Intelligent network connection automobile safety brain system based on automobile-cloud interaction

Through the intelligent connected vehicle safety brain system of vehicle-cloud interaction, combined with the collaborative work between the vehicle and the cloud, it can identify and deal with the safety risks of autonomous vehicles in real time, solve the problem of insufficient traditional safety protection and realize the self-driving vehicle All-round safety monitoring and protection.

CN118977731AActive Publication Date: 2024-11-19国家市场监督管理总局缺陷产品召回技术中心 +1
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
CN202411068518.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-06
Publication Date
2024-11-19
Estimated Expiration
2044-08-06

AI Technical Summary

Technical Problem

Traditional safety protection methods are not enough to provide comprehensive safety protection for autonomous vehicles and cannot effectively respond to the growing security challenges, especially in complex traffic scenarios and information security.

Method used

Adopt intelligent connected car safety brain system based on car-cloud interaction, through the collaborative work of the vehicle-end and cloud-end security brain system, it can identify and respond to functional safety risks and expected functional safety risks in real time, combine big data analysis for all-round monitoring and optimization Safety protection of autonomous vehicles.

Benefits of technology

Real-time comprehensive safety protection for autonomous vehicles is achieved, the level of safety is improved, and the rapid response of the vehicle end and the powerful computing power of the cloud complement each other, the contradiction between real-time response and big data analysis is solved, and more comprehensive security guarantees are provided.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118977731A_ABST
    Figure CN118977731A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of automatic driving, in particular to an intelligent network connection automobile safety brain system based on automobile-cloud interaction, which comprises an automobile end safety brain system arranged at an automobile end and a cloud end safety brain system arranged at a cloud end, the vehicle end safety brain system is used for acquiring vehicle safety information from the vehicle networking system in real time to carry out expected function safety risk identification and function safety risk identification, determining whether emergency disposal is carried out or not according to an identification result, and if yes, sending generated emergency disposal information to the vehicle execution system to carry out emergency control of the automatic driving vehicle; and the cloud safety brain system is used for carrying out single-vehicle safety analysis and multi-vehicle safety analysis on the automatic driving vehicle and further carrying out comprehensive safety monitoring. According to the scheme, the vehicle-end safety brain system is responsible for real-time processing and emergency response, the cloud-end safety brain system is responsible for big data analysis and safety monitoring, and the vehicle-end safety brain system and the cloud-end safety brain system complement each other, so that comprehensive safety guarantee can be provided for the automatic driving vehicle.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of autonomous driving technology, and in particular to a smart connected vehicle safety brain system based on vehicle-cloud interaction. Background Art

[0002] The widespread application of artificial intelligence technology has promoted the rapid development of vehicle intelligence, and autonomous driving is a typical example of vehicle intelligence. However, the long-tail problem, edge scenarios, and unexplainability of artificial intelligence technology have made the safety of autonomous vehicles a key issue that needs to be solved urgently. In addition, with the development of communication technology, vehicles have become a key link in the communication network, which makes traffic scenarios more complex. Vehicles not only have to interact with the physical environment, but also have to exchange a large amount of data with the outside world, making the information security of automobiles particularly urgent.

[0003] In this context, traditional safety protection measures may no longer be sufficient to cope with the growing security challenges, and there is an urgent need for systematic and comprehensive protection of the safety of autonomous vehicles. The rapid development of Internet of Vehicles technology has provided cars with the ability to communicate with the external environment in real time, and the application of cloud computing and big data analysis has provided new tools for predicting and preventing potential security risks.

[0004] Therefore, there is an urgent need for an intelligent connected vehicle safety brain system based on vehicle-cloud interaction. Summary of the invention

[0005] In order to solve the problem that traditional safety protection measures are not sufficient to provide comprehensive safety protection for autonomous driving vehicles, an embodiment of the present invention provides an intelligent connected vehicle safety brain system based on vehicle-cloud interaction.

[0006] In a first aspect, an embodiment of the present invention provides a smart connected vehicle safety brain system based on vehicle-cloud interaction, comprising: a vehicle-side safety brain system arranged at a vehicle-side and a cloud-side safety brain system arranged at a cloud-side, wherein the vehicle-side safety brain system and the cloud-side safety brain system are respectively connected to the original vehicle system of the smart connected vehicle, and the vehicle-side safety brain system and the cloud-side safety brain system are connected to each other; the original vehicle system of the smart connected vehicle includes a vehicle networking system, a vehicle automatic driving system, and a vehicle execution system;

[0007] The vehicle-side safety brain system is used to obtain vehicle safety information from the vehicle networking system in real time to identify expected functional safety risks and functional safety risks, and determine whether to perform emergency disposal according to the identification results. If so, the generated emergency disposal information is sent to the vehicle execution system to perform emergency control of the autonomous driving vehicle;

[0008] The cloud-based safety brain system is used to obtain fault information, autonomous driving performance data, and driver's operating behavior data of the autonomous driving vehicle from the Internet of Vehicles system and the vehicle-side safety brain system, so as to perform single-vehicle safety analysis on the autonomous driving vehicle, as well as multi-vehicle safety analysis on the same model of autonomous driving vehicles, and further conduct comprehensive safety monitoring of the autonomous driving vehicles.

[0009] In a second aspect, an embodiment of the present invention further provides a method for safety monitoring of an intelligent connected vehicle, comprising:

[0010] The vehicle-side safety brain system is used to obtain vehicle safety information from the Internet of Vehicles system in real time to identify expected functional safety risks and functional safety risks, and determine whether to perform emergency measures based on the identification results. If so, the generated emergency measures information is sent to the vehicle execution system to perform emergency control of the autonomous driving vehicle;

[0011] The cloud-based safety brain system is used to obtain the fault information, autonomous driving performance data and driver's operating behavior data of the autonomous driving vehicle from the Internet of Vehicles system and the vehicle-side safety brain system, so as to conduct single-vehicle safety analysis on the autonomous driving vehicle and multi-vehicle safety analysis on the same model of autonomous driving vehicles, and further conduct comprehensive safety monitoring of the autonomous driving vehicles.

