude, latitude and altitude of the vehicle through the vehicle-mounted GPS module, and the process of obtaining the real-time location information of the vehicle includes:

[0013] The vehicle-mounted GPS module is used to receive radio signals from four GPS satellites through its built-in antenna. The radio signals are filtered, amplified and converted to convert analog signals into digital signals. According to the timestamp and satellite orbit parameters in the satellite signal, the GPS module calculates the propagation time of the signal from the satellite to the receiver, and calculates the pseudorange containing the real geometric distance and atmospheric delay error. Based on the geometric relationship between the satellite position and the receiver position, the GPS module uses the pseudorange measurement values ​​of the four satellites to establish a navigation equation, and adopts the least squares method to solve the nonlinear equation group to obtain the optimal position estimate of the receiver. The current layer and tropospheric delay are corrected by additional observation data. The additional observation data includes atmospheric parameters provided by the ground meteorological station and atmospheric condition information obtained by the meteorological radar and the lidar. Finally, the GPS module converts the calculated longitude, latitude and geoid height into a projected coordinate system to obtain the real-time position information of the vehicle.

[0014] Furthermore, the vehicle road data collection module establishes the MLP model to output the visibility prediction value through the meteorological collection system, including:

[0015] The meteorological information collection system is used, and the visibility monitor in the system emits infrared light with a central wavelength of 870 nanometers and receives its scattered light. When the infrared light beam passes through the atmosphere, it scatters with suspended particles in the atmosphere. The received light signal is converted into an electrical signal by the photoelectric detector inside the monitor, and amplified and filtered. By comparing the level values ​​of the emitted light signal and the received light signal, the intensity difference between the two is calculated. The intensity difference directly reflects the attenuation degree of the light beam in the air. If the particle concentration in the atmosphere increases and the scattered light increases, the visibility decreases. If the particle concentration in the atmosphere decreases and the scattered light weakens, the visibility increases. The collected light signal data is transmitted to the vehicle road data collection module for centralized processing, and an MLP model is constructed to parse and calculate the preprocessed data. The MLP model includes an input layer, a hidden layer and an output layer. The input layer of the MLP model receives the particle concentration and light scattering intensity data in the atmosphere extracted from the meteorological information collection system. The hidden layer performs weighted summation and nonlinear activation on the input data through neurons to extract the deep features of the data. The output layer uses a linear activation function to directly output the visibility prediction value.

[0016] Furthermore, the vehicle road data acquisition module obtains the driver's brain wave signal through the bioelectric sensor, observes the driver's concentration level, verifies the user's identity through face recognition technology, and verifies the device information through registration, including:

[0017] Neuron activation generates ion current, which forms a potential difference inside and outside the cell. When a large number of neurons are activated at the same time, the potential difference generated is superimposed to form an EEG signal that can be detected by the bioelectric sensor. After being attenuated by the scalp, skull and brain tissue, the signal is captured by the electrodes placed on the scalp. Therefore, the bioelectric sensor is used to record the potential changes of the driver's cerebral cortex in real time in a non-invasive way, and the EEG signal is amplified, filtered and converted into a digital signal for transmission;

[0018] Using face recognition technology to verify the user's identity, the camera captures the driver's facial image in the form of analog signals, converts it into digital signals through an A / D converter, and extracts key facial features;

[0019] Registration is performed before the vehicle and related equipment are connected to the Internet of Vehicles system. During the registration process, the equipment submits its basic information to the registration agency of the Internet of Vehicles system. The basic information includes the equipment ID, manufacturer and model, and the registration agency verifies the submitted information.

[0020] Furthermore, the process of visualizing the collected raw data by the vehicle road data encryption module includes:

[0021] By performing fast Fourier transform on the EEG signal, the EEG signal is decomposed into components of different frequencies, and frequency domain features related to the degree of concentration are extracted, including alpha and beta waves of the EEG wave, frequency changes and variability of heart rate. The extracted features are combined with the three attention state labels of normal, fatigue and distraction to construct a training data set. The random forest algorithm is selected as the classifier and its parameters are initialized. The random forest integrates learning by constructing multiple decision trees. When constructing each tree, the algorithm randomly selects a part of the features and sample subsets as training data, recursively searches for the optimal split point at each node to maximize the node purity, and repeats the process until the predetermined node purity is met. In the random forest, each decision tree independently predicts the input data and outputs a prediction result. The random forest classifier summarizes the prediction results of all decision trees and gives the final classification decision by majority voting. Finally, the extracted features are input into the trained random forest classifier to perform real-time prediction of the attention state.

[0022] The neural network processes the input image layer by layer through multi-layer convolution and pooling operations, automatically extracts high-level features in the image that can represent the contour, texture and shape of the face, and converts the face image into a set of representative vectors. Through learning and training, the neural network algorithm automatically extracts facial features in the image, compares the extracted facial features with the user face template collected when the user registered and pre-stored in the database, and uses the Euclidean distance to calculate the distance between the extracted facial feature vector and the user face template feature vector. Since the Euclidean distance between multiple facial feature vectors of the same user will be relatively small, while the Euclidean distance between different users will be relatively large, a distance value that distinguishes most of the differences between different users is preliminarily selected as the preliminary threshold based on the Euclidean distance distribution between feature vectors.

