Vehicle data processing method and device, electronic equipment and storage medium

Abstract

The application relates to a vehicle data processing method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring vehicle data, wherein the vehicle data comprises data of multiple data types, the data types comprise vehicle dynamics data, and the data types are other than the vehicle dynamics data; performing time sequence signal processing on the vehicle dynamics data, and extracting data features affecting changes in driver subjective feelings from the processed data; identifying an abnormal time sequence interval based on the data features and the vehicle data, wherein the absolute value of the data features is greater than or equal to a first preset threshold, and the vehicle data satisfies an abnormal checking condition; extracting a data segment in the abnormal time sequence interval from the data features and the vehicle data, and outputting the data segment. The embodiment of the application fills the analysis gap between objective vehicle dynamics data and driver subjective feelings, and improves the accuracy and reliability of processing the vehicle data.

Description

Technical Field

[0001] This application relates to the field of vehicle data processing technology, specifically to a vehicle data processing method, apparatus, electronic device, and storage medium. Background Technology

[0002] As the level of vehicle intelligence continues to improve, the industry has placed higher demands on the driving comfort of vehicles.

[0003] In related technologies, during vehicle testing, macroscopic statistical analysis is typically performed on the collected test data over a continuous period. However, this macroscopic test data cannot reveal the driver's subjective feelings under various instantaneous micro-management scenarios, such as jerking sensations or lateral discomfort caused by vehicle yaw. There is an analytical breakpoint between objective vehicle dynamics data and the driver's subjective feelings, leading to a decrease in the reliability of vehicle test results and making it difficult to accurately pinpoint the factors affecting ride comfort based on the test results. Summary of the Invention

[0004] This application provides a vehicle data processing method, apparatus, electronic device, and storage medium to solve the problem in related technologies where there is an analytical breakpoint between objective test data and the driver's subjective feelings, leading to a decrease in the reliability of vehicle test results.

[0005] To achieve the above objectives, the technical solution adopted in this application is as follows: In a first aspect, this application provides a vehicle data processing method, the method comprising: Acquire vehicle data, which includes data of multiple data types, including vehicle dynamics data and other data types besides vehicle dynamics data; the vehicle dynamics data is objective data generated during vehicle operation. The vehicle dynamics data is subjected to time-series signal processing, and data features that affect changes in the driver's subjective feelings are extracted from the processed data. Based on the data features and the vehicle data, identify abnormal time intervals where the absolute value of the data features is greater than or equal to a first preset threshold and the vehicle data meets the abnormality verification conditions. Extract data segments within the abnormal time interval from the data features and the vehicle data, and output the data segments.

[0006] Optionally, the step of identifying an abnormal time series interval where the absolute value of the data feature is greater than or equal to a first preset threshold and the vehicle data meets the anomaly verification condition based on the data feature and the vehicle data includes: The data features are timestamped together with the vehicle data. According to the data filtering rules, the comparison time period is determined. The data filtering rules include at least one of the following: using the received specified time period as the comparison time period, using the time period corresponding to the received specified timestamp point as the comparison time period, and using the time period where the absolute value of the data feature is greater than or equal to a first preset threshold as the comparison time period. Within the comparison time period, the time interval in which the absolute value of the data feature is greater than or equal to a first preset threshold and the vehicle data meets the anomaly verification conditions is identified as the abnormal time interval.

[0007] Optionally, data types other than vehicle dynamics data include video data and driver physiological data; The method further includes: Obtain the rate of change and mean of the physiological data; The time interval within the comparison time period where the absolute value of the data feature is greater than or equal to a first preset threshold and the vehicle data meets the anomaly verification conditions is defined as the abnormal time interval, including: A time series interval that simultaneously meets the following conditions is identified as an abnormal time series interval: the absolute value of the data feature is greater than or equal to a first preset threshold, the rate of change is greater than or equal to a second preset threshold, the mean is greater than or equal to a third preset threshold, and the video data meets preset judgment conditions. The preset judgment conditions include at least one of the following: the relative distance between the vehicle and the lane line is less than or equal to a preset approach threshold, the distance between the vehicle and the obstacle is less than or equal to a preset distance threshold, the driver's facial expression features include a first preset abnormal feature, and the driver's body movement features include a second preset abnormal feature.

[0008] Optionally, the vehicle dynamics data includes at least one data sequence, and each data sequence corresponds to a different dynamics data type; The step of performing time-series signal processing on the vehicle dynamics data and extracting data features that affect changes in the driver's subjective perception from the processed data includes: Upon receiving a data sequence selection instruction, the data sequence indicated by the data sequence selection instruction in the vehicle dynamics data is taken as the target data sequence; The target data sequence is subjected to time-series signal processing, and data features that affect changes in the driver's subjective feelings are extracted from the processed data.

[0009] Optionally, the target data sequence includes at least one of a steering wheel angle data sequence, a lateral acceleration data sequence, and a vehicle distance from lane lines data sequence; The step of performing time-series signal processing on the target data sequence and extracting data features that affect changes in the driver's subjective perception from the processed data includes at least one of the following steps: If the target data sequence includes the steering wheel angle data sequence, then perform a first-order differential operation on the steering wheel angle data sequence to generate a steering wheel angle acceleration sequence, and extract the peak value from the steering wheel angle acceleration sequence to obtain the steering abruptness feature as a data feature; If the target data sequence includes the lateral acceleration data sequence, then the lateral acceleration data sequence is subjected to absolute value processing and low-pass filtering processing in sequence to obtain the lateral acceleration data sequence, and the yaw and tilt sensing features are extracted as data features based on the lateral acceleration data sequence. If the target data sequence includes the vehicle-to-lane-line distance data sequence, then the absolute value operation and dynamic threshold filtering are performed on the vehicle-to-lane-line distance data sequence to extract the high-risk moments of lane crossing and obtain the lane approach pressure perception features as data features. The high-risk moments of lane crossing refer to the moments when the driver feels psychological pressure due to the approach of the lane line.

[0010] Optionally, upon receiving a data sequence selection instruction, selecting the data sequence indicated by the data sequence selection instruction in the vehicle dynamics data as the target data sequence includes: Data within the target time interval is extracted from the vehicle dynamics data and used as slice data; Upon receiving a data sequence selection instruction, the data sequence indicated by the data sequence selection instruction in the sliced ​​data is taken as the target data sequence.

[0011] Optionally, before extracting data within a target time interval from the vehicle dynamics data as slice data, the method further includes: Upon receiving a data slicing instruction, obtain the event timestamp carried by the data slicing instruction; The target time interval is determined based on the event timestamp.

[0012] Optionally, before extracting data within a target time interval from the vehicle dynamics data as slice data, the method further includes: If the vehicle dynamics data satisfies the slicing condition, determine the moment when the vehicle dynamics data satisfies the slicing condition. The target time interval is determined based on the time when the vehicle dynamics data meets the slicing conditions.

[0013] Optionally, the vehicle dynamics data includes lateral acceleration and steering wheel angular rate; the method further includes: If the lateral acceleration is greater than a preset acceleration threshold and the steering wheel angular rate is greater than a preset angular rate threshold, the vehicle dynamics data is determined to meet the slicing conditions.

[0014] Secondly, this application provides a vehicle data processing device, the device comprising: The data acquisition module is used to acquire vehicle data, which includes data of multiple data types, including vehicle dynamics data and other data types; the vehicle dynamics data is objective data generated during vehicle operation. The data processing module is used to perform time-series signal processing on the vehicle dynamics data and extract data features that affect changes in the driver's subjective feelings from the processed data. The output module is used to identify, based on the data features and the vehicle data, an abnormal time series interval where the absolute value of the data features is greater than or equal to a first preset threshold and the vehicle data meets the abnormality verification conditions. The output module is further configured to extract data segments within the abnormal time interval from the data features and the vehicle data, and output the data segments.

[0015] Thirdly, this application provides an electronic device, which includes a memory and a processor. The memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory. When the processor executes the computer program, it implements the vehicle data processing method as described in the first aspect.

[0016] Fourthly, this application provides a readable storage medium storing a computer program that, when executed by a processor, implements the vehicle data processing method as described in the first aspect.

[0017] The beneficial effects of this application are: The vehicle data processing method provided in this application acquires vehicle data of multiple data types, including vehicle dynamics data and other data types. The vehicle dynamics data is objective data generated during vehicle operation, providing a data foundation for multi-dimensional data processing. Subsequently, by performing time-series signal processing on the vehicle dynamics data and extracting data features affecting changes in driver subjective feelings from the processed data, it is possible to extract data features reflecting driver subjective feelings from purely objective vehicle dynamics data, establishing a quantitative correlation between objective vehicle dynamics data and driver subjective feelings. This effectively bridges the analytical gap between objective data and driver subjective feelings in whole-vehicle testing and driving behavior research. Finally, based on the data features and vehicle data, identification... An abnormal time series interval where the absolute value of the data feature is greater than or equal to a first preset threshold and the vehicle data meets the anomaly verification conditions can accurately pinpoint the operating period that causes abnormal subjective feelings in drivers based on the joint verification of multi-dimensional data, narrowing the analysis scope, refining the granularity of time series analysis, and effectively improving the efficiency and accuracy of anomaly localization. Finally, data fragments within the abnormal time series interval are extracted from the data features and vehicle data and output for subsequent analysis, forming a complete analytical closed loop from objective data processing and extraction of data features affecting changes in driver subjective feelings to the localization and output of data fragments within the abnormal time series interval. This improves the accuracy and reliability of vehicle comfort test results determined by processing vehicle data, facilitates accurate localization of factors influencing ride comfort, and provides reliable data support for optimizing vehicle ride performance. Attached Figure Description

[0018] Figure 1 A flowchart illustrating the steps of a vehicle data processing method provided in this application embodiment; Figure 2 A schematic diagram of the logical architecture of a computer device provided in an embodiment of this application; Figure 3 A flowchart illustrating the steps of another vehicle data processing method provided in this application embodiment; Figure 4 A schematic diagram of the logical architecture of another computer device provided in this application embodiment; Figure 5 This is a logic block diagram of a vehicle data processing device provided in an embodiment of this application. Detailed Implementation

[0019] The embodiments of this application will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be understood that the preferred embodiments are only for illustrating this application and are not intended to limit the scope of protection of this application.