[0012] An embodiment of the present invention provides an intelligent connected vehicle safety brain system based on vehicle-cloud interaction. The vehicle-side safety brain system identifies expected functional safety risks and functional safety risks in real time. When it is determined that emergency disposal is required, the vehicle-side safety brain system takes precedence over the control of the vehicle's automatic driving system, and uses the emergency disposal information generated by the vehicle-side safety brain system to perform emergency control of the vehicle. The vehicle-side safety brain system of this solution can consider functional safety issues and expected functional safety issues at the same time, and perform real-time comprehensive safety protection for automatic driving vehicles.

[0013] The cloud-based safety brain system conducts single-vehicle safety analysis on autonomous vehicles and multi-vehicle safety analysis on massive autonomous vehicles. It uses the advantages of big data to comprehensively monitor the operating status of intelligent connected vehicles, identify risks and determine defects, and improve the safety level of autonomous vehicles.

[0014] Therefore, this solution combines the rapid real-time response capability of the vehicle-side safety brain system with the powerful computing power of the cloud-side safety brain system, forming a complementary advantage of vehicle-cloud collaboration, effectively resolving the contradiction between real-time response and big data analysis, and thus providing more comprehensive safety protection for autonomous driving vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 It is a schematic diagram of the composition of a smart connected vehicle safety brain system based on vehicle-cloud interaction provided by one embodiment of the present invention;

[0017] Figure 2 It is a schematic diagram of the composition of a vehicle-side safety brain system provided by an embodiment of the present invention;

[0018] Figure 3 It is a schematic diagram of the composition of a cloud security brain system provided by one embodiment of the present invention;

[0019] Figure 4 It is a flow chart of a method for safety monitoring of an intelligent connected vehicle provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0021] Please refer to Figure 1 The embodiment of the present invention provides a smart connected vehicle safety brain system based on vehicle-cloud interaction, comprising: a vehicle-side safety brain system arranged on a vehicle side and a cloud-side safety brain system arranged on a cloud side, wherein the vehicle-side safety brain system and the cloud-side safety brain system are respectively connected to the original vehicle system of the smart connected vehicle, and the vehicle-side safety brain system and the cloud-side safety brain system are connected to each other; the original vehicle system of the smart connected vehicle includes a vehicle networking system, a vehicle automatic driving system and a vehicle execution system;

[0022] The vehicle-side safety brain system is used to obtain vehicle safety information from the Internet of Vehicles system in real time to identify expected functional safety risks and functional safety risks, and determine whether to take emergency measures based on the identification results. If so, the generated emergency measures information will be sent to the vehicle execution system to perform emergency control of the autonomous driving vehicle;

[0023] The cloud-based safety brain system is used to obtain fault information, autonomous driving performance data, and driver's operating behavior data of autonomous driving vehicles from the Internet of Vehicles system and the vehicle-side safety brain system, so as to conduct single-vehicle safety analysis on autonomous driving vehicles, as well as multi-vehicle safety analysis on autonomous driving vehicles of the same model, and further conduct comprehensive safety monitoring of autonomous driving vehicles.

[0024] In an embodiment of the present invention, the vehicle-side safety brain system performs expected functional safety risk identification and functional safety risk identification in real time. When it is determined that emergency disposal is required, the vehicle-side safety brain system takes precedence over the control of the vehicle's automatic driving system and uses the emergency disposal information generated by the vehicle-side safety brain system to perform emergency control of the vehicle. The vehicle-side safety brain system of this solution can consider functional safety issues and expected functional safety issues at the same time, and perform real-time comprehensive safety protection for automatic driving vehicles.

[0025] The cloud-based safety brain system conducts single-vehicle safety analysis on autonomous vehicles and multi-vehicle safety analysis on massive autonomous vehicles. It uses the advantages of big data to comprehensively monitor the operating status of intelligent connected vehicles, identify risks and determine defects, and improve the safety level of autonomous vehicles.

[0026] Therefore, this solution combines the rapid real-time response capability of the vehicle-side safety brain system with the powerful computing power of the cloud-side safety brain system, forming a complementary advantage of vehicle-cloud collaboration, effectively resolving the contradiction between real-time response and big data analysis, and thus providing more comprehensive safety protection for autonomous driving vehicles.

[0027] In an embodiment of the present invention, the vehicle-side security brain system further includes a data encryption module, which is used to encrypt the data required by the cloud security brain system and transmit it to the cloud security brain system;

[0028] The cloud-based security brain system also includes a data parsing module, which is used to parse the encrypted data transmitted by the vehicle-side security brain system.

[0029] In this embodiment, the data encryption module in the vehicle-side security brain system can prevent potential security threats and information tampering during the data transmission and storage life cycle, providing a higher level of security for the entire system. The data analysis module of the cloud security brain system uses data decryption technologies such as symmetric decryption algorithms and asymmetric decryption algorithms to analyze the encrypted data transmitted in the vehicle-side security brain system, and can provide data for other modules.

[0030] Next, the vehicle-side safety brain system and the cloud-side safety brain system are explained respectively.

[0031] (1) Vehicle-side safety brain system

[0032] First, please refer to Figure 1 and Figure 2 The vehicle-side safety brain system includes: risk identification module, safety and defect analysis module and emergency response module;

[0033] The risk identification module is connected to the original vehicle system of the vehicle and is used to obtain vehicle safety information from the original vehicle system to identify expected functional safety risks and functional safety risks; the vehicle safety information includes the vehicle state vector and the state vectors of surrounding traffic participants used for scenario risk assessment, the operation data used to monitor the driver's human-computer interaction behavior, and the status signals of the sensors, communications and controllers of the vehicle's automatic driving system;

[0034] The safety and defect analysis module is used to analyze and trace faults based on the real-time identification results of the risk identification module, and to calculate and classify the comprehensive risk value to determine whether to take emergency measures based on the classification results. If so, an emergency response strategy is formulated based on the vehicle safety information and the classification results.

[0035] The emergency response module is used to generate emergency response information according to the emergency response strategy formulated by the safety and defect analysis module and send it to the vehicle execution system to perform emergency control of the autonomous driving vehicle.