[0023] Furthermore, the vehicle road data encryption module uses log recording and watermarking to track and record the user operation throughout the process, including:

[0024] The collected raw data is cleaned, compressed and converted into a unified format. The JSON log format is used to extract key information from the preprocessed data. The key information includes the timestamp of the operation, the operation type, the operation object and the operation result. According to the determined log format and the extracted key information, log entries are constructed. Each log entry contains complete information. The log entries are appended to the end of the log file using the file operation API in a programmatic way to generate log records.

[0025] Using watermark technology, a serial number is generated as the watermark information to be embedded, and the original data is subjected to DCT transformation. A specific DCT coefficient position is selected under the principle of having little impact on the quality of the original data and being able to resist various attack characteristics. The watermark information is binary-encoded using the DCT transform domain watermark algorithm and then embedded into the original data. The data is stored in a format suitable for DCT processing. In addition, by reading the data to be verified, the data to be verified is divided into multiple small data blocks, and a two-dimensional DCT transformation is performed on each block to obtain a DCT coefficient matrix. The position of the watermark information is located in the DCT coefficient matrix, and the watermark information is extracted from the located DCT coefficients. The extracted watermark information is compared with the binary bits of the original watermark information. When the comparison result tolerance rate is below 2%, it is considered that the original data contains valid watermark information; when the comparison result tolerance rate is 2% or above, it is considered that the original data does not contain valid watermark information.

[0026] Furthermore, the process of the vehicle road data encryption module using encryption technology to encrypt sensitive data includes:

[0027] Based on the collected original information, the system identifies sensitive data that needs to be encrypted, including vehicle location, driver identity and vehicle status information. During the key exchange phase, the server uses the key management system to generate an RSA key pair, and the vehicle terminal also generates a corresponding RSA key pair. The server sends its public key to the vehicle terminal, and the vehicle terminal sends its public key to the server. Both parties verify the authenticity of each other's public key through digital certificates. During the data encryption key negotiation phase, the vehicle terminal generates a random number and sends it to the server. The server encrypts the random number with the public key of the vehicle terminal and sends it to the vehicle terminal. The vehicle terminal uses its own private key to decrypt the received encrypted number. Random number. Based on this shared random number, both parties generate an AES key through the key derivation function KDF. In the data encryption stage, the generated AES key is used. The vehicle terminal encrypts the sensitive data to be transmitted through AES, and sends the encrypted data to the vehicle terminal through a secure communication channel. In the data decryption stage, after the receiver receives the encrypted data, it uses the previously negotiated AES key to decrypt it and verifies the integrity of the decrypted data through a hash function. The receiver uses the sender's public key to verify the signature to confirm the integrity and source of the data. In the digital signature stage, the sender uses its own private key to sign the hash value of the data and sends the signature together with the encrypted data to the receiver.

[0028] Furthermore, the vehicle road data analysis module determines whether to provide a safety reminder to the driver according to the visibility prediction result and the final classification decision of the driver's state, and obtains the user identity authentication result according to the calculated Euclidean distance, including:

[0029] In the process of observing road visibility, through iterative training, the MLP model gradually learns the relationship between attenuation and visibility. If the predicted visibility is less than 200 meters, the system considers that the current road visibility is poor and prompts the driver to slow down and turn on the fog lights. If the predicted visibility is 200 meters or more, the system considers that the current road visibility is good and does not give the driver a safety prompt.

[0030] When observing the driver's concentration, the real-time monitoring system determines the driver's concentration by comparing the driver's biological signal characteristics with the model in the random forest classifier. If the final classification decision shows that the driver is distracted and fatigued, the system will remind the driver through sound prompts to reduce the risk of accidents. If the final classification decision shows that the driver is in a normal state, no safety prompt will be given to the driver.

[0031] When using face recognition to authenticate users, when the calculated Euclidean distance is less than the preliminary threshold, the system considers that the face to be verified and the face in the user template belong to the same person, and the system allows the driver to perform subsequent operations and access services. When the Euclidean distance is greater than the preliminary threshold, it is considered that they do not belong to the same person, and the system denies access and issues a warning.

[0032] Furthermore, the process of the emergency response module evaluating the potential risks during vehicle driving, isolating the leakage source, and issuing an early warning notification includes:

[0033] Based on the various behavioral data collected from users in the network system, the behavioral data includes user operation records and data transmission records. The collected data is cleaned, duplicate, erroneous and irrelevant data are removed, and desensitization is performed. Useful features are extracted from the preprocessed data. The features should reflect the laws and characteristics of user behavior. The extracted features are normalized so that data with different features have the same scale to facilitate processing by the isolation forest algorithm. According to the principle of the isolation forest algorithm, an isolation tree is constructed. Each iTree is a binary tree structure. The data space is divided by randomly selecting features and cutting points until each subspace has only one data point. The extracted feature data is used to train the isolation forest model. During the training process, the model will learn the characteristics of normal user behavior patterns and construct a decision boundary that can distinguish normal behavior from abnormal behavior. An independent test set is used to evaluate the trained model, and the performance of the model in identifying abnormal behavior is checked. The model is optimized based on the evaluation results. Adjust the parameters of the isolation forest algorithm to improve the recognition accuracy and efficiency of the model. At the same time, deploy the trained isolation forest model to the Internet of Vehicles system to monitor user behavior in real time. When the model identifies abnormal information access, transmission and storage behaviors that are inconsistent with normal behavior patterns, the early warning mechanism is triggered. The system performs multi-source verification on the identified abnormal behaviors, evaluates the confirmed abnormal behaviors, determines the risk size, impact scope and potential losses, and evaluates the data sensitivity, system importance and attacker's intentions. If a potential source of information leakage is found, network isolation measures are taken immediately to prevent attackers from further access. The isolation measures include closing related ports, blocking IP addresses and isolating infected devices, removing malware from the affected systems, repairing system vulnerabilities and restoring data backups. The vehicle road safety interaction system also sends early warning notifications to users through on-board devices and mobile phone APPs to remind users of potential security risks and send early warning information to relevant departments.

[0034] The present invention provides a vehicle road safety interaction system based on the Internet of Vehicles. It has the following beneficial effects:

[0035] The system uses advanced encryption technology to encrypt the data transmitted by the vehicle, ensure the security and integrity of the data during the transmission process, and prevent the data from being illegally intercepted. Through a strict identity authentication mechanism, it is ensured that only authorized users can access and use the Internet of Vehicles system, effectively preventing unauthorized access and illegal operations. The system has a real-time monitoring function, which can timely detect potential security threats, such as illegal intrusion, data leakage, etc., and provide the possibility for timely response. When security incidents such as data leakage occur, the system can quickly launch an emergency plan and take effective measures to isolate, restore and report to minimize losses. By strengthening user privacy protection, users' trust in the Internet of Vehicles system is enhanced, and users' concerns about privacy leakage are reduced. On the premise of ensuring data security, the system provides more abundant and convenient application services, such as road condition prediction, emergency rescue and automatic driving assistance, to enhance users' driving experience and satisfaction. In the face of the security problem of user information leakage, the proposal of the present invention will promote technological innovation in the field of Internet of Vehicles in terms of data encryption, identity authentication, emergency response, etc. With the promotion and application of the present invention, it is expected to promote the Internet of Vehicles industry to establish more standardized and strict standards and systems in terms of data security and privacy protection. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is a block diagram of a vehicle road safety interaction system based on the Internet of Vehicles according to an embodiment of the present application. DETAILED DESCRIPTION

[0037] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0038] like Figure 1 As shown, the present invention provides a technical solution: a vehicle road safety interaction system based on the Internet of Vehicles, comprising an emergency response module, wherein the emergency module is communicatively connected to a vehicle road data acquisition module, a vehicle road data encryption module and a vehicle road data analysis module;

[0039] The vehicle road data acquisition module includes an on-board OBD system, coolant and oil temperature sensors, an on-board GPS, a weather collection system, a bioelectric sensor and face recognition technology to monitor the vehicle engine health status and vehicle real-time location information, observe the driver's concentration, verify the user's identity and verify the device information through registration;

[0040] The vehicle road data encryption module visualizes the collected raw data, uses log records and watermarks to track and record user operations throughout the process, and encrypts sensitive data using asymmetric encryption and symmetric encryption algorithms;

[0041] The vehicle road data analysis module is used to determine whether to provide a safety reminder to the driver based on the visibility prediction result and the final classification decision of the driver's state, and to obtain a user identity authentication result based on the calculated Euclidean distance;

[0042] The emergency response module evaluates potential risks during vehicle driving, and when abnormal user information access, transmission and storage behaviors are identified, verification and risk assessment are immediately performed, the source of leakage is isolated, and an early warning notification is issued.

[0043] The vehicle road data acquisition module obtains the vehicle engine health status through the vehicle OBD system, the coolant temperature sensor and the oil temperature sensor, including:

[0044] Utilizing the vehicle-mounted OBD system, the engine speed sensor detects the change in magnetic field generated when the ring gear on the flywheel passes the sensor, and the engine speed is calculated based on the principle of electromagnetic induction. The coolant temperature sensor and the oil temperature sensor monitor the temperature of the coolant and oil in the engine. When the coolant and oil flow through the corresponding sensors, the thermistor inside the corresponding sensor changes its resistance value as the coolant and oil temperature changes. This temperature change is converted into a voltage signal and sent to the ECU. The ECU adjusts the engine operating parameters based on the oil temperature information to keep the engine running at a normal temperature.