[0020] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. Therefore, the drawings only show the components related to this application and are not drawn according to the number, shape and size of the components in actual implementation. In actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0021] Method Implementation Examples In the process of vehicle testing, related technologies typically rely on collected test data to perform macroscopic statistical analysis of continuous test data. However, this approach has several shortcomings: 1) It cannot perform real-time analysis and display of results for multi-dimensional test data; 2) It lacks the function of locating data based on specified data change points; 3) It lacks support for batch processing of data of multiple data types using custom methods. Particularly regarding point 3), these technologies have significant limitations when applied to vehicle driving behavior research and subjective evaluation scenarios for intelligent driving. When the testing system receives full vehicle data (such as vehicle speed, longitudinal acceleration, lateral acceleration, steering wheel angle, steering wheel angular velocity, and distance from the lane line), it can only provide macroscopic statistics and display of continuous test data. This macroscopic test data fails to reveal the driver's subjective feelings under various instantaneous micro-management scenarios, such as jerking sensations and lateral discomfort caused by vehicle yaw, resulting in an analytical gap between objective vehicle dynamics data and the driver's subjective feelings. The lack of a batch processing mechanism capable of fine-grained, custom logical segmentation and mathematical transformation of multi-dimensional vehicle data based on subjective evaluation needs makes it difficult for users to efficiently locate data fragments corresponding to key driving events that trigger physiological discomfort or special experiences from massive amounts of vehicle data. Here, "users" refers to researchers other than drivers.

[0022] Therefore, there is a need to provide a means of processing test data that can achieve synchronous collection of multi-dimensional vehicle test data, and has the functions of real-time analysis, visualization, batch data processing with customizable method configuration, and location based on specified data change points.

[0023] To address the aforementioned problems, this application provides a vehicle data processing method, referring to... Figure 1The vehicle data processing method may specifically include the following steps: Step S101: Obtain vehicle data, which includes data of multiple data types, including vehicle dynamics data and other data types besides vehicle dynamics data; the vehicle dynamics data is objective data generated during vehicle operation.

[0024] Step S102: Perform time-series signal processing on the vehicle dynamics data, and extract data features that affect the driver's subjective feelings from the processed data.

[0025] Step S103: Based on the data features and the vehicle data, identify the abnormal time intervals where the absolute value of the data features is greater than or equal to a first preset threshold and the vehicle data meets the abnormality verification conditions.

[0026] Step S104: Extract data segments within the abnormal time interval from the data features and the vehicle data, and output the data segments.

[0027] The vehicle data processing method provided in this application embodiment can be applied to a vehicle data processing system. The vehicle data processing system includes computer equipment, data acquisition equipment, and network connection equipment. The computer equipment and data acquisition equipment establish a communication connection through the network connection equipment.

[0028] The vehicle data consists of multi-source time-series test data collected during vehicle operation in whole-vehicle testing or driving behavior research scenarios.

[0029] The vehicle data includes data of multiple data types. In some embodiments, the data types include vehicle dynamics data, as well as data types other than vehicle dynamics data. In some embodiments, data types other than vehicle dynamics data include video data and driver physiological data.

[0030] The vehicle dynamics data refers to real-time driving dynamics data collected during vehicle operation. The vehicle dynamics data includes at least one data sequence, each corresponding to a different dynamics data type. For example, the dynamics data type may include, but is not limited to, longitudinal acceleration, lateral acceleration, steering wheel angle, steering wheel angle rate, and vehicle distance from lane lines.

[0031] The video data includes in-vehicle and out-of-vehicle videos; the in-vehicle videos include image data of the area where the driver is located in the cabin, and the out-of-vehicle data includes image data of the environment in front of the vehicle and the environment to the sides of the vehicle.

[0032] The driver's physiological data is used to reflect the driver's real-time physiological state; physiological data may include, but is not limited to, electrocardiogram signals, electromyogram signals, and electrodermal signals.

[0033] In some embodiments, the data acquisition device includes a vehicle dynamics data acquisition device, an audio-visual acquisition device, and a physiological data acquisition device.

[0034] Vehicle dynamics data acquisition equipment is used to collect real-time vehicle dynamics data during vehicle operation and transmit the vehicle dynamics data to computer equipment. Vehicle dynamics data acquisition equipment may include, but is not limited to, Controller Area Network (CAN) bus acquisition devices, inertial measurement units, steering wheel angle sensors, lane vision perception modules, etc.

[0035] Audio and video acquisition equipment is used to collect video data of a vehicle during its operation and transmit the video data to a computer device; audio and video acquisition equipment may include, but is not limited to, in-vehicle cabin cameras, front monocular cameras, side cameras, recording microphones, etc.

[0036] Physiological data acquisition devices are used to collect the driver's physiological data during vehicle operation and transmit the physiological data to computer equipment; physiological data acquisition devices may include, but are not limited to, electrocardiogram sensors, electromyography sensors, and skin conductance sensing wristbands.

[0037] Reference Figure 2 The computer equipment includes a control module 11, a data acquisition module 12, a data processing module 13, and an output module 14.

[0038] The control module 11 is used to coordinate the data acquisition module 12, the data processing module 13 and the output module 14, manage the network connection (including automatic reconnection), the on / off status of the data acquisition equipment, the data acquisition process of starting and stopping vehicles and the storage process, and handle abnormal situations such as computer equipment failure.

[0039] The data acquisition module 12 is used to acquire vehicle data through the data acquisition device. The data acquisition module 12 can establish a communication connection with the data acquisition device through a network connection device, receive vehicle data transmitted by the data acquisition device in real time, and support the parallel reception of vehicle data of different data types.

[0040] For example, the data acquisition module 12 can receive physiological data transmitted by the physiological data acquisition device through third-party software. The third-party software is a matching acquisition program for the physiological data acquisition device, used to read the raw sensor signals output by the physiological data acquisition device, complete signal parsing and data encapsulation, and then push out a standardized data stream. The data acquisition module 12 simultaneously acquires video data through multiple cameras in the audio and video acquisition device.

[0041] Data processing module 13 includes a first processing submodule and a second processing submodule. The first processing submodule is used to perform step S102 to extract data features that affect changes in the driver's subjective feelings; the second processing submodule is used to process physiological data in real time.

[0042] Output module 14 includes a visualization submodule and an output submodule. The visualization submodule displays real-time change curves of structured data such as vehicle dynamics and physiological data in the vehicle data through dynamic charts, providing users with an intuitive view of the real-time trends of vehicle data. It also displays status information such as network connection and vehicle data reception, and provides interactive buttons for connection services, baseline acquisition, task control, and data acquisition equipment debugging. The output submodule executes step S103 to identify abnormal time series intervals and executes step S104 to extract data fragments within the abnormal time series intervals from data features and vehicle data, and outputs the data fragments.

[0043] It is understood that in the embodiments of this application, the data acquisition devices are all mounted on the vehicle under test, and are used to complete in-situ, synchronous vehicle data acquisition under vehicle driving conditions. The computer equipment is communicatively connected to the data acquisition devices and is used to execute the vehicle data processing method provided in this application.

[0044] Computer equipment may include, but is not limited to, in-vehicle edge computing terminals, in-vehicle industrial control computers, in-vehicle industrial control computers, portable host computers, and back-end servers. Among them, in-vehicle computer equipment is deployed inside the vehicle under test and can realize real-time data processing at the in-vehicle end; non-in-vehicle equipment such as portable host computers and back-end servers can be deployed outside the vehicle and receive vehicle data uploaded by the in-vehicle end through wired or wireless communication. This application does not make specific limitations in this regard.

[0045] In step S101, the data acquisition module in the computer device can acquire vehicle data through the data acquisition device. The data acquisition module can receive vehicle data transmitted from the data acquisition device in real time, synchronously mark the acquisition timestamp of each set of vehicle data, and form a time-aligned vehicle data sequence by parsing the data type and sorting it by timestamp. The data types include vehicle dynamics data and other data types, enabling the construction of a multi-dimensional vehicle data system. This provides a complete data foundation for subsequent joint verification and analysis of multi-dimensional vehicle data, avoiding the analytical limitations caused by a single vehicle data dimension.

[0046] Among them, vehicle dynamics data refers to objective data generated during vehicle operation, while data types other than vehicle dynamics data can include objective data and / or subjective data.

[0047] In some embodiments, objective data in data types other than vehicle dynamics data include physiological data and video data.

