Vehicle performance evaluation method and device, electronic equipment and vehicle

Abstract

The invention discloses a vehicle performance evaluation method and device, electronic equipment and a vehicle. The method comprises the steps that in response to the fact that a vehicle is in an auxiliary driving mode, physiological data of a driver of the vehicle and driving data of the vehicle are collected; based on a first timestamp corresponding to the physiological data and a second timestamp corresponding to the driving data, aligning the physiological data and the driving data on a time axis to obtain an alignment result; based on scene identification information in the driving data, dividing the alignment result to obtain a scene data packet corresponding to at least one driving scene; determining a plurality of evaluation indexes of the vehicle based on the scene data packet; and determining a performance evaluation result of the auxiliary driving mode based on the plurality of evaluation indexes. The technical problem that the performance evaluation accuracy of the vehicle auxiliary driving mode is low is solved.

Description

Technical Field

[0001] This application relates to the field of vehicle technology, and more specifically, to a method, apparatus, electronic device, and vehicle for evaluating vehicle performance. Background Technology

[0002] Currently, with the development of vehicle driver assistance technology, users often experience poor user experience due to performance issues of driver assistance modes. Therefore, it is necessary to conduct accurate evaluations of the performance of vehicle driver assistance modes.

[0003] In related technologies, performance evaluation of assisted driving modes often relies on vehicle operating data or driver subjective questionnaires. Vehicle operating data cannot reflect the user's real experience, while driver subjective questionnaires, due to lag and memory bias, are difficult to make fair judgments, thus making it impossible to accurately evaluate the performance of vehicle assisted driving modes.

[0004] There is currently no effective solution to the technical problem of low accuracy in performance evaluation of vehicle assisted driving modes. Summary of the Invention

[0005] This application provides a vehicle performance evaluation method, apparatus, electronic device, and vehicle to at least solve the technical problem of low accuracy in performance evaluation of vehicle assisted driving modes.

[0006] According to one aspect of the embodiments of this application, a method for evaluating vehicle performance is provided. The method may include: in response to the vehicle being in an assisted driving mode, collecting physiological data of the driver and driving data of the vehicle, wherein the physiological data characterizes the driver's physiological state and behavioral responses, and the driving data characterizes the driving operations and dynamic characteristics performed by the vehicle in the assisted driving mode; aligning the physiological data and driving data on a time axis based on a first timestamp corresponding to the physiological data and a second timestamp corresponding to the driving data to obtain an alignment result, wherein the first timestamp identifies the time when the physiological data was generated, and the second timestamp identifies the time when the driving data was generated; dividing the alignment result based on scene identification information in the driving data to obtain a scene data packet corresponding to at least one driving scenario, wherein the scene data packet includes the physiological data and driving data corresponding to the driving scenario; determining multiple evaluation indicators for the vehicle based on the scene data packet, wherein the multiple evaluation indicators characterize the driver's satisfaction with the assisted driving mode in at least one driving scenario from multiple dimensions; and determining the performance evaluation result of the assisted driving mode based on the multiple evaluation indicators.

[0007] Optionally, based on the first timestamp corresponding to the physiological data and the second timestamp corresponding to the driving data, the physiological data and driving data are aligned on the time axis to obtain an alignment result, including: determining the physiological data and driving data corresponding to the same time based on the first timestamp corresponding to the physiological data and the second timestamp corresponding to the driving data; aligning the physiological data and driving data corresponding to the same time to obtain aligned physiological data and driving data; and determining the aligned physiological data and driving data as the alignment result.

[0008] Optionally, based on the scene identification information in the driving data, the alignment results are divided to obtain scene data packets corresponding to at least one driving scene, including: determining the start time point and end time point of at least one driving scene in the driving data based on the scene identification information in the driving data; and dividing the physiological data and driving data in the alignment results based on the start time point and end time point to obtain scene data packets corresponding to at least one driving scene.

[0009] Optionally, based on the scenario data package, multiple evaluation indicators for the vehicle are determined, including: preprocessing the scenario data package corresponding to at least one driving scenario to obtain a preprocessed scenario data package, wherein the preprocessing is used to characterize the removal of invalid data in the scenario data package; and based on the preprocessed scenario data package, multiple evaluation indicators for the vehicle are determined, wherein the multiple evaluation indicators include at least a comfort evaluation indicator and a trust evaluation indicator, wherein the comfort evaluation indicator is used to characterize the driver's satisfaction with the comfort of the assisted driving mode, and the trust evaluation indicator is used to characterize the driver's satisfaction with the reliability of the assisted driving mode.

[0010] Optionally, based on the preprocessed scene data package, the vehicle comfort evaluation index is determined, including: determining the absolute standard deviation of the vehicle's acceleration based on the driving data in the preprocessed scene data package; and determining the maximum value of the driver's head motion acceleration and the maximum change in the driver's pupil diameter relative to the reference pupil diameter based on the physiological data in the preprocessed scene data package. The absolute standard deviation of acceleration is used to characterize the degree of jerking of the vehicle in the driving scenario, the maximum value of head motion acceleration is used to characterize the degree of body sway of the driver in the driving scenario, and the maximum change in pupil diameter is used to characterize the driver's reaction intensity in the driving scenario. The vehicle comfort evaluation index is determined based on the absolute standard deviation of the vehicle's acceleration, the maximum value of the driver's head motion acceleration, and the maximum change in the driver's pupil diameter.

[0011] Optionally, based on the preprocessed scene data package, the vehicle trust evaluation index is determined, including: based on the physiological data in the preprocessed scene data package, determining the percentage of time the driver's gaze is away from the road the vehicle is traveling on, and the frequency with which the driver glances at the vehicle's rearview mirror per unit time, wherein the percentage of time is used to characterize the proportion of the time the driver's gaze is away from the road the vehicle is traveling on in the driving scenario to the total driving time of the driving scenario; based on the percentage of time and the frequency, the vehicle trust evaluation index is determined.

[0012] Optionally, the performance evaluation result of the assisted driving mode is determined based on multiple evaluation indicators, including: determining the weight coefficients corresponding to the multiple evaluation indicators, wherein the weight coefficients are used to characterize the importance of the evaluation indicators in the performance evaluation process of the assisted driving mode; calculating a weighted average of the multiple evaluation indicators based on the weight coefficients to obtain a weighted calculation result, wherein the weighted calculation result is used to characterize the driver's overall satisfaction with the assisted driving mode; and determining the performance evaluation result of the assisted driving mode based on the weighted calculation result.

[0013] Optionally, the vehicle performance evaluation method may further include: responding to the performance evaluation result indicating that the driver's satisfaction with the assisted driving mode is lower than a preset threshold, performing attribution analysis based on the driver's physiological data and the vehicle's driving data in the scenario data package to obtain attribution analysis results, wherein the attribution analysis results are used to characterize the reasons why the driver's satisfaction with the assisted driving mode is lower than the preset threshold; and generating an evaluation report of the vehicle's assisted driving mode based on the performance evaluation results and the attribution analysis results, wherein the evaluation report is used to provide strategies for improving the vehicle's assisted driving mode.

[0014] According to one aspect of the embodiments of this application, a vehicle performance evaluation device is provided. The device may include: a data acquisition unit, configured to acquire physiological data of the driver and driving data of the vehicle in response to the vehicle being in an assisted driving mode, wherein the physiological data characterizes the driver's physiological state and behavioral responses, and the driving data characterizes the driving operations and dynamic characteristics performed by the vehicle in the assisted driving mode; an alignment unit, configured to align the physiological data and driving data on a time axis based on a first timestamp corresponding to the physiological data and a second timestamp corresponding to the driving data, to obtain an alignment result, wherein the first timestamp identifies the time when the physiological data was generated, and the second timestamp identifies the time when the driving data was generated; a segmentation unit, configured to segment the alignment result based on scene identification information in the driving data, to obtain a scene data packet corresponding to at least one driving scenario, wherein the scene data packet includes physiological data and driving data corresponding to the driving scenario; a first determination unit, configured to determine multiple evaluation indicators of the vehicle based on the scene data packet, wherein the multiple evaluation indicators characterize the driver's satisfaction with the assisted driving mode in at least one driving scenario from multiple dimensions; and a second determination unit, configured to determine the performance evaluation result of the assisted driving mode based on the multiple evaluation indicators.