[0036] In this embodiment, the expected functional safety problem refers to the hazards and risks caused by the lack of multiple functions of the autonomous driving system such as perception, decision-making and control in complex traffic scenarios, as well as the hazards and risks caused by reasonable and foreseeable human misuse; the functional safety problem refers to the problem caused by system failure. Since there is currently a lack of a comprehensive safety protection system that can simultaneously consider the above two safety issues, the vehicle-side safety brain system is equipped with a risk identification module to perform real-time expected functional safety risk identification and functional safety risk identification, including identifying complex dangerous traffic scenarios, identifying driver misoperation and vehicle autonomous driving system failures. Then, the safety and defect analysis module is used to evaluate the comprehensive risk value and risk level. When the risk level is high, an emergency response strategy is formulated based on the vehicle safety information and classification results, so that the emergency response module is used to generate emergency response information to the vehicle execution system, giving priority to the control of the vehicle autonomous driving system, and performing emergency control of the vehicle. Therefore, this solution can simultaneously consider functional safety issues and expected functional safety issues to provide comprehensive safety protection for autonomous driving vehicles, and can provide solid protection for the safe operation of autonomous driving vehicles.

[0037] In some embodiments, reference Figure 2 ,The risk identification module includes a vehicle driving risk identification module, a human-computer interaction monitoring module and a vehicle functional safety risk identification module;

[0038] The vehicle driving risk identification module is used to evaluate the driving risk value in the current traffic scenario based on the risk model, the vehicle state vector and the state vectors of surrounding traffic participants;

[0039] The human-machine interaction monitoring module is used to monitor the operation data of the driver's human-machine interaction behavior to identify the driver's misuse or improper operation behavior and determine the personnel misuse factor;

[0040] The vehicle functional safety risk identification module is used to detect the status signals of sensors, communications and controllers of the vehicle's autonomous driving system, and determine the fault results and functional safety risk values.

[0041] In this embodiment, the vehicle driving risk identification module uses the vehicle state vector collected by the original vehicle system for scene risk assessment and the state vectors of surrounding traffic participants for analysis. The vehicle state vector includes information such as position, heading angle, speed and acceleration. At the same time, similar information of surrounding traffic participants is collected to form multiple state vectors. Then, based on the risk model, the driving risk value of factors such as relative distance, speed, potential collision points between vehicles in the current traffic scenario is evaluated to identify potential dangers or abnormal behaviors. Once it is detected that the driving risk value exceeds the safety threshold, a warning message is immediately sent to the safety and defect analysis module.

[0042] In addition, the module also takes into account environmental factors such as weather conditions, road types, traffic signs and signal lights, making risk identification more comprehensive. The human-machine interaction monitoring module monitors the driver's and passengers' actions of operating vehicle controls, including the moment, force and duration of stepping on the accelerator and brake pedals, the moment, torque and direction of steering wheel operation, and the actions of operating the vehicle system, including key force, key frequency and operation duration, to identify the driver's misuse or improper operation behavior and determine the personnel misuse factor.

[0043] The vehicle functional safety risk identification module is responsible for diagnosing sensor, communication and controller faults of the autonomous driving system, and transmitting the situation to the safety and defect analysis module in real time to ensure the stability and reliability of the system. Among them, sensor fault detection reads and analyzes the fault signals sent by the sensor, including monitoring communication errors, state failures, state blockages and state overtemperatures, to perform fault detection; when detecting communication and controller faults, the vehicle functional safety risk identification module uses advanced communication technology to monitor CAN signals in real time. When a CAN fault occurs, the system can quickly identify the error state in the CAN network and accurately locate the fault through the LabVIEW signal processing program. At the same time, the LabVIEW program also uses virtual instruments (VI) to further detect and diagnose errors on the CAN signal to ensure the stability and reliability of the communication network. The three modules work closely together to perform comprehensive risk identification of expected functional safety and functional safety.

[0044] In some embodiments, the driving risk value is assessed in the following manner:

[0045] Determine the risk indicator value ITTC based on the time to collision (TTC):

[0046]

[0047] Where D x is the lateral distance between the vehicle and other vehicles, D y is the longitudinal distance between the vehicle and other vehicles, L w is the vehicle width, L l is the length of the vehicle, V rx V is the relative lateral speed between the vehicle and other vehicles, ry A is the relative longitudinal speed between the vehicle and other vehicles, rx A is the relative lateral acceleration between the vehicle and the other vehicle, ry is the longitudinal relative acceleration between the vehicle and other vehicles, a is the influence coefficient of the relative acceleration between the vehicle and other vehicles;

[0048] The ratio of the vehicle's speed to the road speed limit is used as the correlation coefficient λ for calculating the driving risk value. The calculation formula is:

[0049]

[0050] In the formula, v is the speed of the vehicle, v l is the road speed limit, for example, the maximum speed on a highway cannot exceed 120km / h, then v l is 120. The larger the λ is, the faster the vehicle speed is relative to the road speed limit, and the greater the impact on the driving risk value;

[0051] Then the driving risk value r under the current traffic scenario is:

[0052] r = exp(-λ·ITTC).

[0053] In some embodiments, the human misuse factor is determined as follows: if in the current detection cycle, the human-machine interaction monitoring module identifies that the driver has misused or improperly operated the device, the human misuse factor is determined to be 2; if no human misuse occurs, the human misuse factor is 1.

[0054] In some embodiments, the functional safety risk value is determined by referring to the ASIL automobile safety integrity level, and evaluating the functional safety risk value based on three factors: severity (Severity), exposure rate (Exposure) and controllability (Controllability) of the fault determined by the vehicle functional safety risk identification module in the current detection cycle. The functional safety risk value can be 1, 2, 3, and 4 from low to high.

[0055] In some embodiments, the vehicle functional safety risk identification module is also used to adjust the fault detection priority based on the high-risk traffic scenario when the vehicle driving risk identification module assesses a high-risk traffic scenario, so as to more quickly identify the source of the fault and prioritize the operation of key safety components in the current traffic scenario.