[0048] In some embodiments, subjective data, in addition to vehicle dynamics data, includes subjectively labeled data, which includes a timestamp and labeling information. In this scenario, the data acquisition device also includes an external input device, such as an external keyboard. During vehicle operation, when the driver experiences physiological discomfort or a specific experience, they can label the current driving event using the external input device. The external input device is equipped with different function keys, each corresponding to different labeling information. For example, key "1" corresponds to the labeling information "tension and pressure," key "2" corresponds to the labeling information "lateral bump discomfort," and key "3" corresponds to the labeling information "obvious steering jerking." After detecting the labeling operation, the computer device records the timestamp of the labeling operation as the timestamp and simultaneously acquires the labeling information corresponding to the operation. Based on the timestamp and labeling information, subjectively labeled data is generated and stored.

[0049] In some embodiments, the vehicle data obtained through step S101 includes vehicle dynamics data, physiological data, and subjective label data, which are sorted in time sequence and stored as independent data sequences according to their data types; while the video data is stored independently in audio and video file formats.

[0050] In some embodiments, the computer device can integrate vehicle dynamics data, physiological data, and subjective marker data from vehicle data into a unified data table after aligning them by timestamps, and simultaneously store the file path and timestamp mapping relationship of video data in the data table; the visualization submodule in the computer device can directly call the data table, display the change curves of various vehicle data over time in real time through dynamic charts, and synchronously retrieve the video data of the corresponding time period for interactive display.

[0051] In vehicle data, video data is stored in Waveform Audio File Format (WAV) and Audio Video Interleaved (AVI) formats; vehicle dynamics data, physiological data, and subjectively labeled data are stored in Comma-Separated Values ​​(CSV) format.

[0052] The first processing submodule in the computer device serves as a hub connecting objective vehicle dynamics data with the driver's subjective perception, and is designed to address the current analytical gap between vehicle dynamics data and the driver's subjective perception. In step S102, the first processing submodule performs time-series signal processing on the vehicle dynamics data and extracts data features that influence changes in the driver's subjective perception from the processed data.

[0053] Among them, time-series signal processing refers to at least one of the following operations on vehicle dynamics data: filtering, mathematical transformation, noise reduction, and amplitude normalization. These operations are used to remove noise interference from vehicle dynamics data, normalize the time sequence and amplitude format of the data, and provide standardized basic data for subsequent data feature extraction.

[0054] To address the nonlinear mapping relationship between objective vehicle dynamics data and driver subjective feelings, the first processing submodule provides dedicated processing logic for human factors analysis. Based on standardized time-series data after time-series signal processing, it extracts data features that influence changes in driver subjective feelings. Specifically, the first processing submodule can perform target operations on the time-series signal-processed vehicle dynamics data in batches. These target operations include, but are not limited to, first-order differential operations, absolute value operations, and sliding window statistical operations. Through these target operations, the vehicle dynamics data is transformed into quantitative indicators reflecting the rate of data change, the severity of fluctuations, and the intensity of impact. This allows the extraction of data features directly related to changes in driver subjective feelings such as jerking and lateral discomfort. Furthermore, the first processing submodule supports custom configuration of operation rules. Users can adjust the formulas, parameters, and sliding window duration of the target operations through configuration files or an interactive interface to adapt to the subjective-objective mapping requirements of different vehicle models and testing scenarios.

[0055] In step S103, the output module can identify, based on data features and vehicle data, the time intervals where the absolute value of the data features is greater than or equal to a first preset threshold, and the time intervals where the vehicle data meets the anomaly verification conditions. Specifically, an absolute value of the data features greater than or equal to the first preset threshold indicates that the fluctuation intensity of the vehicle dynamics data has reached a perceptual threshold that may cause subjective discomfort to the driver; the vehicle data meeting the anomaly verification conditions indicates that the corresponding indicators of the vehicle data exhibit fluctuations that meet the anomaly verification conditions. When the absolute value of the data features is greater than or equal to the first preset threshold and the vehicle data meets the anomaly verification conditions, it indicates that the abnormal fluctuations of the data features and the abnormal fluctuations of the multi-source data in the vehicle data have temporal consistency. These two conditions mutually corroborate each other, eliminating misjudgments caused by single data noise or random fluctuations. The corresponding time interval is the time interval where abnormal driving comfort truly exists. The output module can then use the time interval where the absolute value of the data features is greater than or equal to the first preset threshold and the vehicle data meets the anomaly verification conditions as the abnormal time interval.

[0056] The first preset threshold is a critical value used to determine whether the absolute value of the data feature is abnormal, and it can be calibrated based on the test data of driver and passenger human factors engineering.

[0057] The anomaly verification conditions are a set of multi-dimensional verification rules built on vehicle data of multiple data types. They can cover at least one of vehicle dynamics threshold, physiological data fluctuation threshold, and video data recognition and judgment rules, and are used to cross-verify vehicle comfort anomalies.

[0058] An abnormal time series interval is a time series interval containing data features whose absolute value is greater than or equal to a first preset threshold and whose vehicle data meets the abnormality verification conditions.

[0059] It is understandable that if the output module determines an abnormal timing interval in step S103, it indicates that the current vehicle's driving comfort cannot meet the requirements, the vehicle's test result is a test failure, and the output module continues to output data segments through step S104.

[0060] If the output module does not identify an abnormal timing interval in step S103, it indicates that the current vehicle's driving comfort meets the requirements, the vehicle's test result is that the test is passed, and there is no need to continue to step S104.

[0061] In step S104, the output module can extract all data within the abnormal time interval from the data features and vehicle data, forming time-aligned data segments and outputting them for subsequent analysis. During the subsequent analysis, the influencing factors on ride comfort can be analyzed and determined based on the data segments. Based on these influencing factors, the vehicle dynamics simulation model is iteratively adjusted until the vehicle's comfort meets the expected requirements, meaning the vehicle's test result is considered passed.

[0062] In some embodiments, the factors influencing ride comfort are analyzed and determined based on the data segments. Specifically, this involves: synchronously comparing multi-source data in time-aligned data segments within the same time window, analyzing the abrupt change moments of vehicle dynamics data, driving conditions recorded by in-vehicle and out-of-vehicle videos, fluctuation trends of driver physiological data, and time-series matching relationships of data features, identifying vehicle dynamics conditions, driving behavior, and cabin environment characteristics that are strongly correlated with abnormal physiological data indicators through correlation analysis, and determining the aforementioned strongly correlated features as factors influencing ride comfort.

[0063] In summary, the vehicle data processing method provided in this application acquires vehicle data of multiple data types, including vehicle dynamics data and other data types. The vehicle dynamics data is objective data generated during vehicle operation, providing a data foundation for multi-dimensional data processing. Subsequently, by performing time-series signal processing on the vehicle dynamics data and extracting data features affecting changes in the driver's subjective feelings from the processed data, data features reflecting the driver's subjective feelings can be extracted from purely objective vehicle dynamics data. This establishes a quantitative correlation between objective vehicle dynamics data and the driver's subjective feelings, effectively bridging the analytical gap between objective data and the driver's subjective feelings in vehicle testing and driving behavior research. Finally, based on the data features and vehicle data, the method identifies data features whose absolute values ​​are greater than or equal to a first preset threshold and whose values ​​are... The abnormal time series intervals that meet the anomaly verification conditions of vehicle data can accurately pinpoint the operating conditions that cause abnormal subjective feelings in drivers based on the joint verification of multi-dimensional data, narrowing the scope of analysis, refining the granularity of time series analysis, and effectively improving the efficiency and accuracy of anomaly localization. Finally, data fragments within the abnormal time series intervals are extracted from data features and vehicle data and output for subsequent analysis, forming a complete analytical closed loop from objective data processing and extraction of data features affecting changes in driver subjective feelings to the localization and output of data fragments within the abnormal time series intervals. This provides micro-granular reconstruction and closed-loop verification of vehicle dynamics data oriented towards driver subjective feelings, improving the accuracy and reliability of vehicle comfort test results determined by processing vehicle data. It facilitates the precise and efficient localization of factors influencing driving comfort that induce physiological discomfort and poor driving experience in drivers, providing reliable data support for optimizing vehicle driving performance.

[0064] Furthermore, in this embodiment, the computer device can also utilize a visualization submodule to display real-time change curves of vehicle data through dynamic charts, overcoming the shortcomings of related technologies that cannot analyze and synchronously display data results in real time. This facilitates users' intuitive observation of the real-time fluctuation trends of various vehicle data and helps to quickly capture instantaneous abnormal driving conditions. The output submodule supports identifying abnormal time-series intervals based on data features and vehicle data where the absolute value of the data features is greater than or equal to a first preset threshold and the vehicle data meets the abnormality verification conditions. It extracts data segments within the abnormal time-series intervals from the data features and vehicle data and outputs the data segments. This enables targeted backtracking verification of key driving event segments, significantly reducing the cost of manual retrieval and comparison of massive amounts of data.

[0065] In an optional embodiment of this application, step S103, which identifies an abnormal time series interval where the absolute value of the data feature is greater than or equal to a first preset threshold and the vehicle data meets the anomaly verification condition based on the data feature and the vehicle data, includes: Step S1031: Align the data features with the vehicle data using timestamps.

[0066] Step S1032: Determine the comparison time period according to the data filtering rules.

[0067] Step S1033: Within the comparison time period, identify the time interval where the absolute value of the data feature is greater than or equal to a first preset threshold and the vehicle data meets the anomaly verification condition, and use this interval as the abnormal time interval.

[0068] In this embodiment, during the process of identifying abnormal time intervals, the computer device can, after aligning data features with vehicle data using timestamps, determine a comparison time period according to data filtering rules. Within the comparison time period, it identifies time intervals where the absolute value of the data features is greater than or equal to a first preset threshold and the vehicle data meets the anomaly verification conditions, as abnormal time intervals. The data filtering rules include at least one of the following: using a received specified time period as the comparison time period; using the time period corresponding to a received specified timestamp point as the comparison time period; or using the time period where the absolute value of the data features is greater than or equal to the first preset threshold as the comparison time period.