[0015] According to another aspect of the embodiments of this application, a processor is also provided. The processor is used to run a program, wherein the program executes the vehicle performance evaluation method of the embodiments of this application.

[0016] According to another aspect of the embodiments of this application, an electronic device is also provided, including: a memory storing an executable program; and a processor for running the program, wherein the program executes the vehicle performance evaluation method of various embodiments of this application when it runs.

[0017] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored program, wherein, when the program is run by a processor, it controls the device where the storage medium is located to execute the vehicle performance evaluation method of the embodiments of this application.

[0018] According to another aspect of the embodiments of this application, a computer program product is also provided. The program product includes computer instructions that, when executed by a processor, implement the vehicle performance evaluation method of the embodiments of this application.

[0019] According to another aspect of the embodiments of this application, a computer program product is also provided, including a non-volatile computer-readable storage medium for storing a computer program, which, when executed by a processor, implements the vehicle performance evaluation method in the embodiments of this application.

[0020] According to another aspect of the embodiments of this application, the embodiments of this application also provide a computer program that, when executed by a processor, implements the vehicle performance evaluation method described in the embodiments of this application.

[0021] According to another aspect of the embodiments of this application, a vehicle is also provided, including: a memory storing an executable program; and a processor for running the program, wherein the program implements the vehicle performance evaluation method of various embodiments of this application when it runs.

[0022] In this embodiment, when the vehicle is in assisted driving mode, the driver's physiological data and the vehicle's driving data can be acquired. Based on a first timestamp corresponding to the physiological data and a second timestamp corresponding to the driving data, the physiological data and driving data are aligned on a timeline. Then, based on scene identification information in the driving data, the aligned physiological data and driving data can be segmented to obtain at least one scene data package corresponding to a driving scene. Each scene data package includes the physiological data and driving data corresponding to that driving scene. After obtaining the scene data packages, multi-dimensional evaluation can be performed to obtain multiple evaluation indicators for the vehicle. Based on these multiple evaluation indicators, the performance evaluation result of the assisted driving mode can be determined. In other words, in this embodiment, by collecting the driver's physiological data and the vehicle's driving data, and synchronizing and segmenting the physiological data and driving data in time, the assisted driving performance of the vehicle is multi-dimensionally evaluated based on the driver's physiological data and the vehicle's driving data in each scene data package, resulting in multiple evaluation indicators. By using these multiple evaluation indicators to assess the assisted driving performance of vehicles, the accuracy of the assessment of vehicle assisted driving modes can be improved, thereby solving the technical problem of the accuracy of performance assessment of vehicle assisted driving modes. Attached Figure Description

[0023] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0024] Figure 1 This is a flowchart of a vehicle performance evaluation method according to an embodiment of this application;

[0025] Figure 2 This is a flowchart of a performance evaluation method for an intelligent driving system based on driver state monitoring, according to an embodiment of this application.

[0026] Figure 3 This is a schematic diagram of a vehicle performance evaluation device according to an embodiment of this application. Detailed Implementation

[0027] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present application.

[0028] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0029] According to an embodiment of this application, a method for evaluating the performance of a vehicle is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0030] Figure 1 This is a flowchart of a vehicle performance evaluation method according to an embodiment of this application, such as... Figure 1 As shown, the method may include the following steps.

[0031] Step S101: In response to the vehicle being in assisted driving mode, collect the driver's physiological data and the vehicle's driving data.

[0032] In the technical solution provided by step S101 of this application, the aforementioned assisted driving mode can be represented as a driving mode that enables the vehicle to autonomously complete driving tasks without direct driver operation. The aforementioned physiological data can be used to characterize the driver's physiological state and behavioral responses. For example, driver head movement data, eye movement information data, and facial expression data. The aforementioned driving data can be used to characterize the driving operations and dynamic characteristics performed by the vehicle in the assisted driving mode. For example, vehicle speed, acceleration, engine speed, steering angle, and braking pressure.

[0033] In this embodiment, in response to the vehicle being in assisted driving mode, the driver's physiological data and the vehicle's driving data are collected. For example, a Driver Monitoring System (DMS) can be used to collect the driver's physiological state and behavioral responses. The DMS collects data from in-vehicle cameras and biometric sensors to create a head model and an eye model of the driver. The head model may include the corners of the eyes, corners of the mouth, tip of the nose, and facial contours, while the eye model may include the corners of the eyes, eyelid contours, iris contours, and pupil center. The driver's head movement data can be obtained by detecting the motion state of the head model (e.g., pitch angle, yaw angle, and roll angle of the head model rotation). The driver's eye movement information data can be obtained by calculating the relative position of the pupil center and the corner of the eye in the eye model. After obtaining the driver's head movement data and eye movement information data, these data can be identified as the driver's physiological data.

[0034] Optionally, physiological data can be analyzed using a Facial Action Coding System (FACS) deployed in the DMS to decompose the driver's facial expressions into multiple Action Units (AUs), each corresponding to the movement of a facial muscle or group of facial muscles. By identifying the activation level of a specific AU, the driver's current facial expression data can be obtained. For example, when the corrugator supercilii muscle unit is highly activated and the zygomaticus major muscle unit is lowly activated, it indicates that the driver is frowning but not smiling. In this case, the driver's current facial expression can be determined to be "tense".

[0035] Optionally, after obtaining the driver's head movement data, eye movement information data, and facial expression data, the driver's head movement data, eye movement information data, and facial expression data can be identified as the driver's physiological data.

[0036] Optionally, when collecting physiological data of the vehicle's driver, a microsecond-level timestamp can be added to the physiological data, which can be used to record the exact time when the physiological data was generated.

[0037] Optionally, when collecting vehicle driving data, data sent by the vehicle's electronic control unit (ECU) can be read via the vehicle's Controller Area Network (CAN) bus. This includes data such as vehicle speed, acceleration, engine speed, steering angle, and braking pressure, thus obtaining the vehicle's driving data.

[0038] Optionally, when collecting vehicle driving data, a microsecond-level timestamp can be added to the driving data, which can be used to record the exact time when the driving data was generated.

[0039] In the above steps, by collecting the driver's physiological data and the vehicle's driving data, and by adding microsecond-level timestamps to the physiological data and driving data, a data foundation is provided for the accurate collection of subsequent physiological data and driving data.

[0040] Step S102: Based on the first timestamp corresponding to the physiological data and the second timestamp corresponding to the driving data, align the physiological data and the driving data on the time axis to obtain the alignment result.

[0041] In the technical solution provided by step S102 of this application, as can be seen from the description of step S101, the first timestamp can be used to identify the time when the physiological data was generated, and the second timestamp can be used to identify the time when the driving data was generated.

[0042] Optionally, after collecting physiological and driving data, the clocks of the camera and biometric sensor connected to the DMS, as well as the ECU connected to the CAN bus, can be synchronized to the same time reference source using the Precision Time Protocol (PTP). Then, based on the first timestamp of the physiological data and the second timestamp of the driving data, the physiological and driving data can be aligned on the time axis to obtain the alignment result.

[0043] In the above steps, physiological data and driving data are aligned on the timeline to ensure that the collected physiological and driving data are synchronized in time. That is, the alignment result not only contains information about the data itself, but more importantly, it presents the temporal correspondence between physiological and driving data.

[0044] Step S103: Based on the scene identification information in the driving data, the alignment results are divided to obtain scene data packets corresponding to at least one driving scene.

[0045] In the technical solution provided in step S103 of this application, the scene identification information can be represented as information used to distinguish different driving scenarios. For example, the scene identification information can be a scene label signal. The scene data packet can include the physiological data and driving data corresponding to the driving scenario.

[0046] In this embodiment, after obtaining the alignment result, the alignment result can be divided based on the scene identification information in the driving data to obtain scene data packets corresponding to at least one driving scene. Optionally, the vehicle's driving process can be divided into at least one driving scene based on the scene information identification in the driving data. By extracting the physiological data and driving data contained in each driving scene from the alignment result to form a data packet corresponding to each driving scene, scene data packets corresponding to at least one driving scene are obtained.