[0056] In some embodiments, the safety and defect analysis module includes at least a fault tracing analysis module, a defect analysis module, a safety calculation and classification module, and an emergency response strategy module;

[0057] The input end of the fault tracing analysis module is connected to the vehicle functional safety risk identification module, which uses a hybrid algorithm based on physical models and data-driven to establish a system fault propagation model and use artificial intelligence technology based on long short-term memory network (LSTM) to calculate the causal relationship between time series, locate the fault location and type, and complete fault tracing;

[0058] The input end of the defect analysis module is connected to the vehicle functional safety risk identification module, which is used to collect vehicle defect clues through multiple channels, and use artificial intelligence technology based on supervised learning to perform defect correlation analysis, classify the same or similar faults that occur in different safety events of the vehicle, and then use association analysis to mine and summarize the rules with common attributes in multi-source data to determine the root cause of the fault;

[0059] The safety calculation and classification module is used to determine the comprehensive risk value based on the driving risk value output by the vehicle driving risk identification module, the personnel misuse factor output by the human-machine interaction monitoring module and the functional safety risk value output by the vehicle functional safety risk identification module, and complete the safety classification through the clustering algorithm based on fuzzy theory;

[0060] The emergency response strategy module is used to formulate a driving path based on vehicle safety information and determine the braking urgency of the driving path based on the classification results.

[0061] In this embodiment, when the vehicle functional safety risk identification module identifies that there is a fault in the vehicle's automatic driving system, the fault tracing analysis module and the defect analysis module trace and analyze the fault. If no fault is identified, the fault tracing analysis module and the defect analysis module do not run.

[0062] The safety calculation and classification module determines the comprehensive risk value R in the following way:

[0063] R=β×γ×r

[0064] In the formula, β is the functional safety risk value, γ is the personnel misuse factor, and r is the driving risk value. The larger the comprehensive risk value R is, the higher the vehicle risk level is.

[0065] In addition, the emergency response strategy module not only considers the current risk level and fault status, but also predicts the risk development trend within a certain prediction time window in the future, so as to make a more predictive safety analysis. Based on the analysis and prediction, it determines the safe driving path, as well as deceleration, lane change, parking or initiation of emergency avoidance procedures. At the same time, in special circumstances, if the system detects a serious fault or irreversible risk, the emergency response strategy module may also decide to safely guide the vehicle to a parking state and initiate an emergency assistance procedure.

[0066] In some embodiments, the security and defect analysis module further includes a security log generation module;

[0067] The input end of the security log generation module is connected to the risk identification module, the fault tracing analysis module, the defect analysis module, the security calculation and classification module and the emergency response strategy module, and the output end is connected to the security brain system located in the cloud; the security log generation module is used to record the vehicle operating status, driver operation data, and vehicle automatic driving system failures when a security incident occurs, and after being encrypted by the data encryption module, it is reported to the cloud security brain system for big data analysis.

[0068] In some embodiments, the emergency response module includes at least: a lateral response module and a longitudinal response module;

[0069] The lateral response module is used to generate emergency response information for the vehicle steering system based on the emergency response strategy to control the direction of the vehicle;

[0070] The longitudinal response module is used to generate emergency response information for vehicle speed and acceleration based on the emergency response strategy to control the vehicle's power output and braking force output.

[0071] In this embodiment, the lateral response module is responsible for the direction and path following of the vehicle, including operations such as steering, lane changing and lane keeping. By controlling the vehicle's steering system, it ensures that the vehicle can accurately drive along the safe driving path set by the emergency response strategy module, so as to maintain the stability of the vehicle and maintain good handling performance in emergency avoidance or complex road conditions. The longitudinal response module mainly adjusts the speed and acceleration of the vehicle, including operations such as acceleration, deceleration and emergency braking. When the emergency response strategy module determines that the speed needs to be changed or stopped, the longitudinal control will intervene and control the power output and braking force output of the vehicle.

[0072] In some embodiments, the emergency response module also includes a safety warning module; the safety warning module obtains risk information from the safety and defect analysis module, and provides safety warning information to the driver in an emergency situation through head-up display (HUD) information prompts, steering wheel and seat vibrations, and voice prompts inside the vehicle, to provide safety warnings and feedback.

[0073] In some embodiments, the emergency response module is also used to: when a fault causes the communication data between the systems to fail, use the vehicle safety information at the last moment before the fault occurs to generate emergency response information to control the vehicle.

[0074] In this embodiment, the emergency response module, as the last line of defense for the safety protection system, has strong fault tolerance and robustness in extreme cases, in addition to executing the conventional functions of the emergency response strategy output of the safety and defect analysis module. Even if the key data from the safety and defect analysis module is missing, the vehicle's safety response can be achieved only through the vehicle's own vehicle status information. When the failure of the automatic driving system or the internal failure of the system of the embodiment of the present invention causes the communication data between the systems to fail, the vehicle safety information at the last moment before the failure occurs is used to predict the surrounding environment based on the kinematic model and control the vehicle.

[0075] In an embodiment of the present invention, when a fault causes the communication data between systems to fail, the longitudinal response module is used to calculate the safe driving distance in real time using the deceleration model, and decelerate based on the sliding mode control algorithm to stop the vehicle within a safe distance; the lateral response module uses the effective expected path finally issued by the vehicle's automatic driving system to control the vehicle's steering system, thereby ensuring basic safety response functions.

[0076] Therefore, the vehicle-side safety brain system for autonomous vehicles in the embodiment of the present invention covers three key links: risk identification, safety and defect analysis, and emergency response. In the risk identification part, the system implements a comprehensive risk identification strategy that can accurately and real-time detect and evaluate the safety risks of traffic scenes, monitor human-computer interaction behaviors, and perform fault diagnosis on various functions of the autonomous driving system. The safety and defect analysis module comprehensively analyzes the traffic scene data, human-computer interaction behavior data, and vehicle failure conditions obtained from the risk identification module, and performs a comprehensive analysis based on this information, calculates the comprehensive risk value and safety level, intelligently selects the most appropriate safety strategy, and generates a safety log to cope with the current traffic environment and system status. Furthermore, the safety and defect analysis module can also complete fault tracing for factors such as fault location and type, use artificial intelligence technology to perform correlation analysis, accurately identify the associated factors of defect occurrence, and the generation of safety logs can record vehicle operation data, event information, and driver and passenger behavior, providing a reliable data source for accident reconstruction. In terms of emergency response, on the basis of conventional functions, when the autonomous driving system and the safety and defect analysis module fail, the vehicle can still be effectively controlled, thanks to the reference model embedded in the emergency response module, which can be safely controlled even in extreme situations. In addition, there is a safety warning module that issues safety warnings in emergency situations, prompting the driver and passengers to take necessary operations and protective measures to ensure driving safety.