[0069] Therefore, by unifying the temporal reference of vehicle data and data features of various data types through timestamp alignment, the temporal deviation between multi-source heterogeneous data is eliminated, placing data of different data types under the same temporal coordinate system. This improves the accuracy of abnormal time series intervals from the perspective of time reference and avoids distortion of abnormal time series interval identification results due to time series misalignment. Multiple optional data filtering rules flexibly determine the comparison time period, supporting both user-specified time periods and time periods corresponding to timestamp points as comparison time periods, or automatically defining time periods where the absolute value of data features is greater than or equal to a first preset threshold as comparison time periods. This eliminates the need for point-by-point traversal of all vehicle data, effectively reducing the data processing scope and significantly improving the identification efficiency of abnormal time series intervals. Furthermore, the abnormal time series intervals identified within the comparison time period simultaneously possess driver subjective perception characteristics and multi-source vehicle data support, providing accurate temporal interval basis for subsequent data segment extraction and output.

[0070] In step S1031: In some embodiments, the computer device can use the vehicle data acquisition timestamp as a reference to perform interpolation and timing matching on the data features to complete global timestamp alignment; in some embodiments, the computer device can use the feature mutation timestamp of the data features as a reference to extract vehicle data within a fixed time window before and after to complete local timestamp alignment, wherein the feature mutation of the data features refers to the absolute value of the data features being greater than or equal to a first preset threshold; in some embodiments, the computer device uniformly uses the hardware clock of the data acquisition device as the global reference time, performs offset correction on the original timestamps of all data streams, and achieves timestamp synchronization alignment after eliminating the clock difference between multiple devices.

[0071] In step S1032, the data filtering rules include at least one of the following: using the received specified time period as the comparison time period, using the time period corresponding to the received specified timestamp point as the comparison time period, and using the time period in which the absolute value of the data feature is greater than or equal to a first preset threshold as the comparison time period.

[0072] The received specified time period refers to the user-defined time period received by the computer device. In step S1032, the output module can use the received specified time period as a comparison time period, which helps reduce the cost of manual retrieval and comparison of massive amounts of vehicle data and improves the efficiency of processing vehicle data to determine abnormal time intervals.

[0073] The received specified timestamp point refers to the timestamp point that the computer device receives from user-defined input.

[0074] In some embodiments, the timestamp in the subjectively marked data is the received specified timestamp point. In step S1032, the output module can use the time period corresponding to the received specified timestamp point as the comparison time period, which solves the problem that related technologies lack the function of locating specified data change points. It realizes the directional backtracking verification of key driving event segments marked by the driver's subjective mark, which helps to reduce the cost of manual retrieval and comparison of massive data and improves the efficiency of processing vehicle data to determine abnormal time series intervals.

[0075] The time period corresponding to the specified timestamp point refers to the continuous time interval formed by expanding in the directions before and after the specified timestamp point according to the preset time window parameters, with the specified timestamp point as the reference center point.

[0076] The time period in which the absolute value of a data feature is greater than or equal to the first preset threshold refers to the continuous time interval formed by the timestamps of the absolute values ​​of data features whose absolute values ​​exceed the first preset threshold.

[0077] In some embodiments, the first processing submodule can automatically identify several time periods in which the absolute value of data features exceeds a first preset threshold, and the visualization submodule can highlight these time periods in the data charts on the display interface. After detecting a user's double-click operation on the highlighted area, the output submodule at the bottom layer of the computer device will precisely and synchronously jump to the video frame corresponding to the highlighted area (e.g., a video frame showing a sudden shaking of the steering wheel), so that the user can simultaneously compare the vehicle dynamics data, physiological data, data features, and video data at the same moment, intuitively reproducing the complete driving scenario that induces physiological discomfort and negative driving experience for the driver. This is beneficial for quickly locating the causes of various poor driving comfort conditions and obtaining the influencing factors of driving comfort.

[0078] In step S1033, the computer device can identify, within the comparison time period, a time interval in which the absolute value of the data feature is greater than or equal to a first preset threshold and the vehicle data meets the anomaly verification conditions, as an abnormal time interval.

[0079] In one alternative embodiment of this application, data types other than vehicle dynamics data include video data and driver physiological data.

[0080] The vehicle data processing method provided in this application also includes: Step A11: Obtain the rate of change and mean of the physiological data.

[0081] Step S1033 involves identifying, within the comparison time period, time intervals where the absolute value of the data feature is greater than or equal to a first preset threshold and the vehicle data meets the anomaly verification conditions, as abnormal time intervals, including: Step A12: Identify time intervals that simultaneously meet the following conditions as abnormal time intervals: the absolute value of the data feature is greater than or equal to a first preset threshold, the rate of change is greater than or equal to a second preset threshold, the mean is greater than or equal to a third preset threshold, and the video data meets preset judgment conditions.

[0082] In this embodiment, data types other than vehicle dynamics data include video data and driver physiological data. Before determining abnormal time intervals, the computer device can first execute step A11 to obtain the rate of change and mean of physiological data, and then execute step A12 to identify time intervals that simultaneously meet the following conditions: the absolute value of the data features is greater than or equal to a first preset threshold, the rate of change is greater than or equal to a second preset threshold, the mean is greater than or equal to a third preset threshold, and the video data meets the preset judgment conditions. These time intervals are then considered abnormal time intervals. By introducing two types of quantitative physiological indicators, the rate of change and mean of physiological data, this application can accurately quantify the amplitude and trend of the driver's physiological fluctuations, distinguish between slight physiological fluctuations and persistent and significant physiological discomfort, and avoid misjudging abnormal working conditions based solely on instantaneous physiological data. Simultaneously, by introducing multi-dimensional preset judgment conditions covering vehicle driving status and driver behavior to analyze video data, including multi-dimensional judgment rules for lane approach, obstacle distance warning, abnormal facial expressions, and abnormal body movements, the determination of abnormal time intervals is extended from a purely numerical indicator dimension to the dimension of real working conditions and behavioral performance. This, combined with factors affecting the driver's subjective feelings, further enhances the understanding of abnormal time intervals. The changing data characteristics form a three-layer cross-verification mechanism of "subjective perception data characteristics - physiological quantitative indicators - visual scene verification". The judgment results of each dimension corroborate and complement each other, which can effectively eliminate misjudgments caused by single-dimensional data noise and random fluctuations. It is conducive to clarifying the temporal correspondence and causal relationship between vehicle driving status, driving and riding visual scene, human physiological fluctuations, and subjective discomfort, and improve the accuracy of abnormal time sequence interval identification. On this basis, it is conducive to accurately identifying the core causes of abnormal driving and riding comfort, improving the accuracy of judging the factors affecting driving and riding comfort, and providing reliable data support for vehicle driving and riding performance optimization.

[0083] In step A11, the second processing submodule in the computer device can process the physiological data in real time to obtain the rate of change and mean of the physiological data. The process of the second processing submodule processing the physiological data in real time to obtain the rate of change and mean of the physiological data specifically involves: first, parsing the multiple types of physiological data and sorting them by timestamp; then, obtaining the baseline value corresponding to each type of physiological data; and finally, using a sliding window algorithm to calculate the mean and rate of change of the physiological data within the sliding window, where the formula for calculating the rate of change is: Rate of change = (Mean - Baseline value) / Baseline value. By synchronously outputting the mean and rate of change of the physiological data through a sliding window streaming real-time computing mechanism, the magnitude of the driver's physiological data deviating from a relaxed and stable state can be quantified, providing standardized quantitative indicators for the subsequent identification of abnormal time series intervals.

[0084] In addition, during vehicle testing, the second processing submodule can automatically trim expired data within the sliding window to optimize memory usage.

[0085] In some embodiments, the control module in the computer device has network reconnection, device fault handling, and graceful shutdown mechanisms to adapt to complex field application scenarios such as vehicle road testing. Specifically, in device fault scenarios such as network interruption or offline data acquisition equipment, the control module can first perform a local backup of all vehicle data currently in the cache, then release all currently occupied data cache resources, and initiate link reconnection to access the test network environment. After the network link is reconnected, the control module retrieves the pre-backed-up vehicle data, restores the complete data resources cached before the fault disconnection, and continues the vehicle data acquisition and processing process. Thus, in the scenario of unexpected disconnection of the vehicle data acquisition link, all data before the fault occurs can be completely retained, avoiding the loss of massive amounts of vehicle data due to instantaneous disconnection. Through the processing logic of clearing resources, reconnecting, and restoring data after a fault, there is no need to manually reconfigure the acquisition process or repeat the whole vehicle test, which helps to ensure the continuity and integrity of the vehicle data processing process and improve the robustness of the vehicle data processing system in complex road test environments.

[0086] The second processing submodule supports manual input of baseline values ​​or automatic acquisition of baseline values. The baseline value of physiological data refers to the benchmark value of physiological data acquired when the driver is in a relaxed and stable state.

[0087] In step A12, the output module can compare the absolute value of the data features with a first preset threshold, the rate of change with a second preset threshold, and the mean with a third preset threshold through the output submodule, and match the video data with preset judgment conditions. It identifies time intervals that simultaneously satisfy: the absolute value of the data features is greater than or equal to the first preset threshold, the rate of change is greater than or equal to the second preset threshold, the mean is greater than or equal to the third preset threshold, and the video data meets the preset judgment conditions, and designates this time interval as an abnormal time interval.