[0047] In the above steps, the performance of a vehicle's assisted driving capabilities may vary across different driving scenarios. For example, it may be very stable and smooth while cruising on a highway, but may encounter more challenges in urban congestion or complex intersections. By segmenting scenarios, the assisted driving performance of the vehicle in each specific scenario can be evaluated based on the scenario data package corresponding to each driving scenario, ensuring the relevance and effectiveness of the evaluation results. That is, based on scenario identification information, the alignment results are divided into at least one scenario data package, allowing the aligned physiological and driving data to be separated according to different driving scenarios, thus laying the foundation for subsequent accurate evaluation of the vehicle's assisted driving modes.

[0048] Step S104: Based on the scenario data package, determine multiple evaluation indicators for the vehicle.

[0049] In the technical solution provided by step S104 of this application, the above-mentioned multiple evaluation indicators can be used to characterize the driver's satisfaction with the above-mentioned assisted driving mode in at least one of the above-mentioned driving scenarios from multiple dimensions.

[0050] In this embodiment, after obtaining a scene data packet corresponding to at least one driving scenario, multiple evaluation indicators of the vehicle can be determined based on the scene data packet.

[0051] Optionally, by using preset calculation rules, physiological and driving data in at least one scenario data package are calculated to obtain different calculation results, thereby obtaining multiple evaluation indicators for the vehicle. For example, comfort indicators, trust indicators, and safety indicators.

[0052] In the above steps, multiple evaluation indicators of the vehicle are determined based on the scenario data package, realizing a multi-dimensional evaluation of the vehicle, thus laying the foundation for subsequent performance evaluation of the assisted driving mode.

[0053] Step S105: Determine the performance evaluation results of the assisted driving mode based on multiple evaluation indicators.

[0054] In the technical solution provided in step S105 of this application, after determining multiple evaluation indicators of the vehicle based on the scenario data packet corresponding to at least one driving scenario, the performance evaluation result of the vehicle's assisted driving mode in at least one driving scenario can be determined based on the multiple evaluation indicators corresponding to each driving scenario.

[0055] Optionally, by performing a weighted average calculation on the aforementioned comfort index, trust index, and safety index, a weighted average result can be obtained. Based on this weighted average result, the performance evaluation result of the vehicle's assisted driving mode in at least one driving scenario can be obtained. Then, by comprehensively considering the performance evaluation results of the vehicle in at least one driving scenario, the final performance evaluation result of the vehicle's assisted driving mode can be determined.

[0056] In steps S101 to S105 above, physiological data of the driver and driving data of the vehicle are collected, and the physiological and driving data are synchronized in time and segmented into scenarios. Then, based on the driver's physiological data and vehicle driving data in each scenario data package, the vehicle's assisted driving performance is evaluated in multiple dimensions, resulting in multiple evaluation indicators. By using these multiple evaluation indicators to assess the vehicle's assisted driving performance, the accuracy of the evaluation of the vehicle's assisted driving modes can be improved, thereby solving the technical problem of accurately evaluating the performance of vehicle assisted driving modes.

[0057] The method described in this embodiment will be further described below.

[0058] As an optional embodiment, step S102, based on the first timestamp corresponding to the physiological data and the second timestamp corresponding to the driving data, aligns the physiological data and the driving data on the time axis to obtain an alignment result, including: determining the physiological data and driving data corresponding to the same time based on the first timestamp corresponding to the physiological data and the second timestamp corresponding to the driving data; aligning the physiological data and driving data corresponding to the same time to obtain aligned physiological data and driving data; and determining the aligned physiological data and driving data as the alignment result.

[0059] In this embodiment, after collecting the driver's physiological data and the vehicle's driving data, the physiological data and driving data corresponding to the same time can be determined based on the first timestamp corresponding to the physiological data and the second timestamp corresponding to the driving data.

[0060] Optionally, after determining the physiological data and driving data corresponding to the same time, the physiological data and driving data corresponding to the same time can be aligned to obtain aligned physiological data and driving data.

[0061] Optionally, after aligning the physiological data and driving data at each time point, the aligned physiological data and driving data can be determined as the alignment result.

[0062] In the above steps, the physiological data and driving data are aligned at the same time according to the first timestamp corresponding to the physiological data and the second timestamp corresponding to the driving data, so as to obtain the alignment result, ensuring the correspondence between the physiological data and driving data in time, and providing a data basis for subsequent performance evaluation of the assisted driving system.

[0063] As an optional embodiment, step S103 involves dividing the alignment results based on the scene identification information in the driving data to obtain scene data packets corresponding to at least one driving scene, including: determining the start time and end time of at least one driving scene in the driving data based on the scene identification information in the driving data; and dividing the physiological data and driving data in the alignment results based on the start time and end time to obtain scene data packets corresponding to at least one driving scene.

[0064] In this embodiment, after collecting the vehicle's driving data, the start and end times of at least one driving scenario in the driving data can be determined based on the scene identification information in the driving data.

[0065] Optionally, a series of scenario triggering conditions are preset based on vehicle behavior, road characteristics, traffic conditions, and environmental factors. When a specific scenario triggering condition is detected, it indicates the entry into or termination of a specific scenario. Recording the time point at which a specific scenario triggering condition is detected yields the start and end times of at least one driving scenario. For example, when a vehicle is detected cutting in front (hereinafter referred to as a Cut-in), the vehicle switches from cruise mode to Cut-in mode, indicating that the vehicle has entered a Cut-in scenario. The time point at which the vehicle switches from cruise mode to Cut-in mode can be considered the start time point of entering the Cut-in scenario. While the vehicle is in a Cut-in scenario, when the distance to the vehicle in front is detected to be greater than a preset value, the vehicle can switch back from Cut-in mode to cruise mode, indicating that the vehicle has terminated the Cut-in scenario. The time point at which the vehicle switches back from Cut-in mode to cruise mode can be considered the end time point of the Cut-in scenario. Using this method, the start and end times of at least one driving scenario can be identified from the driving data.

[0066] Optionally, after determining the start and end times of at least one driving scenario in the driving data, the physiological data and driving data in the alignment results can be separated based on the start and end times to obtain a scenario data package corresponding to at least one driving scenario. For example, all aligned physiological data and driving data within the time period from the start time to the end time of the Cut-in scenario can be extracted to obtain the scenario data package corresponding to the Cut-in scenario.

[0067] In the above steps, the alignment results are divided according to the scene identification information, which can obtain at least one scene data package corresponding to the driving scene. The aligned physiological data and driving data are separated according to different driving scenes, which can lay the foundation for the accurate evaluation of the vehicle's assisted driving mode in the future.

[0068] As an optional implementation method, multiple evaluation indicators of a vehicle are determined based on scene data packets, including: preprocessing scene data packets corresponding to at least one driving scenario to obtain preprocessed scene data packets; and determining multiple evaluation indicators of the vehicle based on the preprocessed scene data packets.

[0069] In this embodiment, the preprocessing described above can be used to characterize the removal of invalid data from the scene data packet. The aforementioned evaluation metrics include at least a comfort evaluation metric and a trust evaluation metric. Specifically, the comfort evaluation metric can be used to characterize the driver's satisfaction with the comfort of the assisted driving mode, and the trust evaluation metric can be used to characterize the driver's satisfaction with the reliability of the assisted driving mode.

[0070] Optionally, after obtaining the scene data packets corresponding to at least one driving scenario, the scene data packets corresponding to at least one driving scenario can be preprocessed to obtain preprocessed scene data packets. For example, when the driver's facial feature point loss rate exceeds a preset loss rate due to a large head turn or strong light or glare, all physiological data within that time period are considered invalid data and are discarded.

[0071] Optionally, in a vehicle driving scenario, when the driver actively controls the vehicle, all driving data within the time period from the point when the driver actively controls the vehicle to the end time of the driving scenario can be marked as "driver takeover data" and then discarded. The indicator of the driver actively controlling the vehicle could be the driver pressing the brake or the driver turning the steering wheel.

[0072] Optionally, after obtaining the preprocessed scene data package, multiple evaluation indicators for the vehicle can be determined based on the preprocessed scene data package. For example, the preprocessed scene data package can be evaluated from multiple dimensions to obtain at least the vehicle's comfort evaluation indicator and trust evaluation indicator.

[0073] In this embodiment, by preprocessing the scene data packets corresponding to at least one driving scenario and removing invalid data, preprocessed scene data packets can be obtained, thereby avoiding erroneous evaluations of the performance of the assisted driving mode and laying the foundation for accurate evaluation of the performance of the assisted driving mode in the future.