[0077] (2) Cloud Security Brain System

[0078] Please refer to Figure 1 and Figure 3 ,The cloud security brain system includes: digital twin model, vehicle safety monitoring module, safety situation awareness and defect determination module;

[0079] The digital twin model is used to build physical models of various types of autonomous driving vehicles, so as to construct digital twin virtual scenes using the collected surrounding environment, vehicle status and driving behavior data, so as to simulate and analyze various types of autonomous driving vehicles;

[0080] The vehicle safety monitoring module is used to conduct single-vehicle safety analysis on autonomous vehicles based on the digital twin model, fault information, and status performance data of each vehicle subsystem, and to conduct multi-vehicle safety analysis on the same model of autonomous vehicles based on fault information, autonomous driving performance data, and driver operation behavior data of each autonomous driving vehicle of the same model;

[0081] The safety situation awareness and defect assessment module is used to assess and grade the operational safety risks of each autonomous driving vehicle based on the single-vehicle safety analysis results and multi-vehicle safety analysis results.

[0082] It can be understood that the cloud-based security brain system can simultaneously connect to the intelligent connected vehicle original vehicle systems of several autonomous driving vehicles, and the models of autonomous driving vehicles can be multiple, with multiple vehicles for each model.

[0083] In the embodiment of the present invention, if only the data of a single vehicle is analyzed, the purpose of comprehensive safety protection cannot be achieved, and it is also necessary to use big data to analyze multiple vehicles of the same model. Therefore, this solution conducts single-vehicle safety analysis on autonomous driving vehicles, and conducts multi-vehicle safety analysis on the fault information, autonomous driving performance data and driver's operation behavior data of a large number of autonomous driving vehicles, comprehensively monitors the operating status of intelligent networked vehicles, identifies risks and determines defects, thereby improving the safety level of autonomous driving vehicles.

[0084] In some embodiments, reference Figure 1 and Figure 3 ,The cloud security monitoring system also includes an application service module; the application service module includes a defect recall module, a sandbox supervision module and a quality and safety management module;

[0085] When the risk level assessed by the safety situation awareness and defect determination module exceeds the first threshold and is less than the second threshold, the defect recall module is used to trigger the defect recall process and notify the vehicle manufacturer and the owner to carry out defect repairs;

[0086] When the risk level assessed by the safety situation awareness and defect determination module exceeds a second threshold, the quality and safety management module is used to issue an alarm to notify the vehicle manufacturer and the vehicle owner to conduct quality and safety management;

[0087] The sandbox supervision module is used to create a sandbox environment to simulate the running status of new functions.

[0088] In this embodiment, the risk level in the safety situation awareness and defect determination module is evaluated in real time. When the risk level exceeds the first threshold and is less than the second threshold, it indicates that the autonomous driving vehicle has defects and needs to be repaired, and the defect recall module is activated. When the risk level exceeds the second threshold, it indicates that the safety problem of the autonomous driving vehicle is more serious. For example, this model of vehicle has had an accident, which means that the autonomous driving function of this model of autonomous driving vehicle may have safety hazards and urgent quality and safety management is required.

[0089] In this embodiment, the quality and safety management module regularly checks the compliance of the vehicle's software and hardware components by strictly following safety standards and continuous monitoring to ensure that the vehicle system always meets relevant safety specifications. Once the system finds any non-compliance with the standards, the application service will immediately issue an alarm and take necessary corrective measures. At the same time, the service enhances the owner's understanding of the vehicle's safety features and their correct operation methods by regularly updating safety information and providing safety guidance. This not only helps to keep the vehicle in good condition, but also effectively improves the overall safety level.

[0090] In addition, the sandbox supervision module creates a "sandbox" environment in which new functions and updates are simulated and tested. It monitors the operating status of new functions without negatively affecting the normal operation of the vehicle, so as to improve the security and reliability of new vehicle software and updates before they are officially deployed.

[0091] In some embodiments, reference Figure 1 and Figure 3 ,The cloud security monitoring system also includes a security knowledge base;

[0092] The safety knowledge base is used to extract safety features from the collected fault data, defect data and accident data, and to clean, classify and grade the extracted key features to establish a typical fault library, a typical defect library and a typical accident library.

[0093] In this embodiment, the safety knowledge base extracts key safety features from the fault data, defect data and accident data collected from the vehicle-side safety brain system, the vehicle safety monitoring module and the original vehicle system, such as the time, location and type of the fault, defect and accident, and cleans, classifies and grades the data based on its key features, completes the import and conversion of data, and establishes a typical fault library, a typical defect library and a typical accident library in the safety knowledge base. At the same time, the effective storage and management of safety knowledge data is designed and implemented, and the safety knowledge base is tested, maintained and optimized regularly. The typical fault library systematically collects and stores various common fault information that may occur in vehicle operation, records the characteristics, causes, manifestations, detection methods and repair measures of the faults in detail, and provides an information library for rapid diagnosis and fault repair, which can promote manufacturers to prevent these problems in future designs. The typical defect library records and analyzes typical defect information that occurs in the design, manufacturing or material use process of the vehicle, including defect description, impact range, detection methods and corrective measures, and serves the setting of industry safety quality standards. The typical accident database stores vehicle accident cases, including the cause of the accident, specific circumstances, vehicle models and behaviors involved, accident consequences and post-accident analysis. It analyzes the commonalities and patterns of accidents and can provide valuable data resources for government regulators and automobile manufacturers to enhance the targetedness of safety strategies and measures.