[0088] In this application embodiment, the preset determination conditions include at least one of the following: the relative distance between the vehicle and the lane line is less than or equal to a preset approach threshold, the distance between the vehicle and the obstacle is less than or equal to a preset distance threshold, the driver's facial expression features include a first preset abnormal feature, and the driver's body movement features include a second preset abnormal feature.

[0089] The first preset abnormal feature refers to facial expression features obtained based on in-vehicle video image recognition that are associated with the driver's discomfort or physiological stress response. The first preset abnormal feature can intuitively reflect the driver's emotional fluctuations and physiological stress state. The first preset abnormal feature may include, but is not limited to, frowning, eyelid tension, clenched teeth, and facial muscle stiffness.

[0090] The second preset abnormal feature refers to the limb movement features obtained based on in-vehicle video image recognition that correspond to the driver's discomfort or stress state. The second preset abnormal feature can reflect the driver's subconscious behavioral reaction caused by abnormal comfort. The second preset abnormal feature may include, but is not limited to, gripping the steering wheel tightly, frequently adjusting the sitting posture, leaning forward, and subconsciously supporting the cockpit structure with the hands.

[0091] It is understandable that within the same time interval, if the absolute value of a data feature is less than the first preset threshold, the rate of change is less than the second preset threshold, the mean is less than the third preset threshold, or the video data does not meet at least one of the preset judgment conditions, then the time interval is not considered an abnormal time interval.

[0092] In one optional embodiment of this application, the vehicle dynamics data includes at least one data sequence, each data sequence corresponding to a different dynamics data type; step S102, which involves performing time-series signal processing on the vehicle dynamics data and extracting data features affecting changes in the driver's subjective perception from the processed data, includes: Step S1021: Upon receiving a data sequence selection instruction, the data sequence indicated by the data sequence selection instruction in the vehicle dynamics data is taken as the target data sequence.

[0093] Step S1022: Perform time-series signal processing on the target data sequence, and extract data features that affect the driver's subjective feelings from the processed data.

[0094] In this embodiment, the vehicle dynamics data includes at least one data sequence, each corresponding to a different dynamics data type. In step S1021, upon receiving a data sequence selection instruction, the first data processing submodule in the computer device can use the data sequence indicated by the data sequence selection instruction in the vehicle dynamics data as the target data sequence. The data sequence selection instruction is a data sequence selection instruction for vehicle dynamics data issued by the user to the computer device through a configuration file or interactive interface.

[0095] In step S1022, the first data processing submodule in the computer device can perform time-series signal processing on the target data sequence through the method described in step S102, and extract data features that affect the driver's subjective feelings from the processed data. This is beneficial for users to flexibly select the dynamic data types involved in data feature extraction according to the differentiated needs of the test target and test scenario, thereby improving the scenario adaptability and application flexibility of this application embodiment. At the same time, time-series signal processing and data feature extraction operations are only performed on the selected target data sequence, eliminating the need for redundant calculations on the full vehicle dynamic data sequence, which can effectively reduce the consumption of computing resources and improve the processing efficiency of data feature extraction. In addition, it can also eliminate the interference of irrelevant dynamic data, strengthen the correlation between data features and subjective feelings, and provide more reliable and effective data feature support for subsequent identification of abnormal time-series intervals and the location of factors affecting driving comfort.

[0096] In one optional embodiment of this application, the target data sequence includes at least one of a steering wheel angle data sequence, a lateral acceleration data sequence, and a vehicle distance from lane lines data sequence; step S1022, which involves performing time-series signal processing on the target data sequence and extracting data features that affect changes in the driver's subjective perception from the processed data, includes at least one of the following steps: Step B11: If the target data sequence includes the steering wheel angle data sequence, perform a first-order differential operation on the steering wheel angle data sequence to generate a steering wheel angle acceleration sequence, and extract the peak value from the steering wheel angle acceleration sequence to obtain the steering abruptness feature as a data feature.

[0097] Step B12: If the target data sequence includes the lateral acceleration data sequence, then the lateral acceleration data sequence is subjected to absolute value processing and low-pass filtering processing in sequence to obtain the lateral acceleration data sequence, and the yaw and tilt sensing features are determined as data features based on the lateral acceleration data sequence.

[0098] Step B13: If the target data sequence includes the vehicle distance data sequence from the lane line, then perform absolute value operation and dynamic threshold filtering on the vehicle distance data sequence from the lane line to identify high-risk moments of lane crossing and obtain lane approach pressure perception features as data features. The high-risk moments of lane crossing refer to the moments when the driver feels psychological pressure due to the approach of the lane line.

[0099] In this embodiment of the application, the target data sequence includes at least one of the following: steering wheel angle data sequence, lateral acceleration data sequence, and vehicle distance from lane line data sequence.

[0100] The first processing submodule includes a multi-dimensional mathematical transformation and filtering unit, which is under the control module's scheduling. During the time-series signal processing of the target data sequence and the extraction of data features affecting changes in the driver's subjective perception from the processed data, the multi-dimensional mathematical transformation and filtering unit can: if the target data sequence includes a steering wheel angle data sequence, execute step B11 to use the steering abruptness feature as a data feature; if the target data sequence includes a lateral acceleration data sequence, execute step B12 to use the yaw and roll perception feature as a data feature; if the target data sequence includes a vehicle-to-lane distance data sequence, execute step B13 to use the lane approach pressure perception feature as a data feature. Therefore, data features that directly reflect changes in the driver's subjective feelings can be extracted from objective vehicle dynamics data. This transforms objective vehicle dynamics data into quantifiable and horizontally comparable data features at the human perception level, establishing a correlation between vehicle dynamics data and the driver's subjective feelings. Simultaneously, based on the dynamics data type of the target data sequence, the operational logic of time-series signal processing and data feature extraction is differentiated. For three typical driving comfort impact conditions—steering impact, lateral bumps, and lane line approach—corresponding data features are extracted, which helps improve the accuracy of data features in representing changes in the driver's subjective feelings. This facilitates the accurate identification of abnormal time-series intervals and factors influencing driving comfort.

[0101] In step B11, when the target data sequence includes a steering wheel angle data sequence, firstly, the multi-dimensional mathematical transformation and filtering unit performs a first-order differential operation on the steering wheel angle data sequence to generate a steering wheel angle acceleration sequence characterizing the change in steering operation rate. Then, the multi-dimensional mathematical transformation and filtering unit performs absolute value processing on the steering wheel angle acceleration sequence and extracts the peak value from the absolute value processed steering wheel angle acceleration sequence to obtain the steering abruptness feature. The steering abruptness feature is used to quantify the intensity of the "abruptness" in the driver's operation.

[0102] In step B12, when the target data sequence includes a lateral acceleration data sequence, firstly, the multi-dimensional mathematical transformation and filtering unit performs absolute value processing and low-pass filtering on the lateral acceleration data sequence to eliminate high-frequency road vibration noise interference and extract a lateral acceleration data sequence that reflects the driver's actual perception. Then, the multi-dimensional mathematical transformation and filtering unit determines the yaw and roll perception features based on the lateral acceleration data sequence. These yaw and roll perception features are used to assess the lateral instability felt by the driver.

[0103] In step B13, when the target data sequence includes a vehicle-to-lane-line distance data sequence, the multi-dimensional mathematical transformation and filtering unit performs absolute value operation and dynamic threshold filtering on the vehicle-to-lane-line distance data sequence. Based on the dynamic threshold filtering results, it automatically identifies and marks high-risk moments when the driver feels psychological pressure due to lane line approach, thus obtaining lane approach pressure perception features.

[0104] In some embodiments, the data features extracted by the computer device in step S1022 include at least one of steering abruptness features, yaw and roll perception features, and lane approach pressure perception features.

[0105] In some embodiments, the data features extracted through at least one of steps B11 to B13 can be output and stored in the form of a new data table.

[0106] In an optional embodiment of this application, step S1021, upon receiving a data sequence selection instruction, involves using the data sequence indicated by the data sequence selection instruction in the vehicle dynamics data as the target data sequence, which includes: Step C11: Extract data within the target time interval from the vehicle dynamics data as slice data.

[0107] Step C12: Upon receiving a data sequence selection instruction, the data sequence indicated by the data sequence selection instruction in the sliced ​​data is taken as the target data sequence.

[0108] In this embodiment of the application, the first processing submodule further includes a custom data slicing unit, and the custom data slicing unit is placed under the scheduling of the control module.

[0109] In determining the target data sequence, the first processing submodule can use a custom data slicing unit to execute step C11 to extract data within the target time interval from the vehicle dynamics data as slice data. This cuts the original long sequence of vehicle dynamics data (which is tens of minutes long) into data segments with high information density centered on effective driving events, thereby eliminating redundant non-operational cruise data intervals. This refines the analysis granularity from the statistical analysis of the entire continuous vehicle dynamics data to the second-level driving micro-operation interval, providing focused data input for subsequent time-series signal processing and data feature extraction. Then, step C12 can be executed. Upon receiving a data sequence selection instruction, the data sequence indicated by the data sequence selection instruction in the slice data can be used as the target data sequence.