[0074] As an optional embodiment, the vehicle comfort evaluation index is determined based on the preprocessed scene data package, including: determining the absolute standard deviation of the vehicle's acceleration based on the driving data in the preprocessed scene data package; and determining the maximum value of the driver's head motion acceleration and the maximum change in the driver's pupil diameter relative to the reference pupil diameter based on the physiological data in the preprocessed scene data package; and determining the vehicle comfort evaluation index based on the absolute standard deviation of the vehicle's acceleration, the maximum value of the driver's head motion acceleration, and the maximum change in the driver's pupil diameter.

[0075] In this embodiment, the absolute standard deviation of the acceleration can be used to characterize the degree of jerking of the vehicle in a driving scenario, the maximum value of the head motion acceleration can be used to characterize the degree of body sway of the driver in a driving scenario, and the maximum change in pupil diameter can be used to characterize the intensity of the driver's reaction in a driving scenario.

[0076] Optionally, after obtaining the preprocessed scene data package, the absolute standard deviation of the vehicle's acceleration can be determined based on the driving data in the preprocessed scene data package. Furthermore, the maximum value of the driver's head motion acceleration and the maximum change in the driver's pupil diameter relative to a reference pupil diameter can be determined based on the physiological data in the preprocessed scene data package. For example, an inertial measurement unit (IMU) deployed in the vehicle can acquire the vehicle's acceleration and angular velocity data in three-dimensional space. The IMU may include an accelerometer and a gyroscope. By acquiring data from the IMU via the CAN bus and recording it in the driving data, the absolute standard deviation of the vehicle's acceleration can be determined.

[0077] Optionally, the DMS can record the driver's head position information by collecting image data from cameras deployed in the vehicle, and calculate the linear and angular acceleration of the head based on this information. This allows the determination of the maximum value of the driver's head motion acceleration.

[0078] Optionally, the DMS can continuously record the driver's pupil diameter data by collecting image data from cameras deployed in the vehicle, and calculate the average value of the pupil diameter data to obtain a reference pupil diameter. Based on the continuously recorded reference pupil diameter, the maximum change in the driver's pupil diameter relative to the reference pupil diameter can be determined.

[0079] Optionally, after determining the absolute standard deviation of the vehicle's acceleration, the maximum value of the driver's head motion acceleration, and the maximum change in the driver's pupil diameter, a vehicle comfort evaluation index can be determined based on these parameters. For example, a normalization function can be used to normalize the absolute standard deviation of the vehicle's acceleration, the maximum value of the driver's head motion acceleration, and the maximum change in the driver's pupil diameter, obtaining normalized values. Weighted calculations can then be performed on these normalized values ​​to determine the vehicle comfort evaluation index.

[0080] For example, using a normalization function, the absolute standard deviation of the vehicle's acceleration, the maximum value of the driver's head motion acceleration, and the maximum change in the driver's pupil diameter can be mapped between 0 and 100. For instance, Normalize(x) = (x / x_max) 100. Where x can represent the absolute standard deviation of the vehicle's acceleration, the maximum value of the driver's head motion acceleration, or the maximum change in the driver's pupil diameter. x_max can represent the first preset maximum value of the absolute standard deviation of the vehicle's acceleration, the second preset maximum value of the driver's head motion acceleration, or the third preset maximum value of the maximum change in the driver's pupil diameter.

[0081] Optionally, after obtaining the normalized result, the result calculated using the normalization function can be weighted using the following formula to obtain the vehicle's comfort score. For example:

[0082] S_comfort=100-w1 Normalize(Jerk_std) -w2 Normalize(Head_motion_max) -w3 Normalize(Pupil_dilate)

[0083] Here, S_comfort can be represented as the vehicle's comfort score; Jerk_std can be represented as the absolute standard deviation of the vehicle's acceleration; Head_motion_max can be represented as the maximum value of the driver's head motion acceleration; Pupil_dilate can be represented as the maximum change in the driver's pupil diameter; w1 can be represented as the weighting coefficient of Jerk_std; w2 can be represented as the weighting coefficient of Head_motion_max; w3 can be represented as the weighting coefficient of Pupil_dilate, and the sum of w1, w2, and w3 is equal to 1.

[0084] In the above steps, based on the absolute standard deviation of the vehicle's acceleration, the maximum value of the driver's head movement acceleration, and the maximum change in the driver's pupil diameter, the vehicle's comfort evaluation index can be determined, thus providing data support for the comfort dimension for the subsequent performance evaluation of the assisted driving mode.

[0085] As an optional implementation method, the vehicle trust evaluation index is determined based on the preprocessed scene data packet, including: determining the percentage of time the driver's gaze is away from the road on which the vehicle is traveling, and the frequency of the driver scanning the vehicle's rearview mirror per unit time, based on the physiological data in the preprocessed scene data packet; and determining the vehicle trust evaluation index based on the percentage of time and the frequency.

[0086] In this example, the aforementioned time percentage can be used to characterize the proportion of the total driving time in a driving scenario where the driver's gaze is off the road the vehicle is traveling on.

[0087] Optionally, after obtaining the preprocessed scene data package, the physiological data in the preprocessed scene data package can be used to determine the percentage of time the driver's gaze is off the road and the frequency with which the driver glances at the vehicle's rearview mirror per unit time. For example, by collecting image data from cameras deployed in the vehicle, the DMS can record the driver's pupil center position. By analyzing the driver's pupil center position in real time, it can be determined whether the driver's gaze is on the road and whether the driver glances at the vehicle's rearview mirror.

[0088] Optionally, after determining the percentage of time the driver's gaze is off the road and the frequency with which the driver glances at the rearview mirror per unit time, a vehicle trustworthiness evaluation index can be determined based on the percentage of time and frequency. For example, by using a normalization function to normalize the percentage of time and frequency, and then weighting the normalized values, a vehicle trustworthiness evaluation index can be determined.

[0089] For example, a normalization function can be used to map the above duration percentages and frequencies to a range of 0-100. For instance, Normalize(x) = (x / x_max) 100. Where x can represent the percentage of time spent or the frequency. x_max can represent the fourth preset maximum value of the percentage of time spent or the fifth preset maximum value of the frequency.

[0090] Alternatively, the trust score of a vehicle can be obtained by weighting the results calculated using the above normalization function using the following formula.

[0091] S_trust=100-w4 Normalize(Gaze_off_road_ratio)-w5 Normalize(Check_mirror_freq)

[0092] Here, S_trust can be represented as the vehicle's trust score; Gaze_off_road_ratio can be represented as the duration percentage; Check_mirror_freq can be represented as the frequency; w4 can be represented as the weight coefficient of Gaze_off_road_ratio; w5 can be represented as the weight coefficient of Check_mirror_freq, and the sum of w4 and w5 is equal to 1.

[0093] In the above steps, based on the percentage of time the driver's gaze is off the road the vehicle is traveling on, and the frequency with which the driver scans the vehicle's rearview mirror per unit time, the vehicle's trust rating index can be determined, thus providing data support for the trust dimension for subsequent performance evaluation of the assisted driving mode.

[0094] As an optional embodiment, step S105, determining the performance evaluation result of the assisted driving mode based on multiple evaluation indicators, includes: determining the weight coefficients corresponding to each of the multiple evaluation indicators; performing a weighted average calculation on the multiple evaluation indicators based on the weight coefficients to obtain a weighted calculation result; and determining the performance evaluation result of the assisted driving mode based on the weighted calculation result.

[0095] In this embodiment, the aforementioned weighting coefficients can be used to characterize the importance of evaluation indicators in the performance evaluation of the assisted driving mode. The weighted calculation results can be used to characterize the driver's overall satisfaction with the assisted driving mode.

[0096] Optionally, after determining multiple evaluation indicators for the vehicle, weight coefficients can be determined for each indicator. For example, different weight coefficients can be assigned to the multiple evaluation indicators according to different evaluation modes. When conducting a basic evaluation of the performance of the assisted driving mode, α=0.4, β=0.3, γ=0.3; when conducting a comfort-specific evaluation of the assisted driving mode's performance, α=0.7, β=0.2, γ=0.1; when conducting a trust-specific evaluation of the assisted driving mode's performance, α=0.2, β=0.6, γ=0.2. Here, α can represent the weight coefficient of the comfort evaluation indicator; β can represent the weight coefficient of the trust evaluation indicator; γ can represent the weight coefficient of the safety evaluation indicator, and the sum of α, β, and γ equals 1.