[0094] In some embodiments, reference Figure 1 and Figure 3 ,The cloud security monitoring system also includes a model optimization module; the model optimization module includes a vehicle-side security brain model optimization module and a cloud-side security brain model optimization module;

[0095] The vehicle-side safety brain model optimization module is connected to the vehicle-side safety brain system and the safety knowledge base, and is used to perform intensive testing on the models of various modules in the vehicle-side safety system based on the data collected from the vehicle-side safety brain system and the safety knowledge base, so as to continuously update and optimize the model parameters;

[0096] The cloud-based safety brain model optimization module is connected to the vehicle safety monitoring module, digital twin model and safety knowledge base. It is used to conduct intensive testing on each model in the vehicle safety monitoring module based on the data collected from the vehicle safety monitoring module and the safety knowledge base, so as to continuously update and optimize the model parameters.

[0097] In this embodiment, the model optimization module collects data from the vehicle-side safety brain system, the vehicle safety monitoring module and the safety knowledge base, intensively tests the models in the vehicle-side safety brain system and the vehicle safety monitoring module to identify performance limitations, adjusts the internal parameter weights and settings of the model through parameter updates and optimizations, uses data enhancement to expand and improve the training data set of the artificial intelligence model, and adopts model fusion to combine the advantages of multiple models to continuously improve the robustness and generalization ability of the model and achieve higher prediction accuracy and efficiency.

[0098] In some embodiments, reference Figure 3 ,The modules used for single vehicle safety analysis in the vehicle safety monitoring module include: vehicle cumulative fault analysis module, vehicle functional safety early warning module and vehicle quality status physical examination module;

[0099] The vehicle cumulative fault analysis module is used to store the fault information of the autonomous driving vehicle in each time period, and to perform statistical analysis on the stored and generated cumulative fault data to determine the high-frequency faults of the autonomous driving vehicle;

[0100] The vehicle functional safety warning module is used to determine the correlation between each fault and the vehicle functional safety based on the accumulated fault data and digital twin model of the autonomous vehicle, so as to warn of the vehicle functional safety issues that are likely to be caused by the current fault when real-time fault data is obtained;

[0101] The vehicle quality status inspection module is used to predict the component quality status of the autonomous driving vehicle based on the artificial intelligence technology of the recurrent neural network model and the mileage, loss data, and status performance data of each subsystem of the autonomous driving vehicle.

[0102] In this embodiment, the vehicle cumulative fault analysis module collects and stores fault information within the historical time period, including the time of fault occurrence, fault duration, fault location, fault cause, etc., and performs statistical analysis on the stored and generated cumulative fault data to identify high-frequency faults of each autonomous driving vehicle. The vehicle functional safety early warning module determines the correlation between each fault and the functional safety of the vehicle based on the cumulative fault data and digital twin model of the autonomous driving vehicle. When obtaining real-time fault data, it determines whether the current fault has a strong correlation with the functional safety of the vehicle, that is, the probability that the fault only affects the vehicle performance or may cause functional safety problems. If it involves a fault that may cause functional safety problems, corresponding early warning measures need to be taken to improve the safety and reliability of the vehicle. The vehicle quality status physical examination module comprehensively collects the driving time, mileage, tire wear, tire pressure, and the status and performance data of the vehicle engine system, transmission system, electronic system and brake system within the historical period of the vehicle, and uses artificial intelligence technology based on the recurrent neural network (RNN) network to predict the status of the collected sequence data, infer the quality problems that may be caused by the aging of its components over time, and score the status of each system.

[0103] In some embodiments, reference Figure 3 ,The modules used for multi-vehicle safety analysis in the vehicle safety monitoring module include: batch defect identification module, autonomous driving performance evaluation module and driving behavior analysis module;

[0104] The batch defect identification module is used to identify vehicle batch defects based on the fault information of each autonomous driving vehicle of the same model and the production time of each autonomous driving vehicle of the same model;

[0105] The autonomous driving performance evaluation module is used to evaluate the autonomous driving performance of the same model of vehicles based on the autonomous driving performance data and scoring model of each autonomous driving vehicle;

[0106] The driving behavior analysis module is used to judge the abnormal status and unsafe behavior of drivers based on the operating behavior data of each driving group of each autonomous driving vehicle of the same model.

[0107] In this embodiment, in addition to the functional safety issues caused by system failures, the expected functional safety issues of autonomous driving vehicles are also very important. Therefore, in order to provide comprehensive safety protection for both functional safety issues and expected functional safety issues, a batch defect recognition module, an autonomous driving performance evaluation module, and a driving behavior analysis module are designed to provide comprehensive safety protection for both functional safety issues and expected functional safety issues.

[0108] The batch defect recognition module uses deep learning-based artificial intelligence technology to deeply mine batch fault vehicle data, including fault type, fault time, fault cause, fault frequency and environmental factors, identify vehicle batch defects, determine the batch characteristics of defects, and thus prevent possible risks in advance, for example, a certain model and a certain batch have emission problems. The autonomous driving performance evaluation module uses a scoring model to evaluate the autonomous driving performance of different vehicle groups by considering the autonomous driving performance of the same model of autonomous driving vehicles in different driving scenarios, such as perception, decision planning and control, to ensure its reliability and safety during use. The driving behavior analysis module focuses on the driving behavior of the group. It analyzes multi-dimensional driving behavior data through artificial intelligence technology based on unsupervised learning algorithms, including the driver's own attention state, the driver's hand and foot movements during driving, and the multi-faceted interaction between the driver and the vehicle. It analyzes the driving behavior paradigm of driving groups with different characteristics, such as different age groups, different driving experience or different driving styles, to better understand the driving behavior of different driving groups, and timely discover the abnormal state or unsafe behavior of the driver, and reduce traffic accidents caused by human factors.

[0109] In some embodiments, when the autonomous driving performance evaluation module evaluates the autonomous driving performance of the same model of vehicles based on the autonomous driving performance data and scoring model of each autonomous driving vehicle, it is specifically used to:

[0110] Step S1, for each research period, is performed: obtaining the autonomous driving performance data within the research period, and dividing the autonomous driving performance data within the research period into several research cycles; wherein the autonomous driving performance data includes the lateral and longitudinal acceleration of the autonomous driving, the driver's takeover time, and the number of dangerous events that occur, and the dangerous event refers to a failure of the autonomous driving system or a collision of the vehicle.