[0045] The vehicle road data acquisition module calculates the longitude, latitude and geodetic height of the vehicle through the vehicle-mounted GPS module, and the process of obtaining the real-time location information of the vehicle includes:

[0046] The vehicle-mounted GPS module is used to receive radio signals from four GPS satellites through its built-in antenna. The radio signals are filtered, amplified and converted to convert analog signals into digital signals. According to the timestamp and satellite orbit parameters in the satellite signal, the GPS module calculates the propagation time of the signal from the satellite to the receiver, and calculates the pseudorange containing the real geometric distance and atmospheric delay error. Based on the geometric relationship between the satellite position and the receiver position, the GPS module uses the pseudorange measurement values ​​of the four satellites to establish a navigation equation, and adopts the least squares method to solve the nonlinear equation group to obtain the optimal position estimate of the receiver. The current layer and tropospheric delay are corrected by additional observation data. The additional observation data includes atmospheric parameters provided by the ground meteorological station and atmospheric condition information obtained by the meteorological radar and the lidar. Finally, the GPS module converts the calculated longitude, latitude and geoid height into a projected coordinate system to obtain the real-time position information of the vehicle.

[0047] The process of the vehicle road data acquisition module establishing the MLP model output visibility prediction value through the meteorological acquisition system includes:

[0048] The meteorological information collection system is used, and the visibility monitor in the system emits infrared light with a central wavelength of 870 nanometers and receives its scattered light. When the infrared light beam passes through the atmosphere, it scatters with suspended particles in the atmosphere. The received light signal is converted into an electrical signal by the photoelectric detector inside the monitor, and amplified and filtered. By comparing the level values ​​of the emitted light signal and the received light signal, the intensity difference between the two is calculated. The intensity difference directly reflects the attenuation degree of the light beam in the air. If the particle concentration in the atmosphere increases and the scattered light increases, the visibility decreases. If the particle concentration in the atmosphere decreases and the scattered light weakens, the visibility increases. The collected light signal data is transmitted to the vehicle road data collection module for centralized processing, and an MLP model is constructed to parse and calculate the preprocessed data. The MLP model includes an input layer, a hidden layer and an output layer. The input layer of the MLP model receives the particle concentration and light scattering intensity data in the atmosphere extracted from the meteorological information collection system. The hidden layer performs weighted summation and nonlinear activation on the input data through neurons to extract the deep features of the data. The output layer uses a linear activation function to directly output the visibility prediction value.

[0049] The vehicle road data acquisition module obtains the driver's brain electrical signal through the bioelectric sensor, observes the driver's concentration level, verifies the user's identity through face recognition technology, and verifies the device information through registration, including:

[0050] Neuron activation generates ion current, which forms a potential difference inside and outside the cell. When a large number of neurons are activated at the same time, the potential difference generated is superimposed to form an EEG signal that can be detected by the bioelectric sensor. After being attenuated by the scalp, skull and brain tissue, the signal is captured by the electrodes placed on the scalp. Therefore, the bioelectric sensor is used to record the potential changes of the driver's cerebral cortex in real time in a non-invasive way, and the EEG signal is amplified, filtered and converted into a digital signal for transmission;

[0051] Using face recognition technology to verify the user's identity, the camera captures the driver's facial image in the form of analog signals, converts it into digital signals through an A / D converter, and extracts key facial features;

[0052] Registration is performed before the vehicle and related equipment are connected to the Internet of Vehicles system. During the registration process, the equipment submits its basic information to the registration agency of the Internet of Vehicles system. The basic information includes the equipment ID, manufacturer and model, and the registration agency verifies the submitted information.

[0053] The process of visualizing the collected raw data by the vehicle road data encryption module includes:

[0054] By performing fast Fourier transform on the EEG signal, the EEG signal is decomposed into components of different frequencies, and frequency domain features related to the degree of concentration are extracted, including alpha and beta waves of the EEG wave, frequency changes and variability of heart rate. The extracted features are combined with the three attention state labels of normal, fatigue and distraction to construct a training data set. The random forest algorithm is selected as the classifier and its parameters are initialized. The random forest integrates learning by constructing multiple decision trees. When constructing each tree, the algorithm randomly selects a part of the features and sample subsets as training data, recursively searches for the optimal split point at each node to maximize the node purity, and repeats the process until the predetermined node purity is met. In the random forest, each decision tree independently predicts the input data and outputs a prediction result. The random forest classifier summarizes the prediction results of all decision trees and gives the final classification decision by majority voting. Finally, the extracted features are input into the trained random forest classifier to perform real-time prediction of the attention state.

[0055] The neural network processes the input image layer by layer through multi-layer convolution and pooling operations, automatically extracts high-level features in the image that can represent the contour, texture and shape of the face, and converts the face image into a set of representative vectors. Through learning and training, the neural network algorithm automatically extracts facial features in the image, compares the extracted facial features with the user face template collected when the user registered and pre-stored in the database, and uses the Euclidean distance to calculate the distance between the extracted facial feature vector and the user face template feature vector. Since the Euclidean distance between multiple facial feature vectors of the same user will be relatively small, while the Euclidean distance between different users will be relatively large, a distance value that distinguishes most of the differences between different users is preliminarily selected as the preliminary threshold based on the Euclidean distance distribution between feature vectors.

[0056] The vehicle road data encryption module uses log recording and watermarking to track and record user operations throughout the process, including:

[0057] The collected raw data is cleaned, compressed and converted into a unified format. The JSON log format is used to extract key information from the preprocessed data. The key information includes the timestamp of the operation, the operation type, the operation object and the operation result. According to the determined log format and the extracted key information, log entries are constructed. Each log entry contains complete information. The log entries are appended to the end of the log file using the file operation API in a programmatic way to generate log records.