[0097] Optionally, after determining the weight coefficients corresponding to multiple evaluation indicators, a weighted average can be calculated based on these weight coefficients to obtain the weighted calculation result. For example:

[0098] Overall_Score=α S_comfort+β S_trust+γ S_safety

[0099] Here, Overall_Score can be represented as the weighted calculation result; S_safety can be represented as the vehicle's safety score.

[0100] Optionally, after obtaining the weighted calculation results, the performance evaluation results of the assisted driving mode can be determined based on the weighted calculation results.

[0101] In the above steps, by weighting the vehicle's comfort score, trust score, and safety score, a comprehensive multi-dimensional evaluation index can be used to assess the performance of the vehicle's assisted driving mode, thereby achieving the technical effect of accurately evaluating the performance of the vehicle's assisted driving mode.

[0102] As an optional embodiment, the vehicle performance evaluation method may further include: responding to the performance evaluation result indicating that the driver's satisfaction with the assisted driving mode is lower than a preset threshold, performing attribution analysis based on the driver's physiological data in the scenario data package and the vehicle's driving data to obtain attribution analysis results; and generating an evaluation report of the vehicle's assisted driving mode based on the performance evaluation results and the attribution analysis results.

[0103] In this embodiment, the attribution analysis results described above can be used to characterize why the driver's satisfaction with the assisted driving mode is lower than a preset threshold. The evaluation report described above can be used to provide strategies for improving the vehicle's assisted driving mode.

[0104] Optionally, after determining the performance evaluation results of the assisted driving mode, if the performance evaluation results indicate that the driver's satisfaction with the assisted driving mode is lower than a preset threshold, attribution analysis can be performed based on the driver's physiological data and the vehicle's driving data in the scenario data package to obtain the attribution analysis results. For example, when it is detected that the driver's satisfaction with the assisted driving mode is lower than the preset threshold, the comfort score and / or trust score calculated above can be compared to obtain the low comfort score and / or trust score. Then, through the low comfort score and / or trust score, the physiological data and / or driving data causing the low score can be located. Finally, based on the physiological data and / or driving data causing the low score, the reason why the driver's satisfaction with the assisted driving mode is lower than the preset threshold can be determined, thereby obtaining the attribution analysis results.

[0105] Optionally, after obtaining the attribution analysis results, an evaluation report of the vehicle's assisted driving mode can be generated based on the performance evaluation results and the attribution analysis results. For example, the performance evaluation results and attribution analysis results can be used to generate an evaluation report of the vehicle's assisted driving mode in the form of a time series diagram. The evaluation report of the vehicle's assisted driving mode may include the aforementioned weighted calculation results for at least one driving scenario, comfort score, trust score, and physiological and / or driving data that caused the low score.

[0106] Optionally, uploading the test reports of the vehicle's assisted driving modes to the automaker's cloud big data platform can provide strategic support for the developers of the assisted driving modes when making subsequent improvements.

[0107] In this embodiment, physiological data of the driver and driving data of the vehicle are collected, and the physiological and driving data are synchronized in time and segmented into scenarios. Then, based on the driver's physiological data and vehicle driving data in each scenario data package, the vehicle's assisted driving performance is evaluated in multiple dimensions, resulting in multiple evaluation indicators. By using these multiple evaluation indicators to assess the vehicle's assisted driving performance, the accuracy of the evaluation of the vehicle's assisted driving modes can be improved, thereby solving the technical problem of accurately evaluating the performance of vehicle assisted driving modes.

[0108] The technical solutions of the embodiments of this application will be illustrated below with reference to preferred embodiments.

[0109] Users often experience poor performance when using assisted driving modes due to performance issues, necessitating accurate performance evaluation of these modes. However, performance evaluations often rely on vehicle operating data or driver subjective questionnaires. Vehicle operating data fails to reflect the user's true experience, and driver subjective questionnaires, due to time lag and memory bias, are difficult to provide impartial judgments. Therefore, there are technical issues with the low accuracy of performance evaluations for vehicle assisted driving modes.

[0110] To address the aforementioned technical problems, this application proposes a vehicle performance evaluation method. This method involves collecting physiological data of the driver and driving data of the vehicle, and synchronizing these data in time. Based on scene identification information in the driving data, the synchronized physiological and driving data are segmented to obtain scene data packets. Subsequently, the scene data packets are evaluated in multiple dimensions to obtain multiple evaluation indicators. By comprehensively considering multiple evaluation indicators, the final evaluation result can be determined, thereby effectively improving the accuracy of evaluating the vehicle's assisted driving mode and solving the technical problem of low accuracy in performance evaluation of vehicle assisted driving modes.

[0111] Figure 2 This is a flowchart illustrating a performance evaluation method for an intelligent driving system based on driver state monitoring, according to an embodiment of this application. Figure 2 As shown, the method may include the following steps.

[0112] Step S201: Obtain vehicle CAN bus data.

[0113] In this embodiment, data from the CAN bus in the vehicle is acquired, which may include the vehicle's speed, acceleration, and distance from other vehicles.

[0114] Step S202: Obtain intelligent driving system data.

[0115] In this embodiment, the intelligent driving system data may include DMS data. The DMS system can capture driver head movement data, eye movement information data, and facial expression information data, among others.

[0116] Step S203: Data time synchronization.

[0117] In this embodiment, all the data obtained in steps S201 and S202 are timestamped at the microsecond level based on a high-precision master clock, thereby achieving time synchronization of all the above data.

[0118] Alternatively, PTP can be used, or hardware synchronization between all acquisition devices in the CAN bus and DMS system can be achieved through hardware trigger signals, thereby realizing time synchronization of all the above data.

[0119] Step S204: Scene recognition and intelligent driving behavior analysis.

[0120] In this embodiment, driving scenarios are identified by monitoring scenario label signals, dividing the driving process into different scenario slices. For example, when the vehicle changes from cruise mode to cut-in mode, it is marked as the start of the cut-in scenario; when the distance to the vehicle cutting in front is detected to be greater than a preset value, the vehicle state switches back to cruise mode, and the cut-in scenario ends. The system automatically extracts all data within this time window from start to end, forming a "scenario data package". Within the "scenario data package", each time series data point of the DMS directly corresponds to the vehicle data (e.g., vehicle speed, acceleration, and distance to the cutting vehicle) at the same moment.

[0121] This embodiment also includes analyzing intelligent driving behavior. Intelligent driving behavior analysis can be used to remove invalid data from the "scene data package". Optionally, removing invalid data from the "scene data package" may include removing data from periods when DMS data is lost. For example, when the DMS causes a facial feature point loss rate >30% due to the driver turning their head sharply, obstruction (e.g., touching their face), or strong glare, all DMS data for that period is considered invalid and will not be included in the scoring calculation.

[0122] Optionally, removing invalid data from the "scenario data package" may also include removing data during periods when the system is not at fault. For example, during a driving scenario, if the driver actively intervenes (e.g., by applying the brakes or turning the steering wheel), all subsequent data from the moment of intervention is marked as "driver takeover," and this data will be removed from this system evaluation.

[0123] Optionally, removing invalid data from the "scenario data package" may also include removing data with physiological baseline differences. For example, to eliminate individual differences (e.g., some people naturally have larger pupils), the system will record the driver's various physiological baseline values ​​(e.g., average pupil diameter, average blink frequency, etc.) during a period of calm driving baseline (e.g., highway cruising) before the test begins. Subsequent analysis will focus more on the amount of change relative to the individual baseline, rather than the absolute value of the individual baseline.

[0124] Step S205, driver status quantification.

[0125] In this embodiment, the driver's head movement data, eye movement information data, and facial expression information data captured by the DMS system in step S202 are quantified to obtain the driver's state quantification index.

[0126] Optionally, driver state quantization can include head motion state quantization. Dozens of key points, such as the corners of the eyes, mouth, nose tip, and facial contours, are located on the face image using a neural network to obtain facial feature points. After obtaining the facial feature points, a 3D face model is built based on them, with the 3D coordinates corresponding to the 2D feature points labeled. The 2D feature points detected in the current frame are matched with a standard 3D face model. Using the Perspective-n-Point (PNP) algorithm, the rotation and translation vectors of the DMS camera relative to the face are calculated, describing the head's posture. The rotation vectors are converted into Euler angles, such as pitch (for nodding), yaw (for shaking), and roll (for tilting). After conversion to Euler angles, head motion state quantization indicators can be output. These indicators can include real-time posture and motion intensity. Motion intensity can be calculated by determining the angular velocity through the rate of change of Euler angles between adjacent frames, or by calculating the linear acceleration of the head through the translation vector.