[0111] In this embodiment, the research period can be set to 3 days, 5 days, one week, or after the vehicle uses the automatic driving function for the first time, etc., so as to obtain the automatic driving performance data of each vehicle of the same model during this research period. It can be understood that the automatic driving performance data can only be collected when the driver uses the automatic driving function.

[0112] Step S2, determining the performance score of each study period based on the lateral and longitudinal accelerations and the driver takeover time in each study period.

[0113] In some embodiments, the performance score for each study cycle is assessed as follows:

[0114]

[0115] In the formula, s iis the performance score of the i-th study cycle, t is the time the driver takes over during the study cycle, T is the duration of the study cycle, and m is the number of times the lateral and longitudinal accelerations exceed the discrimination threshold.

[0116] In this embodiment, the judgment criteria of the lateral and longitudinal accelerations are shown in the following table:

[0117]

[0118] The lateral and longitudinal stability of an autonomous vehicle during driving can be evaluated in the above manner.

[0119] Step S3, based on the performance score of each research period and the number of dangerous events that occurred during the research period, calculate the comprehensive score of the autonomous driving performance during the research period.

[0120] In some embodiments, the comprehensive score of the autonomous driving performance during the study period is evaluated in the following manner:

[0121]

[0122] Where S is the comprehensive score of autonomous driving performance during the research period, s i is the performance score of the ith study period, n is the number of study periods in the study period, and p is the number of hazardous events that occurred in the study period.

[0123] It can be seen that the higher S is, the better the autonomous driving performance is. When S is less than 0, it means that the autonomous driving vehicle does not meet normal driving requirements and a single vehicle defect recall or quality and safety management is required.

[0124] It can be understood that through the above-mentioned comprehensive scoring formula for autonomous driving performance, the number of times dangerous events occur is discussed by situation, which greatly improves the impact of dangerous events on the scoring results, so that the occurrence of dangerous events has a veto function. As long as S is less than 0, the risk level evaluated by the safety situation awareness and defect judgment module can initiate defect recall or quality and safety management services.

[0125] Therefore, in the intelligent connected vehicle safety brain system based on vehicle-cloud interaction provided by the embodiment of the present invention, the vehicle-side safety brain system is responsible for real-time processing and emergency response, while the cloud-side safety brain system is responsible for big data analysis and safety monitoring. The two complement each other and can provide comprehensive safety protection for autonomous driving vehicles.

[0126] like Figure 4 As shown, an embodiment of the present invention further provides a method for safety monitoring of an intelligent networked vehicle based on a system of any embodiment in the specification, the method comprising:

[0127] Step 400: Use the vehicle-side safety brain system to obtain vehicle safety information from the Internet of Vehicles system in real time to identify expected functional safety risks and functional safety risks, and determine whether to perform emergency measures based on the identification results. If so, send the generated emergency measures information to the vehicle execution system to perform emergency control of the autonomous driving vehicle;

[0128] Step 402, using the cloud-based safety brain system to obtain the fault information, autonomous driving performance data and driver's operating behavior data of the autonomous driving vehicle from the Internet of Vehicles system and the vehicle-side safety brain system, so as to conduct single-vehicle safety analysis on the autonomous driving vehicle, as well as multi-vehicle safety analysis on the same model of autonomous driving vehicles, and further conduct comprehensive safety monitoring of the autonomous driving vehicles.

[0129] The content of the above method is based on the same concept as the embodiment of the system of the present invention. For the specific content, please refer to the description in the embodiment of the system of the present invention, and it will not be repeated here.

[0130] It should be noted that, in this article, relational terms such as first and second 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.

[0131] A person of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiments; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk, etc., various media that can store program codes.

[0132] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A smart connected vehicle safety brain system based on vehicle-cloud interaction, characterized in that: include: A vehicle-side safety brain system arranged on the vehicle side and a cloud-side safety brain system arranged on the cloud side, wherein the vehicle-side safety brain system and the cloud-side safety brain system are respectively connected to the original vehicle system of the intelligent network-connected vehicle, and the vehicle-side safety brain system and the cloud-side safety brain system are connected to each other; the original vehicle system of the intelligent network-connected vehicle includes a vehicle networking system, a vehicle automatic driving system and a vehicle execution system; The vehicle-side safety brain system is used to obtain vehicle safety information from the vehicle networking system in real time to identify expected functional safety risks and functional safety risks, and determine whether to perform emergency disposal according to the identification results. If so, the generated emergency disposal information is sent to the vehicle execution system to perform emergency control of the autonomous driving vehicle; The cloud-based safety brain system is used to obtain fault information, autonomous driving performance data, and driver's operating behavior data of the autonomous driving vehicle from the Internet of Vehicles system and the vehicle-side safety brain system, so as to perform single-vehicle safety analysis on the autonomous driving vehicle, as well as multi-vehicle safety analysis on the same model of autonomous driving vehicles, and further conduct comprehensive safety monitoring of the autonomous driving vehicles.

2. The system according to claim 1, characterized in that The vehicle-side safety brain system includes: a risk identification module, a safety and defect analysis module and an emergency response module; The risk identification module is connected to the original vehicle system of the automobile and is used to obtain vehicle safety information from the original vehicle system of the automobile to identify expected functional safety risks and functional safety risks; the vehicle safety information includes the vehicle state vector and the state vectors of surrounding traffic participants for scenario risk assessment, operation data for monitoring the driver's human-computer interaction behavior, and the status signals of sensors, communications and controllers of the vehicle's automatic driving system; The safety and defect analysis module is used to perform fault analysis and source tracing according to the real-time identification results of the risk identification module, and to calculate and classify the comprehensive risk value, so as to determine whether to perform emergency disposal according to the classification results, and if so, formulate an emergency disposal strategy based on the vehicle safety information and the classification results; The emergency response module is used to generate emergency response information according to the emergency response strategy formulated by the safety and defect analysis module and send it to the vehicle execution system to perform emergency control of the autonomous driving vehicle.