[0110] It is understood that vehicle dynamics data includes at least one data sequence, each corresponding to a different dynamics data type. For example, the dynamics data type includes longitudinal acceleration, lateral acceleration, steering wheel angle, steering wheel angular rate, and vehicle distance from lane lines. In step C11, the custom data slicing unit can synchronously slice the longitudinal acceleration sequence, lateral acceleration sequence, steering wheel angle sequence, steering wheel angular rate sequence, and vehicle distance from lane lines sequence within the target time interval of the vehicle dynamics data, generating time-aligned slice data. The slice data includes time-aligned longitudinal acceleration data sequences, lateral acceleration data sequences, steering wheel angle data sequences, steering wheel angular rate data sequences, and vehicle distance from lane lines data sequences.

[0111] It is understandable that the time series interval of the slice data obtained through step C11 is aligned with the target time series interval.

[0112] In step C12, upon receiving a data sequence selection instruction, the first processing submodule can use a method similar to step S1021 to select the data sequence indicated by the data sequence selection instruction in the sliced ​​data as the target data sequence.

[0113] In an optional embodiment of this application, before step C11, which involves extracting data within a target time interval from the vehicle dynamics data as slice data, the method further includes: Step C13: Upon receiving a data slicing instruction, obtain the event timestamp carried by the data slicing instruction.

[0114] Step C14: Determine the target time interval based on the event timestamp.

[0115] Specifically, upon receiving a data slicing instruction, the custom data slicing unit first obtains the event timestamp carried by the data slicing instruction, and then determines the target time series interval based on the event timestamp. This facilitates flexible custom configuration of the target time series interval, adapts to the differentiated analysis needs of different human factors analysis and driving comfort tests, and helps to broaden the applicable scenarios for fine-grained segmentation of vehicle data.

[0116] The data slicing command is a command input by the user into the computer device, and it carries an event timestamp. The event timestamp can be a single timestamp or a continuous time interval consisting of two timestamps.

[0117] In some embodiments, the event timestamp is a single timestamp point. In step C14, the custom data slice unit can define a continuous time interval centered on the event timestamp as the target time interval. The duration of the continuous time interval can be customized according to the time length requirements of the slice data.

[0118] In some embodiments, the event timestamp is a continuous time interval consisting of two timestamp points. In step C14, the custom data slicing unit can use this continuous time interval as the target time interval.

[0119] In an optional embodiment of this application, before step C11, which involves extracting data within a target time interval from the vehicle dynamics data as slice data, the method further includes: Step C15: If the vehicle dynamics data meets the slicing conditions, determine the moment when the vehicle dynamics data meets the slicing conditions.

[0120] Step C16: Determine the target time interval based on the time when the vehicle dynamics data meets the slicing conditions.

[0121] Specifically, when the vehicle dynamics data meets the slicing conditions, the custom data slicing unit first determines the time when the vehicle dynamics data meets the slicing conditions, and then determines the target time series interval based on the time when the vehicle dynamics data meets the slicing conditions. Thus, the process of determining the target time series interval can be fully automated without the need for the user to monitor the process in real time and manually issue slicing instructions. This helps to reduce the cost of manual operation and improve the automation level and overall analysis efficiency of fine-grained segmentation processing of multi-dimensional vehicle dynamics data.

[0122] Slicing conditions refer to pre-configured data threshold conditions used to determine whether to trigger the automatic extraction of slices from vehicle dynamics data.

[0123] When the computer device determines that the vehicle dynamics data meets the slicing conditions, it can synchronously record the moment when the vehicle dynamics data meets the slicing conditions, so that the custom data slicing unit can find the moment when the vehicle dynamics data meets the slicing conditions during the execution of step C15.

[0124] In step C16, the custom data slicing unit defines a continuous time interval centered on the moment when the vehicle dynamics data meets the slicing conditions, which serves as the target time interval. The duration of this continuous time interval can be customized according to the time length requirements of the sliced ​​data.

[0125] In one optional embodiment of this application, the vehicle dynamics data includes lateral acceleration and steering wheel angular rate; the vehicle data processing method provided in this application further includes: Step C17: If the lateral acceleration is greater than a preset acceleration threshold and the steering wheel angular rate is greater than a preset angular rate threshold, determine that the vehicle dynamics data meets the slicing condition.

[0126] In this embodiment of the application, in a scenario where vehicle dynamics data includes lateral acceleration and steering wheel angular rate, the computer device can match the lateral acceleration with a preset acceleration threshold and the steering wheel angular rate with a preset angular rate threshold.

[0127] If the lateral acceleration exceeds a preset acceleration threshold and the steering wheel angular rate exceeds a preset angular rate threshold, step C17 is executed to determine that the vehicle dynamics data meets the slicing conditions. For example, the preset acceleration threshold is 0.3 m / s². 2 The preset turning rate threshold is 50 deg / s.

[0128] If the lateral acceleration is less than or equal to a preset acceleration threshold, and / or the steering wheel angular rate is less than or equal to a preset angular rate threshold, the vehicle dynamics data is determined not to meet the slicing conditions.

[0129] In this embodiment of the application, by determining whether the lateral acceleration in the vehicle dynamics data is greater than a preset acceleration threshold and whether the steering wheel angular rate is greater than a preset angular rate threshold, the automatic determination of whether the vehicle dynamics data meets the slicing conditions can be achieved. This process does not require manual intervention by the user, which helps to reduce the cost of manual operation and improve the automation level and overall analysis efficiency of fine-grained segmentation processing of multi-dimensional vehicle dynamics data.

[0130] As an example, refer to Figure 3 The diagram illustrates a flowchart of another vehicle data processing method provided in this application embodiment. This vehicle data processing method may specifically include the following steps: Step S201: Obtain vehicle data.

[0131] Specifically, the computer device can acquire vehicle data using the same method as in step S101 described above. The vehicle data includes the driver's physiological data, in-vehicle video, out-of-vehicle video, vehicle dynamics data, and subjective events. The vehicle data can be divided into three dimensions: person (physiological data, in-vehicle video, subjective events), vehicle (vehicle dynamics data), and road (out-of-vehicle video).

[0132] Subjective events include subjectively labeled data.

[0133] Step S202: Timestamp alignment and data storage.

[0134] Specifically, the computer device can timestamp-align different types of test data in the vehicle data obtained through step S201 and store the timestamp-aligned vehicle data.

[0135] Step S203 involves slicing, removing, and filtering data, and extracting data features through absolute value analysis, filtering, and differentiation.

[0136] Specifically, the first processing submodule in the computer device can first extract slice data from the vehicle dynamics data using the same method as in step C11, thereby eliminating redundant non-operational cruise data intervals and refining the analysis granularity from the entire continuous segment of vehicle dynamics data to the second-level driving micro-operation interval, providing focused data input for subsequent time-series signal processing and data feature extraction processes; then, the target data sequence is selected from the slice data using the same method as in step C12; finally, the target data sequence is processed using the same methods as in steps B11 to B13, through absolute value, filtering, differentiation, etc., and data features affecting changes in the driver's subjective feelings are extracted from the processed data.

[0137] Step S204: Match the timestamps of subjective events and generate a list of candidate events.

[0138] Specifically, computer equipment can generate a list of candidate events in chronological order based on the timestamps of subjectively labeled data in vehicle data.

[0139] Step S205: Closed-loop verification of people-vehicle-road system.

[0140] Specifically, the output module in the computer device can perform closed-loop verification of vehicle data across the three dimensions of people, vehicles, and roads according to the candidate event list generated in step S204. Specifically, the visualization submodule can display physiological data charts, vehicle dynamics data charts, and video data in real time. The output submodule, using the same method as steps S1031 to S1033 and steps A11 to A12, performs jump-back comparisons of the candidate event list, physiological data charts, vehicle dynamics data charts, and video data. It identifies abnormal time intervals where the absolute value of data features is greater than or equal to a first preset threshold, the rate of change of physiological data is greater than or equal to a second preset threshold, the mean of physiological data is greater than or equal to a third preset threshold, and the video data meets preset judgment conditions. This enables retrospective comparative analysis based on user-marked or selected discomfort events.

[0141] Step S206: Export data fragments.

[0142] Specifically, the output module in the computer device can extract data segments within the abnormal time interval from the data features and vehicle data using the same method as in step S104, and output the data segments for subsequent analysis.

[0143] Reference Figure 4 This illustrates a schematic diagram of the logical architecture of another computer device provided in an embodiment of this application, such as... Figure 4 As shown, the logical architecture of the computer device includes a data acquisition module, a first processing submodule, a visualization submodule, a control module, an output submodule, a second processing submodule, and a network module. The system comprises the following modules: a data acquisition module for acquiring physiological data, video data, vehicle dynamics data, and subjectively labeled data; a second processing submodule for parsing, calculating, and storing physiological data; a visualization submodule for displaying real-time change curves of vehicle data through dynamic charts, showing status information, and providing interactive buttons; a control module for managing the on / off status of data acquisition devices (cameras, microphones) and the start / stop and storage of video data recording; an output submodule for identifying abnormal time intervals based on video data, vehicle dynamics data, physiological data, and data features, where the absolute value of data features is greater than or equal to a first preset threshold, the rate of change of physiological data is greater than or equal to a second preset threshold, the mean of physiological data is greater than or equal to a third preset threshold, and the video data meets preset judgment conditions; an output submodule for extracting data segments within abnormal time intervals from data features and vehicle data and outputting the data segments; a first processing submodule for performing time-series signal processing on vehicle dynamics data using absolute value, filtering, and other methods, and extracting data features that affect changes in the driver's subjective feelings from the processed data; and a network module for network connection, heart rate maintenance, and automatic network reconnection.