[0058] Using watermark technology, a serial number is generated as the watermark information to be embedded, and the original data is subjected to DCT transformation. A specific DCT coefficient position is selected under the principle of having little impact on the quality of the original data and being able to resist various attack characteristics. The watermark information is binary-encoded using the DCT transform domain watermark algorithm and then embedded into the original data. The data is stored in a format suitable for DCT processing. In addition, by reading the data to be verified, the data to be verified is divided into multiple small data blocks, and a two-dimensional DCT transformation is performed on each block to obtain a DCT coefficient matrix. The position of the watermark information is located in the DCT coefficient matrix, and the watermark information is extracted from the located DCT coefficients. The extracted watermark information is compared with the binary bits of the original watermark information. When the comparison result tolerance rate is below 2%, it is considered that the original data contains valid watermark information; when the comparison result tolerance rate is 2% or above, it is considered that the original data does not contain valid watermark information.

[0059] The process of the vehicle road data encryption module using encryption technology to encrypt sensitive data includes:

[0060] Based on the collected original information, the system identifies sensitive data that needs to be encrypted, including vehicle location, driver identity and vehicle status information. During the key exchange phase, the server uses the key management system to generate an RSA key pair, and the vehicle terminal also generates a corresponding RSA key pair. The server sends its public key to the vehicle terminal, and the vehicle terminal sends its public key to the server. Both parties verify the authenticity of each other's public key through digital certificates. During the data encryption key negotiation phase, the vehicle terminal generates a random number and sends it to the server. The server encrypts the random number with the public key of the vehicle terminal and sends it to the vehicle terminal. The vehicle terminal uses its own private key to decrypt the received encrypted number. Random number. Based on this shared random number, both parties generate an AES key through the key derivation function KDF. In the data encryption stage, the generated AES key is used. The vehicle terminal encrypts the sensitive data to be transmitted through AES, and sends the encrypted data to the vehicle terminal through a secure communication channel. In the data decryption stage, after the receiver receives the encrypted data, it uses the previously negotiated AES key to decrypt it and verifies the integrity of the decrypted data through a hash function. The receiver uses the sender's public key to verify the signature to confirm the integrity and source of the data. In the digital signature stage, the sender uses its own private key to sign the hash value of the data and sends the signature together with the encrypted data to the receiver.

[0061] The vehicle road data analysis module determines whether to provide a safety reminder to the driver based on the visibility prediction result and the final classification decision of the driver's state, and obtains the user identity authentication result based on the calculated Euclidean distance, including:

[0062] In the process of observing road visibility, through iterative training, the MLP model gradually learns the relationship between attenuation and visibility. If the predicted visibility is less than 200 meters, the system considers that the current road visibility is poor and prompts the driver to slow down and turn on the fog lights. If the predicted visibility is 200 meters or more, the system considers that the current road visibility is good and does not give the driver a safety prompt.

[0063] When observing the driver's concentration, the real-time monitoring system determines the driver's concentration by comparing the driver's biological signal characteristics with the model in the random forest classifier. If the final classification decision shows that the driver is distracted and fatigued, the system will remind the driver through sound prompts to reduce the risk of accidents. If the final classification decision shows that the driver is in a normal state, no safety prompt will be given to the driver.

[0064] When using face recognition to authenticate users, when the calculated Euclidean distance is less than the preliminary threshold, the system considers that the face to be verified and the face in the user template belong to the same person, and the system allows the driver to perform subsequent operations and access services. When the Euclidean distance is greater than the preliminary threshold, it is considered that they do not belong to the same person, and the system denies access and issues a warning.

[0065] The process of the emergency response module evaluating the potential risks during vehicle driving, isolating the leakage source, and issuing an early warning notification includes:

[0066] Based on the various behavioral data collected from users in the network system, the behavioral data includes user operation records and data transmission records. The collected data is cleaned, duplicate, erroneous and irrelevant data are removed, and desensitization is performed. Useful features are extracted from the preprocessed data. The features should reflect the laws and characteristics of user behavior. The extracted features are normalized so that data with different features have the same scale to facilitate processing by the isolation forest algorithm. According to the principle of the isolation forest algorithm, an isolation tree is constructed. Each iTree is a binary tree structure. The data space is divided by randomly selecting features and cutting points until each subspace has only one data point. The extracted feature data is used to train the isolation forest model. During the training process, the model will learn the characteristics of normal user behavior patterns and construct a decision boundary that can distinguish normal behavior from abnormal behavior. An independent test set is used to evaluate the trained model, and the performance of the model in identifying abnormal behavior is checked. The model is optimized based on the evaluation results. Adjust the parameters of the isolation forest algorithm to improve the recognition accuracy and efficiency of the model. At the same time, deploy the trained isolation forest model to the Internet of Vehicles system to monitor user behavior in real time. When the model identifies abnormal information access, transmission and storage behaviors that are inconsistent with normal behavior patterns, the early warning mechanism is triggered. The system performs multi-source verification on the identified abnormal behaviors, evaluates the confirmed abnormal behaviors, determines the risk size, impact scope and potential losses, and evaluates the data sensitivity, system importance and attacker's intentions. If a potential source of information leakage is found, network isolation measures are taken immediately to prevent attackers from further access. The isolation measures include closing related ports, blocking IP addresses and isolating infected devices, removing malware from the affected systems, repairing system vulnerabilities and restoring data backups. The vehicle road safety interaction system also sends early warning notifications to users through on-board devices and mobile phone APPs to remind users of potential security risks and send early warning information to relevant departments.

