Search

CN-122024350-A - Automatic driving record data extraction and analysis system and method

CN122024350ACN 122024350 ACN122024350 ACN 122024350ACN-122024350-A

Abstract

The invention provides an automatic driving record data extraction and analysis system and method, and relates to the technical field of automatic driving. The system comprises five modules, namely an OBD interface, safety authentication, data transmission and data verification and exception handling, wherein the OBD interface is intelligently dynamically adapted DOIP to an OBD-II protocol and automatically switches the protocol according to a DSSAD communication architecture, the safety authentication module introduces DSSAD hardware unique identifiers, generates dynamic challenge values through a national cipher SM2 algorithm to realize multi-dimensional protection of the hardware identifiers, passwords and algorithms, the data transmission module dynamically allocates bandwidths according to real-time control classes, sensor original classes and log interaction classes to support incremental data transmission after network interruption, the data verification module utilizes a CRC32 algorithm to carry out multi-layer verification on hash values of the national cipher SM3 and anchor with NTP time sources, and the exception handling module carries out hierarchical processing on warning, serious and fatal exceptions. The system meets the real-time requirement of automatic driving data extraction, ensures the safety and the integrity of data, and supports the production line test and batch verification of an automobile factory.

Inventors

  • LIU HAIZHU

Assignees

  • 刘海柱

Dates

Publication Date
20260512
Application Date
20260115

Claims (10)

  1. 1. The automatic driving record data extraction and analysis system is characterized by comprising an OBD interface module, a safety authentication module, a data transmission module, a data verification module and an abnormality processing module; The OBD interface module is internally provided with a protocol detection algorithm and is used for automatically identifying the communication architecture of the automatic driving take-over system and dynamically switching the adapted communication protocol; The security authentication module introduces DSSAD a hardware unique identifier as a second authentication factor; the data checking module adopts a CRC32 algorithm checking, national cipher SM3 hash comparison and NTP time anchored multi-layer checking mechanism; The exception handling module is used for dividing the exception data into warning exceptions, serious exceptions and deadly exceptions.
  2. 2. The system of claim 1, wherein the OBD interface module transmits laser radar point cloud and camera image large bandwidth data using DOIP protocol when identifying an Ethernet architecture; When the vehicle speed, steering angle and position basic state data are transmitted by adopting an OBD-II protocol when the vehicle speed, steering angle and position basic state data are identified as a CAN bus architecture; the self-adaptive level conversion circuit is also arranged in the automobile factory production line pole end test environment, is used for matching the electrical characteristics of DSSAD in the voltage range of 9V-16V, has high and low temperature protection capability of-40 ℃ to 85 ℃.
  3. 3. The system of claim 1, wherein the security authentication module obtains DSSAD hardware ID first, then generates a dynamic challenge value through SM2 cryptographic algorithm, and combines with the seed and the secret key to complete multidimensional authentication; in the authentication process, the operation mode and the key length of the cryptographic SM4 encryption algorithm are dynamically negotiated with DSSAD, and different temporary encryption keys are generated for each authentication session.
  4. 4. The system of claim 1, wherein the real-time control class data transmitted by the data transmission module comprises a torque request and a deceleration request command for the autopilot system; The sensor original class data transmitted by the data transmission module comprises a laser radar point cloud and a camera image, and the log interaction class data comprises a task instruction and an execution log issued by the test platform; When the network is congested, the module preferentially guarantees the transmission bandwidth of the real-time control data, avoids the delay of key control instructions, supports the increment synchronization after the network is interrupted, records the transmitted data check points, and only transmits unfinished increment data segments after the connection is restored.
  5. 5. The system of claim 1, wherein the data verification module performs integrity verification on the original data by using a CRC32 algorithm, generates SM3 hash values for key parameters of vehicle speed, steering angle and position, compares the SM3 hash values with DSSAD ends, and anchors a data timestamp with a test platform NTP time source; The deviation between the data time stamp and the time of the test platform is controlled within 10ms through NTP time anchoring, so that the problem of data time sequence disorder caused by transmission delay is accurately identified, and if the SM3 Ha Xibi is inconsistent or the time deviation exceeds a threshold value, the data is judged to be abnormal, and a re-extraction request is triggered.
  6. 6. The system of claim 1, wherein the exception handling module performs gradient processing on key error class exceptions: The 1 st-2 nd time of error is judged as warning abnormality, and only logs are recorded; The 3 rd error is judged to be serious abnormality, the secret key is automatically regenerated, and authentication is initiated; and 5 times of errors are accumulated to judge that the error is deadly abnormal, the communication between the system and the corresponding DSSAD is locked, abnormal data are prevented from being recorded, and the manager is required to restore after unlocking.
  7. 7. The automatic driving record data extraction and analysis method is characterized by comprising the following steps of: s1, a system sends an expansion session request instruction, receives DSSAD a response and then enters a diagnosis expansion session, sends a security access seed request instruction, receives DSSAD fed back seed, sends a key instruction, completes security authentication after receiving the response, and acquires data extraction authority; s2, reading DSSAD stored data elements including vehicle states, sensor data and control instructions by the system through interaction instructions; S3, carrying out integrity check on the extracted original data by adopting a CRC32 algorithm, and re-executing the step S2 if the check fails, and after the check passes, analyzing the original data into a standardized format comprising a time stamp, a data type identifier, a device identifier, a parameter name and a parameter value to generate a CSV data file and an interaction log; And S4, uploading the CSV data file, the interaction log and the test result to a test management platform through a WEBSOCKET interface.
  8. 8. The method of claim 7, wherein in step S1, the system further uses SM2 cryptographic algorithm to digitally sign the seed and DSSAD hardware ID fed back by DSSAD before sending the key, generates a dynamic challenge value, and sends the challenge value and the key together to the DSSAD through the instruction, the DSSAD re-verifies the key by requiring prior verification of the signature validity of the challenge value, and the DSSAD returns a positive response after the double verification is passed.
  9. 9. The method of claim 7, wherein in step S3, if the verification is passed, the data is parsed to generate a CSV format data file, an interaction log is generated, and the instruction response and the file length are recorded.
  10. 10. The method of claim 7, wherein in step S4, the JWT token sent by the system is decoded before the test management platform receives the data, and the validity period of the token signature are verified; and meanwhile, whether the deviceCode, MAC address and the IP address of the system are consistent with preset information or not is compared, and uploading data is received after the double verification is passed, so that legal traceability of data sources is ensured.