[0127] Optionally, driver state quantification may also include eye movement information state quantification. By accurately locating eye feature points (e.g., eyelid contour, iris contour, and pupil center), the driver's gaze direction, pupil diameter, and eyelid opening can be calculated. The gaze direction can be obtained by calculating the relative position vector between the pupil center and the corner of the eye. For example, through a calibration process (having the driver look at several predetermined points on the screen), a mapping model from eye features to screen fixation points can be established. Applying this mapping model, real-time eye features can be converted into fixation point coordinates within the cockpit (e.g., in front of the road, on the central control screen, or in the rearview mirror), thereby obtaining the driver's gaze direction.

[0128] Optionally, the driver's pupil diameter can be calculated by precisely locating eye feature points.

[0129] Optionally, the driver's eyelid opening can be calculated by accurately locating eye feature points. For example, by measuring the pixel distance between the feature points of the upper and lower eyelids, and normalizing this distance with the driver's baseline values ​​when the eyes are wide open and completely closed, a ratio between 0 and 1 can be obtained, thus yielding the driver's eyelid opening.

[0130] Optionally, the aforementioned gaze direction, pupil diameter, and eyelid opening can be used as quantitative indicators of eye movement information status.

[0131] Optionally, driver state quantification may also include facial expression state quantification. Through facial motion coding systems and deep learning, facial muscle movements can be converted into probabilities of emotional states.

[0132] Optionally, facial expressions can be quantified based on action unit (AU) analysis. FACS decomposes facial expressions into 44 action units, each corresponding to the movement of a muscle or group of muscles. The algorithm analyzes the geometric changes of facial feature points to identify specific AUs (e.g., AU4 - corrugator supercilii, AU12 - zygomaticus major) and estimate their intensity (typically graded from 0 to 5). For example, "tension" might be manifested as a high intensity of AU4 (frowning) and a low intensity of AU12 (smiling).

[0133] Optionally, facial expressions are quantified using end-to-end classification based on deep learning. The cropped facial image region is directly input into a pre-trained convolutional neural network. The output layer of this network is a softmax classifier that outputs probability values ​​for various emotion categories (e.g., "neutral," "nervous," "surprised," "disgusted").

[0134] Optionally, after quantifying facial expressions as described above, facial expression quantification metrics can be output. These metrics may include action unit intensity and emotion probability.

[0135] Step S206, multimodal data correlation analysis.

[0136] In this embodiment, multimodal data correlation analysis is performed to determine the analysis results. The multimodal data may include "scene data packages" and quantitative indicators of driver state. Optionally, the multimodal data correlation analysis can be conducted from dimensions such as comfort and trust.

[0137] Alternatively, comfort scores can be calculated using the following formula for comfort analysis of multimodal data.

[0138] S_comfort=100-w1 Normalize(Jerk_std) -w2 Normalize(Head_motion_max) - w3

[0139] Normalize(Pupil_dilate)

[0140] Normalize(x) is a normalization function that maps the index x to a score between 0 and 100. For example: Normalize(x) = (x / x_max) 100, where x_max is the acceptable upper limit of the metric.

[0141] In the above formula, S_comfort can be represented as the comfort score; Jerk_std can be used to represent the absolute standard deviation of acceleration within the scene (which can be used to reflect the degree of vehicle jerking); Head_motion_max can be used to represent the maximum value of head motion acceleration within the scene (which can reflect body swaying); Pupil_dilate can be used to represent the maximum expansion of the pupil diameter relative to the baseline within the scene.

[0142] In the above formula, w1, w2, and w3 can be used to represent weighting coefficients, and w1 + w2 + w3 = 1. These weighting coefficients can be determined by expert scoring or the analytic hierarchy process, for example: w1 = 0.5, w2 = 0.3, w3 = 0.2.

[0143] Alternatively, for trust analysis of multimodal data, the trust score can be calculated using the following formula.

[0144] S_trust=100-w4 Normalize(Gaze_off_road_ratio)-w5 Normalize(Check_mirror_freq)

[0145] In the above formula, S_trust can be represented as the trust score; Gaze_off_road_ratio can be used to represent the total percentage of time that the view is away from the road in the scene; Check_mirror_freq can be used to represent the frequency of scanning the rearview mirror per unit time.

[0146] In the above formula, w4 and w5 can be used to represent weighting coefficients, and w4 + w5 = 1.

[0147] Optionally, the aforementioned comfort score and trust score can be used as the results of multimodal data association analysis.

[0148] Step S207: Score generation.

[0149] After determining the results of the multimodal data association analysis, a weighted average can be calculated on the scores in the multimodal data association analysis results to generate a performance score for the intelligent driving system.

[0150] Alternatively, a weighted average of the comfort score and the trust score can be calculated using the following formula.

[0151] Overall_Score=α S_comfort+β S_trust+γ S_safety

[0152] Here, Overall_Score can be used to represent the performance score; S_safety can be used to represent the safety score, and α, β, and γ are the weights of the scores in each dimension, with a sum of 1. These weights can be dynamically adjusted according to the purpose of the evaluation. For example: in basic evaluation: α=0.4, β=0.3, γ=0.3 (balanced focus); in comfort-specific evaluation: α=0.7, β=0.2, γ=0.1 (emphasis on comfort); in HMI trust evaluation: α=0.2, β=0.6, γ=0.2 (emphasis on trust).

[0153] Step S208, global problem analysis and optimization.

[0154] In this embodiment, after generating a performance score for the intelligent driving system, a global problem analysis and optimization can be performed on the system based on the low scores. Optionally, the low scores can be traced back to the specific algorithm module causing the problem, generating a diagnostic report that can directly guide development, and the intelligent driving system can be optimized based on the diagnostic report.

[0155] Step S209, Personalized learning and adaptation.

[0156] In this embodiment, after capturing the intelligent driving system data of different drivers through the DMS system, different performance scores for the intelligent driving system can be generated based on the different intelligent driving system data. The intelligent driving system can then perform personalized learning and adaptation for different drivers based on these different performance scores.

[0157] Step S210: Result presentation and feedback.

[0158] In this embodiment, the evaluation results of the intelligent driving system performance can be presented and fed back through an overall rating dashboard or time-series curves and associated views.

[0159] Optionally, a timeline can be used to simultaneously display: vehicle data curves (e.g., speed, acceleration); DMS indicator curves (e.g., pupil diameter, head movement); and subjective rating curves. Above the curves, "key events" identified by the system are marked, such as "maximum acceleration point" or "pupil dilation moment," and a description of the deduction item for that event is displayed.

[0160] Step S211: Upload data to the cloud platform.

[0161] In this embodiment, the in-vehicle system can package the evaluation results of each scenario into a structured data packet. This data packet may include: scenario identifier (ID); timestamp; vehicle identification number (VIN); DMS indicator sequence (e.g., time-series data such as pupil diameter and head movement); vehicle data sequence (e.g., acceleration and steering angle); correlation analysis results (e.g., a strong correlation between "braking acceleration -0.15g / s³" and "forward head acceleration 0.2g"); and scores for each dimension and the performance rating of the intelligent driving system.

[0162] Optionally, after obtaining the data packet, it can be encrypted and uploaded to the automaker's cloud big data platform when the vehicle is idle (e.g., when parked or charging) via a telematics box (T-Box) and cellular network (4G / 5G).

[0163] Optionally, after uploading the data package to the automaker's cloud-based big data platform, a cloud-based data lake and data visualization panel can be generated. The cloud-based data lake can incorporate all data, which R&D personnel can access through the visual data visualization panel. The data visualization panel can be used for: filtering by scenario (e.g., quickly viewing the average score of all "Cut-in" scenarios); locating low-scoring clusters (e.g., automatically identifying specific scenarios and parameter combinations with scores significantly below average); and drill-down analysis (e.g., clicking on any data point can simultaneously replay the event's exterior video, vehicle signal curves, and driver's DMS video clips).