[0067] When using the present invention, the system collects vehicle driving and road environment data in real time through on-board sensors, combines biometric technology to monitor the driver's attention and verify the user and device identity, and then the data processing module visualizes the collected data, records the user's operation trajectory and encrypts sensitive information; the data analysis module judges the driver's status and road visibility based on the data, issues safety prompts in a timely manner, and verifies the user's identity through Euclidean distance calculation. The emergency response module continuously evaluates driving risks. Once abnormal information access, transmission and storage behavior is found, the verification procedure is immediately started to evaluate the risk level, isolate potential leakage sources, and immediately issue early warning notifications to ensure safe vehicle-road interaction.

[0068] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions. The sentence "includes an element defined by ... does not exclude the existence of other identical elements in the process, method, article or device including the element".

[0069] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A vehicle road safety interactive system based on the Internet of Vehicles, 1. comprising an emergency response module, characterized in that: The emergency module is communicatively connected to a vehicle road data acquisition module, a vehicle road data encryption module and a vehicle road data analysis module; The vehicle road data acquisition module includes an on-board OBD system, coolant and oil temperature sensors, an on-board GPS, a weather collection system, a bioelectric sensor and face recognition technology to monitor the vehicle engine health status and vehicle real-time location information, observe the driver's concentration, verify the user's identity and verify the device information through registration; The vehicle road data encryption module visualizes the collected raw data, uses log records and watermarks to track and record user operations throughout the process, and encrypts sensitive data using asymmetric encryption and symmetric encryption algorithms; The vehicle road data analysis module is used to determine whether to provide a safety reminder to the driver based on the visibility prediction result and the final classification decision of the driver's state, and to obtain a user identity authentication result based on the calculated Euclidean distance; The emergency response module evaluates potential risks during vehicle driving, and when abnormal user information access, transmission and storage behaviors are identified, verification and risk assessment are immediately performed, the source of leakage is isolated, and an early warning notification is issued.

2. The vehicle road safety interaction system based on the Internet of Vehicles according to claim 1, characterized in that: The vehicle road data acquisition module obtains the vehicle engine health status through the vehicle OBD system, the coolant temperature sensor and the oil temperature sensor, including: Utilizing the vehicle-mounted OBD system, the engine speed sensor detects the change in magnetic field generated when the ring gear on the flywheel passes the sensor, and the engine speed is calculated based on the principle of electromagnetic induction. The coolant temperature sensor and the oil temperature sensor monitor the temperature of the coolant and oil in the engine. When the coolant and oil flow through the corresponding sensors, the thermistor inside the corresponding sensor changes its resistance value as the coolant and oil temperature changes. This temperature change is converted into a voltage signal and sent to the ECU. The ECU adjusts the engine operating parameters based on the oil temperature information to keep the engine running at a normal temperature.

3. The vehicle road safety interaction system based on the Internet of Vehicles according to claim 2 is characterized in that: The vehicle road data acquisition module calculates the longitude, latitude and altitude of the vehicle through the vehicle-mounted GPS module, and the process of obtaining the real-time location information of the vehicle includes: The vehicle-mounted GPS module is used to receive radio signals from four GPS satellites through its built-in antenna. The radio signals are filtered, amplified and converted to convert analog signals into digital signals. According to the timestamp and satellite orbit parameters in the satellite signal, the GPS module calculates the propagation time of the signal from the satellite to the receiver, and calculates the pseudorange containing the real geometric distance and atmospheric delay error. Based on the geometric relationship between the satellite position and the receiver position, the GPS module uses the pseudorange measurement values ​​of the four satellites to establish a navigation equation, and adopts the least squares method to solve the nonlinear equation group to obtain the optimal position estimate of the receiver. The current layer and tropospheric delay are corrected by additional observation data. The additional observation data includes atmospheric parameters provided by the ground meteorological station and atmospheric condition information obtained by the meteorological radar and the lidar. Finally, the GPS module converts the calculated longitude, latitude and geoid height into a projected coordinate system to obtain the real-time position information of the vehicle.