Description

Automatic driving record data extraction and analysis system and method Technical Field The invention relates to the technical field of automatic driving, in particular to an automatic driving record data extraction and analysis system and method. Background The automatic driving technology is an important development trend in the automobile industry in recent years, and an automatic driving data recording system is a system capable of monitoring, collecting and recording various key data in the running process of an automatic driving vehicle, wherein the key data may include information of sensor readings, vehicle states, control inputs and surrounding environments. With the rapid development of automatic driving technology, the acquisition and analysis of vehicle behavior data becomes critical, and vehicles are equipped with various sensors including lidar, cameras, radar, etc. to sense the surrounding environment. The massive data generated in the sensors are key to understanding the performance of the autopilot system, optimizing algorithms, ensuring safety and performing accident analysis. Thus, automated driving data logging system extraction tools have evolved to efficiently process and analyze these massive data. The sensor data comprises various types such as images, point clouds, radar signals and the like, the data volume is huge, the formats are not uniform, the existing tools are difficult to extract and analyze efficiently, the real-time performance of the data is required by an automatic driving system, the response delay of the existing extraction method is high, the real-time performance of the data is required by the automatic driving system, the response delay of the existing extraction method is high, and meanwhile, a uniform safety authentication mechanism is lacking in the data transmission and storage process, so that the risk of data leakage or tampering is easy to cause. China is pushing DSSAD (Data Storage System for Automated Driving) to mandatory national standard formulation, but at present, an automobile factory production line testing method and a systematic tool aiming at DSSAD data are lacking in the industry, and data formats and interface standards are not uniform, so that data of different manufacturers are difficult to share, and therefore, an integrated driving record data extraction and analysis tool is needed. Aiming at the defects in the prior art, the invention provides an automatic driving record data extraction and analysis system and method, wherein physical connection is established between an OBD port and DSSAD, and data interaction is realized based on a vehicle-mounted Ethernet diagnosis (Diagnostic Communication over Internet Protocol, DOIP) protocol. The method realizes high-efficiency extraction and standardized analysis of mass multi-type data of DSSAD systems, meets the real-time requirement of an automatic driving system on data extraction, ensures the safety and the integrity of data transmission and storage, provides an automobile factory production line test method by adapting DSSAD mandatory national standard requirements, supports test task issuing, result uploading and batch verification, establishes a cooperative interaction mechanism of tools and a test platform, realizes data synchronization, instruction response and test flow automation, meets the data application requirement, protects user privacy, and meets the requirements of legal rules on data acquisition and use. Disclosure of Invention The invention aims to solve the technical problem of providing an automatic driving record data extraction and analysis system and method, which solve the problem of insufficient extraction of sensor data in the prior art and improve the real-time performance and safety of data extraction in the automatic driving process. The technical scheme adopted by the invention is to provide an automatic driving record data extraction and analysis system and method, wherein the system comprises an OBD interface module, a security authentication module, a data transmission module, a data verification module and an abnormality processing module: The OBD interface module is internally provided with a protocol detection algorithm and is used for automatically identifying the communication architecture of the automatic driving take-over system and dynamically switching the adapted communication protocol; The security authentication module introduces DSSAD a hardware unique identifier as a second authentication factor; the data checking module adopts a CRC32 algorithm checking, national cipher SM3 hash comparison and NTP time anchored multi-layer checking mechanism; The exception handling module is used for dividing the exception data into warning exceptions, serious exceptions and deadly exceptions. Preferably, when the OBD interface module is identified as an Ethernet architecture, DOIP protocols are adopted to transmit laser radar point cloud and 4K camera image large