[0164] Step S212, OTA update.

[0165] In this embodiment, researchers can optimize the intelligent driving system based on the problems reflected in the data display panel and update it using over-the-air (OTA) technology.

[0166] For example, step S212 may include the following steps: Step 1, identifying the problem: the global data display panel shows a low "cornering control comfort" score; Step 2, analyzing the problem: data playback reveals that overshoot of the lateral controller causes excessive swaying of the vehicle when exiting a corner, leading to driver head discomfort; Step 3, modifying the code: algorithm engineers modify the parameters of the proportional-integral-derivative controller (PID) or the weights of the linear quadratic regulator (LQR) for the lateral control; Step 4, simulation verification: in a simulation environment, the modified algorithm is tested using the DMS evaluation model to confirm the score improvement; Step 5, OTA release: the new algorithm is pushed to all vehicles as a global OTA upgrade; Step 6, effect verification: the fleet data is used to verify whether the DMS score for this scenario improves as expected after the upgrade.

[0167] Step S213: Adjust the parameters of the intelligent driving system.

[0168] In this embodiment, after receiving an OTA update, the intelligent driving system can adjust its parameters according to the OTA update content.

[0169] Step S214, to obtain a personalized driving experience.

[0170] In this embodiment, after adjusting the parameters of the intelligent driving system, a personalized driving experience can be obtained based on the adjusted parameters.

[0171] Optionally, after obtaining a personalized driving experience in step S214, you can return to steps S201 and S202 to start a new round of evaluation.

[0172] In steps S201 to S214 above, after acquiring vehicle CAN bus data and intelligent driving system data and synchronizing them in time, scene recognition, intelligent driving behavior analysis, and driver state quantification can be performed on the time-synchronized vehicle CAN bus data and intelligent driving system data to obtain multimodal correlation data. Analyzing the multimodal correlation data can generate a performance score for the intelligent driving system, thereby effectively meeting the purpose of accurately evaluating the performance of vehicle assisted driving modes and solving the technical problem of low accuracy in evaluating the performance of vehicle assisted driving modes.

[0173] According to an embodiment of this application, a vehicle performance evaluation device is also provided. It should be noted that the vehicle performance evaluation device can be used to perform a vehicle performance evaluation method according to the embodiment.

[0174] Figure 3This is a schematic diagram of a vehicle performance evaluation device according to an embodiment of this application. Figure 3 As shown, the vehicle performance evaluation device 300 may include: a data acquisition unit 301, an alignment unit 302, a division unit 303, a first determination unit 304, and a second determination unit 305.

[0175] The acquisition unit 301 is used to acquire the driver's physiological data and the vehicle's driving data in response to the vehicle being in the assisted driving mode. The physiological data is used to characterize the driver's physiological state and behavioral response, and the driving data is used to characterize the driving operations and dynamic characteristics performed by the vehicle in the assisted driving mode.

[0176] Alignment unit 302 is used to align physiological data and driving data on the time axis based on the first timestamp corresponding to the physiological data and the second timestamp corresponding to the driving data to obtain an alignment result. The first timestamp is used to identify the time when the physiological data was generated, and the second timestamp is used to identify the time when the driving data was generated.

[0177] The segmentation unit 303 is used to segment the alignment result based on the scene identification information in the driving data to obtain at least one scene data packet corresponding to the driving scene, wherein the scene data packet includes physiological data and driving data corresponding to the driving scene.

[0178] The first determining unit 304 is used to determine multiple evaluation indicators of the vehicle based on the scenario data package, wherein the multiple evaluation indicators are used to characterize the driver's satisfaction with the assisted driving mode in at least one driving scenario from multiple dimensions.

[0179] The second determining unit 305 is used to determine the performance evaluation results of the assisted driving mode based on multiple evaluation indicators.

[0180] Optionally, the alignment unit 302 may include: a first determining module, used to determine physiological data and driving data corresponding to the same time based on a first timestamp corresponding to physiological data and a second timestamp corresponding to driving data; a first alignment module, used to align the physiological data and driving data corresponding to the same time to obtain aligned physiological data and driving data; and a second determining module, used to determine the aligned physiological data and driving data as the alignment result.

[0181] Optionally, the segmentation unit 303 may include: a third determining module, used to determine the start time and end time of at least one driving scene in the driving data based on the scene identification information in the driving data; and a segmentation module, used to segment the physiological data and driving data in the alignment result based on the start time and end time to obtain a scene data packet corresponding to at least one driving scene.

[0182] Optionally, the first determining unit 304 may include: a preprocessing module, used to preprocess the scene data packets corresponding to at least one driving scenario to obtain preprocessed scene data packets, wherein the preprocessing is used to characterize the removal of invalid data in the scene data packets; and a fourth determining module, used to determine multiple evaluation indicators of the vehicle based on the preprocessed scene data packets, wherein the multiple evaluation indicators include at least a comfort evaluation indicator and a trust evaluation indicator, wherein the comfort evaluation indicator is used to characterize the driver's satisfaction with the comfort of the assisted driving mode, and the trust evaluation indicator is used to characterize the driver's satisfaction with the reliability of the assisted driving mode.

[0183] Optionally, the fourth determining module may include: a first determining submodule, used to determine the absolute standard deviation of the vehicle's acceleration based on the driving data in the preprocessed scene data package, and to determine the maximum value of the driver's head motion acceleration and the maximum change in the driver's pupil diameter relative to the reference pupil diameter based on the physiological data in the preprocessed scene data package, wherein the absolute standard deviation of acceleration is used to characterize the degree of vehicle jerking in the driving scenario, the maximum value of head motion acceleration is used to characterize the degree of body sway of the driver in the driving scenario, and the maximum change in pupil diameter is used to characterize the driver's reaction intensity in the driving scenario; and a second determining submodule, used to determine the vehicle's comfort evaluation index based on the absolute standard deviation of the vehicle's acceleration, the maximum value of the driver's head motion acceleration, and the maximum change in the driver's pupil diameter.

[0184] Optionally, the fourth determining module may include: a third determining submodule, used to determine the percentage of time the driver's gaze is away from the road the vehicle is traveling on, and the frequency with which the driver glances at the vehicle's rearview mirror per unit time, based on the physiological data in the preprocessed scene data package; wherein the percentage of time is used to characterize the proportion of the time the driver's gaze is away from the road the vehicle is traveling on in the driving scenario to the total driving time of the driving scenario; and a fourth determining submodule, used to determine the vehicle's trust rating index based on the percentage of time and the frequency.

[0185] Optionally, the second determining unit 305 may include: a fifth determining module, used to determine the weight coefficients corresponding to multiple evaluation indicators, wherein the weight coefficients are used to characterize the importance of the evaluation indicators in the performance evaluation of the assisted driving mode; a calculation module, used to perform a weighted average calculation on multiple evaluation indicators based on the weight coefficients to obtain a weighted calculation result, wherein the weighted calculation result is used to characterize the driver's overall satisfaction with the assisted driving mode; and a sixth determining module, used to determine the performance evaluation result of the assisted driving mode based on the weighted calculation result.

[0186] Optionally, the vehicle performance evaluation device 300 may further include: an analysis unit, configured to, in response to a performance evaluation result indicating that the driver's satisfaction with the assisted driving mode is lower than a preset threshold, perform attribution analysis based on the driver's physiological data in the scenario data package and the vehicle's driving data to obtain attribution analysis results, wherein the attribution analysis results are used to characterize the reasons why the driver's satisfaction with the assisted driving mode is lower than the preset threshold; and a generation unit, configured to, based on the performance evaluation results and the attribution analysis results, generate an evaluation report of the vehicle's assisted driving mode, wherein the evaluation report is used to provide strategies for improving the vehicle's assisted driving mode.

[0187] This device collects physiological data of the driver and driving data of the vehicle, and synchronizes and segments the physiological and driving data in time. Then, based on the driver's physiological data and vehicle driving data in each scenario data package, a multi-dimensional evaluation of the vehicle's assisted driving performance is performed, resulting in multiple evaluation indicators. By using these multiple evaluation indicators to assess the vehicle's assisted driving performance, the accuracy of evaluating the vehicle's assisted driving modes can be improved, thereby solving the technical problem of accurately evaluating the performance of vehicle assisted driving modes.