3. The system according to claim 1, characterized in that The cloud-based safety brain system includes: a digital twin model, a vehicle safety monitoring module, and a safety situation awareness and defect determination module; The digital twin model is used to build a physical model of each type of autonomous driving vehicle, so as to construct a digital twin virtual scene using the collected surrounding environment, vehicle status and driving behavior data, so as to simulate and analyze each type of autonomous driving vehicle; The vehicle safety monitoring module is used to perform single-vehicle safety analysis on the autonomous driving vehicle based on the digital twin model, the fault information and the status performance data of each subsystem of the vehicle, and to perform multi-vehicle safety analysis on the same model of autonomous driving vehicles based on the fault information of each autonomous driving vehicle of the same model, the autonomous driving performance data and the operation behavior data of the driver; The safety situation awareness and defect determination module is used to evaluate and grade the operational safety risks of each autonomous driving vehicle based on the single-vehicle safety analysis results and the multi-vehicle safety analysis results.

4. The system according to claim 1, characterized in that The vehicle-side safety brain system also includes a data encryption module, which is used to encrypt the data required by the cloud-side safety brain system and transmit it to the cloud-side safety brain system; The cloud-side security brain system also includes a data parsing module for parsing the encrypted data transmitted by the vehicle-side security brain system.

5. The system according to claim 2, characterized in that The risk identification module includes a vehicle driving risk identification module, a human-computer interaction monitoring module and a vehicle functional safety risk identification module; The vehicle driving risk identification module is used to evaluate the driving risk value in the current traffic scenario based on the risk model, the vehicle state vector and the state vectors of surrounding traffic participants; The human-machine interaction monitoring module is used to monitor the operation data of the driver's human-machine interaction behavior to identify the driver's misuse or improper operation behavior and determine the personnel misuse factor; The vehicle functional safety risk identification module is used to detect the status signals of the sensors, communications and controllers of the vehicle's automatic driving system, and determine the fault results and functional safety risk values.

6. The system according to claim 5, characterized in that The driving risk value is evaluated in the following manner: Determine the risk indicator value ITTC based on the collision time: Where D x is the lateral distance between the vehicle and other vehicles, D y is the longitudinal distance between the vehicle and other vehicles, L w is the vehicle width, L l is the length of the vehicle, V rx V is the relative lateral speed between the vehicle and other vehicles, ry A is the relative longitudinal speed between the vehicle and other vehicles, rx A is the relative lateral acceleration between the vehicle and the other vehicle, ry is the longitudinal relative acceleration between the vehicle and other vehicles, a is the influence coefficient of the relative acceleration between the vehicle and other vehicles; The ratio of the vehicle's speed to the road speed limit is used as the correlation coefficient λ for calculating the driving risk value. The calculation formula is: In the formula, v is the speed of the vehicle, v l Set speed limits for roads; Then the driving risk value r under the current traffic scenario is: r = exp(-λ·ITTC).

7. The system according to claim 3, characterized in that The modules for multi-vehicle safety analysis in the vehicle safety monitoring module include: a batch defect recognition module, an automatic driving performance evaluation module and a driving behavior analysis module; The batch defect identification module is used to identify vehicle batch defects based on fault information of each autonomous driving vehicle of the same model and the production time of each autonomous driving vehicle of the same model; The autonomous driving performance evaluation module is used to evaluate the autonomous driving performance of the same model of vehicles based on the autonomous driving performance data and scoring model of each autonomous driving vehicle of the same model; The driving behavior analysis module is used to determine the abnormal state and unsafe behavior of the driver based on the operating behavior data of each driving group of each autonomous driving vehicle of the same model.

8. The system according to claim 7, characterized in that When the autonomous driving performance evaluation module evaluates the autonomous driving performance of the same type of vehicle based on the autonomous driving performance data and scoring model of each autonomous driving vehicle, the autonomous driving performance evaluation module is specifically used to: For each research period, the following are performed: Obtaining the autonomous driving performance data within the research period, and dividing the autonomous driving performance data within the research period into a number of research cycles; wherein the autonomous driving performance data includes the lateral and longitudinal acceleration of the autonomous driving, the driver's takeover time, and the number of dangerous events, wherein the dangerous event refers to a failure of the autonomous driving system or a collision of the vehicle; Determine the performance score for each study period based on the lateral and longitudinal accelerations and the driver takeover time during each study period; Based on the performance score of each study period and the number of dangerous events that occurred during the study period, the comprehensive score of autonomous driving performance during the study period is calculated.

9. The system according to claim 8, characterized in that The overall score of autonomous driving performance during the study period was evaluated as follows: Where S is the comprehensive score of autonomous driving performance during the research period, s i is the performance score of the ith study period, n is the number of study periods in the study period, and p is the number of hazardous events that occurred in the study period.

10. A method for safety monitoring of an intelligent connected vehicle based on the system according to any one of claims 1 to 9, characterized in that: include: The vehicle-side safety brain system is used to obtain vehicle safety information from the Internet of Vehicles system in real time to identify expected functional safety risks and functional safety risks, and determine whether to perform emergency measures based on the identification results. If so, the generated emergency measures information is sent to the vehicle execution system to perform emergency control of the autonomous driving vehicle; The cloud-based safety brain system is used to obtain the fault information, autonomous driving performance data and driver's operating behavior data of the autonomous driving vehicle from the Internet of Vehicles system and the vehicle-side safety brain system, so as to conduct single-vehicle safety analysis on the autonomous driving vehicle and multi-vehicle safety analysis on the same model of autonomous driving vehicles, and further conduct comprehensive safety monitoring of the autonomous driving vehicles.

Citation Information

Patent Citations

  • Safety control method and system based on man-machine co-driving

    CN115092162A

  • Automatic driving vehicle driving safety degree quantification system

    CN115782905A

  • Vehicle supervision method and device

    CN116846937A

  • Fusion safety protection system and control method for intelligent networked automobile

    CN117644880A

  • Risk assessment method and device for traffic safety common sense violation of automatic driving vehicle

    CN117818657A