4. The vehicle road safety interaction system based on the Internet of Vehicles according to claim 3 is characterized in that: The process of the vehicle road data acquisition module establishing the MLP model output visibility prediction value through the meteorological acquisition system includes: The meteorological information collection system is used, and the visibility monitor in the system emits infrared light with a central wavelength of 870 nanometers and receives its scattered light. When the infrared light beam passes through the atmosphere, it scatters with suspended particles in the atmosphere. The received light signal is converted into an electrical signal by the photoelectric detector inside the monitor, and amplified and filtered. By comparing the level values ​​of the emitted light signal and the received light signal, the intensity difference between the two is calculated. The intensity difference directly reflects the attenuation degree of the light beam in the air. If the particle concentration in the atmosphere increases and the scattered light increases, the visibility decreases. If the particle concentration in the atmosphere decreases and the scattered light weakens, the visibility increases. The collected light signal data is transmitted to the vehicle road data collection module for centralized processing, and an MLP model is constructed to parse and calculate the preprocessed data. The MLP model includes an input layer, a hidden layer and an output layer. The input layer of the MLP model receives the particle concentration and light scattering intensity data in the atmosphere extracted from the meteorological information collection system. The hidden layer performs weighted summation and nonlinear activation on the input data through neurons to extract the deep features of the data. The output layer uses a linear activation function to directly output the visibility prediction value.

5. The vehicle road safety interaction system based on the Internet of Vehicles according to claim 4 is characterized in that: The vehicle road data acquisition module obtains the driver's brain electrical signal through the bioelectric sensor, observes the driver's concentration level, verifies the user's identity through face recognition technology, and verifies the device information through registration, including: Neuron activation generates ion current, which forms a potential difference inside and outside the cell. When a large number of neurons are activated at the same time, the potential difference generated is superimposed to form an EEG signal that can be detected by the bioelectric sensor. After being attenuated by the scalp, skull and brain tissue, the signal is captured by the electrodes placed on the scalp. Therefore, the bioelectric sensor is used to record the potential changes of the driver's cerebral cortex in real time in a non-invasive way, and the EEG signal is amplified, filtered and converted into a digital signal for transmission; Using face recognition technology to verify the user's identity, the camera captures the driver's facial image in the form of analog signals, converts it into digital signals through an A / D converter, and extracts key facial features; Registration is performed before the vehicle and related equipment are connected to the Internet of Vehicles system. During the registration process, the equipment submits its basic information to the registration agency of the Internet of Vehicles system. The basic information includes the equipment ID, manufacturer and model, and the registration agency verifies the submitted information.

6. The vehicle road safety interaction system based on the Internet of Vehicles according to claim 5, characterized in that: The process of visualizing the collected raw data by the vehicle road data encryption module includes: By performing fast Fourier transform on the EEG signal, the EEG signal is decomposed into components of different frequencies, and frequency domain features related to the degree of concentration are extracted, including alpha and beta waves of the EEG wave, frequency changes and variability of heart rate. The extracted features are combined with the three attention state labels of normal, fatigue and distraction to construct a training data set. The random forest algorithm is selected as the classifier and its parameters are initialized. The random forest integrates learning by constructing multiple decision trees. When constructing each tree, the algorithm randomly selects a part of the features and sample subsets as training data, recursively searches for the optimal split point at each node to maximize the node purity, and repeats the process until the predetermined node purity is met. In the random forest, each decision tree independently predicts the input data and outputs a prediction result. The random forest classifier summarizes the prediction results of all decision trees and gives the final classification decision by majority voting. Finally, the extracted features are input into the trained random forest classifier to perform real-time prediction of the attention state. The neural network processes the input image layer by layer through multi-layer convolution and pooling operations, automatically extracts high-level features in the image that can represent the contour, texture and shape of the face, and converts the face image into a set of representative vectors. Through learning and training, the neural network algorithm automatically extracts facial features in the image, compares the extracted facial features with the user face template collected when the user registered and pre-stored in the database, and uses the Euclidean distance to calculate the distance between the extracted facial feature vector and the user face template feature vector. Since the Euclidean distance between multiple facial feature vectors of the same user will be relatively small, while the Euclidean distance between different users will be relatively large, a distance value that distinguishes most of the differences between different users is preliminarily selected as the preliminary threshold based on the Euclidean distance distribution between feature vectors.

7. The vehicle road safety interaction system based on the Internet of Vehicles according to claim 6, characterized in that: The vehicle road data encryption module uses log recording and watermarking to track and record user operations throughout the process, including: The collected raw data is cleaned, compressed and converted into a unified format. The JSON log format is used to extract key information from the preprocessed data. The key information includes the timestamp of the operation, the operation type, the operation object and the operation result. According to the determined log format and the extracted key information, log entries are constructed. Each log entry contains complete information. The log entries are appended to the end of the log file using the file operation API in a programmatic way to generate log records. Using watermark technology, a serial number is generated as the watermark information to be embedded, and the original data is subjected to DCT transformation. A specific DCT coefficient position is selected under the principle of having little impact on the quality of the original data and being able to resist various attack characteristics. The watermark information is binary-encoded using the DCT transform domain watermark algorithm and then embedded into the original data. The data is stored in a format suitable for DCT processing. In addition, by reading the data to be verified, the data to be verified is divided into multiple small data blocks, and a two-dimensional DCT transformation is performed on each block to obtain a DCT coefficient matrix. The position of the watermark information is located in the DCT coefficient matrix, and the watermark information is extracted from the located DCT coefficients. The extracted watermark information is compared with the binary bits of the original watermark information. When the comparison result tolerance rate is below 2%, it is considered that the original data contains valid watermark information; when the comparison result tolerance rate is 2% or above, it is considered that the original data does not contain valid watermark information.