[0144] As an example, in Figure 4 Under the logical architecture of the computer device shown, the vehicle data processing method provided in this application may include the following steps: Step S301: Initialize the vehicle data processing system: Load configuration parameters, initialize the camera, network protocol and visualization interface, and check the camera availability.

[0145] Step S302: Establish connection: Establish a connection with third-party software through network communication, send an initial request to obtain the base timestamp and channel configuration, and start the heartbeat thread and receiving thread.

[0146] Step S303, Vehicle Data Acquisition: Receive and parse physiological data, collect video data, record subjective events, and acquire vehicle dynamics data.

[0147] Step S304, Physiological data processing: Collect baseline values, calculate the rate of change of physiological data through a sliding window, and promptly prune outdated data in the sliding window.

[0148] Step S305, Data Visualization: Real-time updates of vehicle data change curves and status information.

[0149] Step S306, Data storage: Store the processed physiological data, raw physiological data, and subjective events as CSV files, and store the video data as video files.

[0150] Step S307, Data Processing: Perform time-series signal processing on the vehicle data stored in step S306 above, and extract data features that affect the driver's subjective feelings from the processed data; based on the data features and vehicle data, identify abnormal time intervals where the absolute value of the data features is greater than or equal to a first preset threshold and the vehicle data meets the abnormality verification conditions; and extract data segments within the abnormal time intervals from the data features and vehicle data, and output the data segments.

[0151] Specifically, the vehicle data processing system includes computer equipment, data acquisition equipment, and network connection equipment. The computer equipment and data acquisition equipment establish a communication connection through the network connection equipment. Among them, the data acquisition equipment includes multiple Universal Serial Bus (USB) cameras and a custom external keyboard.

[0152] When implementing the vehicle data processing method provided in this application, ensure that all devices are correctly connected and powered on. After deploying the data acquisition equipment, start the vehicle data acquisition program. After detecting that the user clicks the "Connect Service" button, the computer device establishes a communication connection with the server. After detecting that the user clicks the "Collect Baseline" button, the computer device receives data from the server in real time and uses the received data as the baseline value corresponding to the physiological data. After the baseline value acquisition is complete, and the computer device detects that the user clicks the "Stop Acquisition" button, the baseline value acquisition process ends, and the program enters the waiting task start phase.

[0153] When the task begins, the computer device detects that the user has clicked the "Start Task" button, reads the internal parameters of the configuration file to obtain the sampling rate of the data acquisition device, uses the sampling rate as the window size of the sliding window, calculates the real-time physiological data, and stores the processed physiological data and the original physiological data in CSV format.

[0154] After the task is started, the computer equipment synchronously controls the data acquisition equipment to collect video data, and stores the audio and video frames in the video data in WAV and AVI formats respectively.

[0155] During task execution, the visualization submodule in the computer device displays real-time change curves of structured data such as vehicle dynamics and physiological data in the current vehicle data in chart form, and simultaneously retrieves video data for the corresponding time period for interactive display. The computer device monitors the marking operation signal of the custom external keyboard in real time. If a marking operation signal is detected, corresponding subjective marking data is generated and stored in CSV format. After detecting that the user clicks the "End Task" button, the computer device stores all vehicle data for this task in a preset designated folder.

[0156] After detecting a user clicking the "Playback" button, the computer device redirects to the data playback page. When the user selects a CSV data file to be played back, the computer device automatically retrieves and loads matching video data based on the file name. The computer device supports user-specified key time periods for targeted playback. The data playback interface uses a horizontal layout, sequentially displaying video footage and data charts, including physiological data charts, vehicle dynamics data charts, and subjectively labeled data. Based on a timestamp alignment mechanism, the computer device synchronously plays back the physiological data charts, vehicle dynamics data charts, subjectively labeled data, and video data. When the computer device detects a user clicking on the timeline in a data chart, it jumps to the corresponding time point in the chart and synchronizes the playback of the corresponding audio and video frames. The computer device also supports exporting data segments corresponding to user-selected time periods for subsequent analysis.

[0157] Furthermore, in a specific embodiment involving vehicle dynamics data analysis, after importing vehicle dynamics data, the computer device can monitor in real time the target calculation instructions for human factors analysis input by the user through a custom interface. For example, the target calculation instruction input by the user includes the following formula: [Steering Abruptness Index] = ABS(DIFF([Steering Wheel Angle])) / [Vehicle Speed]. After detecting the user's click on the batch execution control, the computer device traverses all slice data within seconds and calculates and generates a new dimensional feature curve that reflects the intensity of the driver's operation and potential subjective discomfort. Meanwhile, the computer equipment can automatically identify several abnormal time segments where the value of the characteristic curve exceeds the first preset threshold, and highlight the abnormal time segments in the data chart of the playback interface; after the computer equipment detects the user's double-click operation on the highlighted area, the output submodule of the computer equipment will accurately and synchronously jump to the video frame where the steering wheel shakes rapidly. This function effectively solves the technical blind spot in traditional whole vehicle test analysis where "the test data is known to fluctuate, but the driver's real state and psychological feelings at the corresponding moment cannot be synchronously traced back", and realizes the deep integration and analysis of human-vehicle-road vehicle data.

[0158] In summary, the vehicle data processing method provided in this application acquires vehicle data of multiple data types, including vehicle dynamics data and other data types. The vehicle dynamics data is objective data generated during vehicle operation, providing a data foundation for multi-dimensional data processing. Subsequently, by performing time-series signal processing on the vehicle dynamics data and extracting data features affecting changes in the driver's subjective feelings from the processed data, data features reflecting the driver's subjective feelings can be extracted from purely objective vehicle dynamics data. This establishes a quantitative correlation between objective vehicle dynamics data and the driver's subjective feelings, effectively bridging the analytical gap between objective data and the driver's subjective feelings in vehicle testing and driving behavior research. Finally, based on the data features and vehicle data... The system identifies abnormal time-series intervals where the absolute value of data features is greater than or equal to a first preset threshold and the vehicle data meets the anomaly verification conditions. Based on the joint verification of multi-dimensional data, it can accurately pinpoint the time periods that cause abnormal subjective feelings in drivers, narrowing the analysis scope, refining the granularity of time-series analysis, and effectively improving the efficiency and accuracy of anomaly localization. Finally, it extracts data fragments within the abnormal time-series intervals from data features and vehicle data and outputs these data fragments for subsequent analysis, forming a complete analytical closed loop from objective data processing and extraction of data features affecting changes in drivers' subjective feelings to the localization and output of data fragments within the abnormal time-series intervals. This improves the accuracy and reliability of vehicle comfort test results determined by processing vehicle data, facilitates the accurate identification of factors influencing ride comfort, and provides reliable data support for optimizing vehicle ride performance.

[0159] Furthermore, the first processing submodule includes a custom data slicing unit and a multi-dimensional mathematical transformation and filtering unit. The custom data slicing unit extracts data within a target time interval from vehicle dynamics data as sliced ​​data. The multi-dimensional mathematical transformation and filtering unit performs time-series signal processing on the target data sequence in the sliced ​​data and extracts data features affecting changes in the driver's subjective feelings from the processed data. This provides a batch processing mechanism capable of fine-grained, custom logical segmentation and mathematical transformation of multi-dimensional vehicle data according to subjective needs. This allows users to efficiently locate key driving event segments that trigger physiological discomfort or special experiences from massive amounts of vehicle data. Ultimately, this integrated vehicle data processing method achieves simultaneous multi-dimensional vehicle data acquisition, real-time online analysis, dynamic visualization, custom batch data processing, abnormal time interval identification, and data segment output.

[0160] Device Examples Reference Figure 5 The diagram illustrates a logic block diagram of a vehicle data processing device according to an embodiment of this application. The device includes: The data acquisition module 12 is used to acquire vehicle data, which includes data of multiple data types, including vehicle dynamics data and other data types besides vehicle dynamics data; the vehicle dynamics data is objective data generated during vehicle operation. Data processing module 13 is used to perform time-series signal processing on the vehicle dynamics data and extract data features that affect changes in the driver's subjective feelings from the processed data; Output module 14 is used to identify, based on the data features and the vehicle data, an abnormal time series interval where the absolute value of the data features is greater than or equal to a first preset threshold and the vehicle data meets the abnormality verification conditions. The output module 14 is further configured to extract data segments within the abnormal time interval from the data features and the vehicle data, and output the data segments.

[0161] Optionally, the output module 14 is specifically used for: The data features are timestamped together with the vehicle data. According to the data filtering rules, the comparison time period is determined. The data filtering rules include at least one of the following: using the received specified time period as the comparison time period, using the time period corresponding to the received specified timestamp point as the comparison time period, and using the time period where the absolute value of the data feature is greater than or equal to a first preset threshold as the comparison time period. Within the comparison time period, the time interval in which the absolute value of the data feature is greater than or equal to a first preset threshold and the vehicle data meets the anomaly verification conditions is identified as the abnormal time interval.

[0162] Optionally, the data types other than vehicle dynamics data include video data and driver physiological data; the data processing module 13 includes a second data processing submodule for: acquiring the rate of change and mean of the physiological data; The output module 14 is specifically used to: identify time intervals that simultaneously meet the following conditions as abnormal time intervals: the absolute value of the data feature is greater than or equal to a first preset threshold, the rate of change is greater than or equal to a second preset threshold, the mean is greater than or equal to a third preset threshold, and the video data meets preset judgment conditions. The preset judgment conditions include at least one of the following: the relative distance between the vehicle and the lane line is less than or equal to a preset approach threshold, the distance between the vehicle and the obstacle is less than or equal to a preset distance threshold, the driver's facial expression features include a first preset abnormal feature, and the driver's body movement features include a second preset abnormal feature.