[0188] According to an embodiment of this application, a processor is also provided for running a program, wherein the program is executed by the processor to perform the vehicle performance evaluation method in the embodiment.

[0189] According to an embodiment of this application, an electronic device is also provided, including: a memory storing an executable program; and a processor for running the program, wherein the program executes the vehicle performance evaluation method in the embodiment during runtime.

[0190] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided. The computer-readable storage medium includes a stored program, wherein, when the program is executed, it controls the device where the computer-readable storage medium is located to perform the vehicle performance evaluation method of the embodiment.

[0191] According to an embodiment of this application, a computer program product is also provided, which includes a computer program, wherein when the computer program is executed by a processor, it implements the vehicle performance evaluation method in the embodiment.

[0192] According to an embodiment of this application, a computer program product is also provided, including a non-volatile computer-readable storage medium for storing a computer program. When the computer program is executed by a processor, it implements the vehicle performance evaluation method in the embodiment.

[0193] According to an embodiment of this application, a computer program is also provided, which, when executed by a processor, implements the vehicle performance evaluation method in the embodiment.

[0194] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0195] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0196] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between units or modules may be electrical or other forms.

[0197] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0198] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0199] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to related technologies, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0200] The above are merely preferred embodiments of this application. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for evaluating the performance of a vehicle, characterized in that, include: In response to the vehicle being in assisted driving mode, physiological data of the driver and driving data of the vehicle are collected, wherein the physiological data is used to characterize the driver's physiological state and behavioral response, and the driving data is used to characterize the driving operations and dynamic characteristics performed by the vehicle in the assisted driving mode. Based on the first timestamp corresponding to the physiological data and the second timestamp corresponding to the driving data, the physiological data and the driving data are aligned on the time axis to obtain an alignment result, wherein the first timestamp is used to identify the time when the physiological data was generated and the second timestamp is used to identify the time when the driving data was generated. Based on the scene identification information in the driving data, the alignment result is divided to obtain at least one scene data packet corresponding to the driving scene, wherein the scene data packet includes the physiological data corresponding to the driving scene and the driving data; Based on the scenario data package, multiple evaluation indicators for the vehicle are determined, wherein the multiple evaluation indicators are used to characterize the driver's satisfaction with the assisted driving mode in at least one driving scenario from multiple dimensions. The performance evaluation results of the assisted driving mode are determined based on multiple evaluation indicators.

2. The method according to claim 1, characterized in that, Based on the first timestamp corresponding to the physiological data and the second timestamp corresponding to the driving data, the physiological data and the driving data are aligned on the time axis to obtain the alignment result, including: Based on the first timestamp corresponding to the physiological data and the second timestamp corresponding to the driving data, the physiological data and the driving data corresponding to the same time are determined; Align the physiological data and driving data corresponding to the same time to obtain the aligned physiological data and driving data; The aligned physiological data and the driving data are determined as the alignment result.

3. The method according to claim 1, characterized in that, Based on the scene identification information in the driving data, the alignment result is divided to obtain at least one scene data packet corresponding to a driving scene, including: Based on the scene identification information in the driving data, determine the start time and end time of at least one driving scene in the driving data; Based on the start time point and the end time point, the physiological data and the driving data in the alignment result are divided to obtain at least one scene data packet corresponding to the driving scenario.

4. The method according to claim 3, characterized in that, Based on the scenario data package, multiple evaluation indicators for the vehicle are determined, including: Preprocessing is performed on at least one of the scene data packets corresponding to the driving scenario to obtain preprocessed scene data packets, wherein the preprocessing is used to characterize the removal of invalid data in the scene data packets; Based on the preprocessed scenario data packet, the vehicle's multiple evaluation indicators are determined. The multiple evaluation indicators include at least a comfort evaluation indicator and a trust evaluation indicator. The comfort evaluation indicator is used to characterize the driver's satisfaction with the comfort of the assisted driving mode, and the trust evaluation indicator is used to characterize the driver's satisfaction with the reliability of the assisted driving mode.

5. The method according to claim 4, characterized in that, Based on the preprocessed scene data packet, the comfort evaluation index of the vehicle is determined, including: Based on the driving data in the preprocessed scenario data package, the absolute standard deviation of the vehicle's acceleration is determined. Also, based on the physiological data in the preprocessed scenario data package, the maximum value of the driver's head motion acceleration and the maximum change in the driver's pupil diameter relative to a reference pupil diameter are determined. The absolute standard deviation of the acceleration characterizes the degree of jerking of the vehicle in the driving scenario, the maximum value of the head motion acceleration characterizes the degree of body sway of the driver in the driving scenario, and the maximum change in pupil diameter characterizes the intensity of the driver's reaction in the driving scenario. The comfort evaluation index of the vehicle is determined based on the absolute standard deviation of the vehicle's acceleration, the maximum value of the driver's head movement acceleration, and the maximum change in the driver's pupil diameter.

6. The method according to claim 4, characterized in that, Based on the preprocessed scene data packet, the trust evaluation index of the vehicle is determined, including: Based on the physiological data in the preprocessed scene data package, the percentage of time the driver's gaze is away from the road the vehicle is traveling on, and the frequency with which the driver glances at the vehicle's rearview mirror per unit time are determined. The percentage of time is used to characterize the proportion of the time the driver's gaze is away from the road the vehicle is traveling on in the driving scene to the total driving time of the driving scene. Based on the duration percentage and the frequency, the trust evaluation index of the vehicle is determined.

7. The method according to claim 1, characterized in that, Based on multiple evaluation metrics, the performance evaluation results of the assisted driving mode are determined, including: Determine the weight coefficients corresponding to the plurality of evaluation indicators, wherein the weight coefficients are used to characterize the importance of the evaluation indicators in the performance evaluation of the assisted driving mode; Based on the weighting coefficients, a weighted average is calculated on the multiple evaluation indicators to obtain a weighted calculation result, wherein the weighted calculation result is used to characterize the driver's overall satisfaction with the assisted driving mode; Based on the weighted calculation results, the performance evaluation results of the assisted driving mode are determined.

8. The method according to any one of claims 1 to 7, characterized in that, The method further includes: In response to the performance evaluation result indicating that the driver's satisfaction with the assisted driving mode is lower than a preset threshold, attribution analysis is performed based on the driver's physiological data in the scenario data package and the vehicle's driving data to obtain attribution analysis results, wherein the attribution analysis results are used to characterize the reasons why the driver's satisfaction with the assisted driving mode is lower than the preset threshold. Based on the performance evaluation results and the attribution analysis results, an evaluation report of the vehicle's assisted driving mode is generated, wherein the evaluation report is used to provide strategies for improving the vehicle's assisted driving mode.

9. A vehicle performance evaluation device, characterized in that, include: The data acquisition unit is used to acquire the driver's physiological data and the vehicle's driving data in response to the vehicle being in the assisted driving mode. The physiological data is used to characterize the driver's physiological state and behavioral response, and the driving data is used to characterize the driving operations and dynamic characteristics performed by the vehicle in the assisted driving mode. An alignment unit is used to align the physiological data and the driving data on a time axis based on a first timestamp corresponding to the physiological data and a second timestamp corresponding to the driving data, to obtain an alignment result, wherein the first timestamp is used to identify the time when the physiological data was generated and the second timestamp is used to identify the time when the driving data was generated. A segmentation unit is used to segment the alignment result based on the scene identification information in the driving data to obtain at least one scene data packet corresponding to a driving scene, wherein the scene data packet includes the physiological data corresponding to the driving scene and the driving data; The first determining unit is configured to determine multiple evaluation indicators of the vehicle based on the scenario data packet, wherein the multiple evaluation indicators are used to characterize the driver's satisfaction with the assisted driving mode in at least one driving scenario from multiple dimensions. The second determining unit is used to determine the performance evaluation result of the assisted driving mode based on multiple evaluation indicators.

10. An electronic device, characterized in that, include: Memory, which stores executable programs; A processor for running the program, wherein the program, when running, performs the method according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored executable program, wherein, when the executable program is executed, it controls the device on which the storage medium is located to perform the method according to any one of claims 1 to 8.

12. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method according to any one of claims 1 to 8.

13. A vehicle, characterized in that, include: Memory, which stores executable programs; A processor for running the program, wherein the program, when running, performs the method according to any one of claims 1 to 8.

Images