8. The vehicle road safety interaction system based on the Internet of Vehicles according to claim 7, characterized in that: The process of the vehicle road data encryption module using encryption technology to encrypt sensitive data includes: Based on the collected original information, the system identifies sensitive data that needs to be encrypted, including vehicle location, driver identity and vehicle status information. During the key exchange phase, the server uses the key management system to generate an RSA key pair, and the vehicle terminal also generates a corresponding RSA key pair. The server sends its public key to the vehicle terminal, and the vehicle terminal sends its public key to the server. Both parties verify the authenticity of each other's public key through digital certificates. During the data encryption key negotiation phase, the vehicle terminal generates a random number and sends it to the server. The server encrypts the random number with the public key of the vehicle terminal and sends it to the vehicle terminal. The vehicle terminal uses its own private key to decrypt the received encrypted number. Random number. Based on this shared random number, both parties generate an AES key through the key derivation function KDF. In the data encryption stage, the generated AES key is used. The vehicle terminal encrypts the sensitive data to be transmitted through AES, and sends the encrypted data to the vehicle terminal through a secure communication channel. In the data decryption stage, after the receiver receives the encrypted data, it uses the previously negotiated AES key to decrypt it and verifies the integrity of the decrypted data through a hash function. The receiver uses the sender's public key to verify the signature to confirm the integrity and source of the data. In the digital signature stage, the sender uses its own private key to sign the hash value of the data and sends the signature together with the encrypted data to the receiver.

9. The vehicle road safety interaction system based on the Internet of Vehicles according to claim 8, characterized in that: The vehicle road data analysis module determines whether to provide a safety reminder to the driver based on the visibility prediction result and the final classification decision of the driver's state, and obtains the user identity authentication result based on the calculated Euclidean distance, including: In the process of observing road visibility, through iterative training, the MLP model gradually learns the relationship between attenuation and visibility. If the predicted visibility is less than 200 meters, the system considers that the current road visibility is poor and prompts the driver to slow down and turn on the fog lights. If the predicted visibility is 200 meters or more, the system considers that the current road visibility is good and does not give the driver a safety prompt. When observing the driver's concentration, the real-time monitoring system determines the driver's concentration by comparing the driver's biological signal characteristics with the model in the random forest classifier. If the final classification decision shows that the driver is distracted and fatigued, the system will remind the driver through sound prompts to reduce the risk of accidents. If the final classification decision shows that the driver is in a normal state, no safety prompt will be given to the driver. When using face recognition to authenticate users, when the calculated Euclidean distance is less than the preliminary threshold, the system considers that the face to be verified and the face in the user template belong to the same person, and the system allows the driver to perform subsequent operations and access services. When the Euclidean distance is greater than the preliminary threshold, it is considered that they do not belong to the same person, and the system denies access and issues a warning.

10. The vehicle road safety interaction system based on the Internet of Vehicles according to claim 9, characterized in that: The process of the emergency response module evaluating the potential risks during vehicle driving, isolating the leakage source, and issuing an early warning notification includes: Based on the various behavioral data collected from users in the network system, the behavioral data includes user operation records and data transmission records. The collected data is cleaned, duplicate, erroneous and irrelevant data are removed, and desensitization is performed. Useful features are extracted from the preprocessed data. The features should reflect the laws and characteristics of user behavior. The extracted features are normalized so that data with different features have the same scale to facilitate processing by the isolation forest algorithm. According to the principle of the isolation forest algorithm, an isolation tree is constructed. Each iTree is a binary tree structure. The data space is divided by randomly selecting features and cutting points until each subspace has only one data point. The extracted feature data is used to train the isolation forest model. During the training process, the model will learn the characteristics of normal user behavior patterns and construct a decision boundary that can distinguish normal behavior from abnormal behavior. An independent test set is used to evaluate the trained model, and the performance of the model in identifying abnormal behavior is checked. The model is optimized based on the evaluation results. Adjust the parameters of the isolation forest algorithm to improve the recognition accuracy and efficiency of the model. At the same time, deploy the trained isolation forest model to the Internet of Vehicles system to monitor user behavior in real time. When the model identifies abnormal information access, transmission and storage behaviors that are inconsistent with normal behavior patterns, the early warning mechanism is triggered. The system performs multi-source verification on the identified abnormal behaviors, evaluates the confirmed abnormal behaviors, determines the risk size, impact scope and potential losses, and evaluates the data sensitivity, system importance and attacker's intentions. If a potential source of information leakage is found, network isolation measures are taken immediately to prevent attackers from further access. The isolation measures include closing related ports, blocking IP addresses and isolating infected devices, removing malware from the affected systems, repairing system vulnerabilities and restoring data backups. The vehicle road safety interaction system also sends early warning notifications to users through on-board devices and mobile phone APPs to remind users of potential security risks and send early warning information to relevant departments.

Images