[0163] Optionally, the vehicle dynamics data includes at least one data sequence, each data sequence corresponding to a different dynamics data type; the data processing module 13 includes a first data processing submodule, used for: Upon receiving a data sequence selection instruction, the data sequence indicated by the data sequence selection instruction in the vehicle dynamics data is taken as the target data sequence; The target data sequence is subjected to time-series signal processing, and data features that affect changes in the driver's subjective feelings are extracted from the processed data.

[0164] Optionally, the target data sequence includes at least one of a steering wheel angle data sequence, a lateral acceleration data sequence, and a vehicle distance from lane lines data sequence; the first data processing submodule is specifically used for: If the target data sequence includes the steering wheel angle data sequence, then perform a first-order differential operation on the steering wheel angle data sequence to generate a steering wheel angle acceleration sequence, and extract the peak value from the steering wheel angle acceleration sequence to obtain the steering abruptness feature as a data feature; If the target data sequence includes the lateral acceleration data sequence, then the lateral acceleration data sequence is subjected to absolute value processing and low-pass filtering processing in sequence to obtain the lateral acceleration data sequence, and the yaw and tilt sensing features are extracted as data features based on the lateral acceleration data sequence. If the target data sequence includes the vehicle-to-lane-line distance data sequence, then the absolute value operation and dynamic threshold filtering are performed on the vehicle-to-lane-line distance data sequence to extract the high-risk moments of lane crossing and obtain the lane approach pressure perception features as data features. The high-risk moments of lane crossing refer to the moments when the driver feels psychological pressure due to the approach of the lane line.

[0165] Optionally, the first data processing submodule is specifically used for: Data within the target time interval is extracted from the vehicle dynamics data and used as slice data; Upon receiving a data sequence selection instruction, the data sequence indicated by the data sequence selection instruction in the sliced ​​data is taken as the target data sequence.

[0166] Optionally, the first data processing submodule is further configured to: Upon receiving a data slicing instruction, obtain the event timestamp carried by the data slicing instruction; The target time interval is determined based on the event timestamp.

[0167] Optionally, the first data processing submodule is further configured to: If the vehicle dynamics data satisfies the slicing condition, determine the moment when the vehicle dynamics data satisfies the slicing condition. The target time interval is determined based on the time when the vehicle dynamics data meets the slicing conditions.

[0168] Optionally, the vehicle dynamics data includes lateral acceleration and steering wheel angular rate; the first data processing submodule is further configured to: If the lateral acceleration is greater than a preset acceleration threshold and the steering wheel angular rate is greater than a preset angular rate threshold, the vehicle dynamics data is determined to meet the slicing conditions.

[0169] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.

[0170] This application also provides an electronic device, which includes a memory and a processor. The memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory. When the processor executes the computer program, it implements the vehicle data processing method as described in any of the preceding claims.

[0171] This application also provides a readable storage medium storing a computer program that, when executed by a processor, implements the vehicle data processing method as described in any of the preceding claims.

[0172] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0173] The above embodiments are merely preferred embodiments provided to fully illustrate this application, and the scope of protection of this application is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on this application are all within the scope of protection of this application.

Claims

1. A vehicle data processing method, characterized in that, The method includes: Acquire vehicle data, which includes data of multiple data types, including vehicle dynamics data and other data types besides vehicle dynamics data; the vehicle dynamics data is objective data generated during vehicle operation. The vehicle dynamics data is subjected to time-series signal processing, and data features that affect changes in the driver's subjective feelings are extracted from the processed data. Based on the data features and the vehicle data, identify abnormal time intervals where the absolute value of the data features is greater than or equal to a first preset threshold and the vehicle data meets the abnormality verification conditions. Extract data segments within the abnormal time interval from the data features and the vehicle data, and output the data segments.

2. The method according to claim 1, characterized in that, The step of identifying abnormal time intervals where the absolute value of the data features is greater than or equal to a first preset threshold and the vehicle data meets the anomaly verification conditions based on the data features and the vehicle data includes: The data features are timestamped together with the vehicle data. According to the data filtering rules, the comparison time period is determined. The data filtering rules include at least one of the following: using the received specified time period as the comparison time period, using the time period corresponding to the received specified timestamp point as the comparison time period, and using the time period where the absolute value of the data feature is greater than or equal to a first preset threshold as the comparison time period. Within the comparison time period, the time interval in which the absolute value of the data feature is greater than or equal to a first preset threshold and the vehicle data meets the anomaly verification conditions is identified as the abnormal time interval.

3. The method according to claim 2, characterized in that, Data types other than vehicle dynamics data include video data and driver physiological data; The method further includes: Obtain the rate of change and mean of the physiological data; The time interval within the comparison time period where the absolute value of the data feature is greater than or equal to a first preset threshold and the vehicle data meets the anomaly verification conditions is defined as the abnormal time interval, including: A time series interval that simultaneously meets the following conditions is identified as an abnormal time series interval: the absolute value of the data feature is greater than or equal to a first preset threshold, the rate of change is greater than or equal to a second preset threshold, the mean is greater than or equal to a third preset threshold, and the video data meets preset judgment conditions. The preset judgment conditions include at least one of the following: the relative distance between the vehicle and the lane line is less than or equal to a preset approach threshold, the distance between the vehicle and the obstacle is less than or equal to a preset distance threshold, the driver's facial expression features include a first preset abnormal feature, and the driver's body movement features include a second preset abnormal feature.

4. The method according to claim 1, characterized in that, The vehicle dynamics data includes at least one data sequence, and each data sequence corresponds to a different dynamics data type; The step of performing time-series signal processing on the vehicle dynamics data and extracting data features that affect changes in the driver's subjective perception from the processed data includes: Upon receiving a data sequence selection instruction, the data sequence indicated by the data sequence selection instruction in the vehicle dynamics data is taken as the target data sequence; The target data sequence is subjected to time-series signal processing, and data features that affect changes in the driver's subjective feelings are extracted from the processed data.

5. The method according to claim 4, characterized in that, The target data sequence includes at least one of the following: steering wheel angle data sequence, lateral acceleration data sequence, and vehicle distance from lane line data sequence; The step of performing time-series signal processing on the target data sequence and extracting data features that affect changes in the driver's subjective perception from the processed data includes at least one of the following steps: If the target data sequence includes the steering wheel angle data sequence, then perform a first-order differential operation on the steering wheel angle data sequence to generate a steering wheel angle acceleration sequence, and extract the peak value from the steering wheel angle acceleration sequence to obtain the steering abruptness feature as a data feature; If the target data sequence includes the lateral acceleration data sequence, then the lateral acceleration data sequence is subjected to absolute value processing and low-pass filtering processing in sequence to obtain the lateral acceleration data sequence, and the yaw and tilt sensing features are extracted as data features based on the lateral acceleration data sequence. If the target data sequence includes the vehicle-to-lane-line distance data sequence, then the absolute value operation and dynamic threshold filtering are performed on the vehicle-to-lane-line distance data sequence to extract the high-risk moments of lane crossing and obtain the lane approach pressure perception features as data features. The high-risk moments of lane crossing refer to the moments when the driver feels psychological pressure due to the approach of the lane line.

6. The method according to claim 4, characterized in that, Upon receiving a data sequence selection instruction, the step of selecting the data sequence indicated by the data sequence selection instruction from the vehicle dynamics data as the target data sequence includes: Data within the target time interval is extracted from the vehicle dynamics data and used as slice data; Upon receiving a data sequence selection instruction, the data sequence indicated by the data sequence selection instruction in the sliced ​​data is taken as the target data sequence.

7. The method according to claim 6, characterized in that, Before extracting data within a target time interval from the vehicle dynamics data as slice data, the method further includes: Upon receiving a data slicing instruction, obtain the event timestamp carried by the data slicing instruction; The target time interval is determined based on the event timestamp.

8. The method according to claim 6, characterized in that, Before extracting data within a target time interval from the vehicle dynamics data as slice data, the method further includes: If the vehicle dynamics data satisfies the slicing condition, determine the moment when the vehicle dynamics data satisfies the slicing condition. The target time interval is determined based on the time when the vehicle dynamics data meets the slicing conditions.

9. The method according to claim 8, characterized in that, The vehicle dynamics data includes lateral acceleration and steering wheel angular velocity; the method further includes: If the lateral acceleration is greater than a preset acceleration threshold and the steering wheel angular rate is greater than a preset angular rate threshold, the vehicle dynamics data is determined to meet the slicing conditions.

10. A vehicle data processing device, characterized in that, The device includes: The data acquisition module is used to acquire vehicle data, which includes data of multiple data types, including vehicle dynamics data and other data types; the vehicle dynamics data is objective data generated during vehicle operation. The data processing module is used to perform time-series signal processing on the vehicle dynamics data and extract data features that affect changes in the driver's subjective feelings from the processed data. The output module is used to identify, based on the data features and the vehicle data, an abnormal time series interval where the absolute value of the data features is greater than or equal to a first preset threshold and the vehicle data meets the abnormality verification conditions. The output module is further configured to extract data segments within the abnormal time interval from the data features and the vehicle data, and output the data segments.

11. An electronic device, characterized in that, The system includes a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory, wherein the processor executes the computer program to implement the vehicle data processing method as described in any one of claims 1 to 9.

12. A readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the vehicle data processing method as described in any one of claims 1 to 9.

Images