Method and related device for evaluating in-vehicle infotainment fluency

Abstract

This application discloses a method and related equipment for evaluating the smoothness of in-vehicle infotainment systems, relating to the field of vehicle testing technology. The method includes: acquiring a recorded video stream of the in-vehicle interaction process of a target vehicle; parsing and processing the recorded video stream to obtain temporal data of the visual features of the in-vehicle infotainment system during the interaction process; determining quantitative information on the smoothness of the target in-vehicle infotainment system based on the temporal data of the visual features; and generating a smoothness score characterizing the smoothness of the in-vehicle infotainment system based on touch response latency, animation frame dropping features, and stuttering features. This application, by parsing and extracting temporal data of the visual features from the recorded video stream of the in-vehicle interaction process, quantifying touch response latency, animation frame dropping features, and stuttering features, and generating a unified smoothness score, can achieve an objective, detailed, and comprehensive evaluation of the smoothness of the in-vehicle infotainment system.

Description

Technical Field

[0001] This application relates to the field of vehicle testing technology, and more specifically, to a method and related equipment for evaluating vehicle system smoothness. Background Technology

[0002] With the rapid development of intelligent cockpits and in-vehicle infotainment systems, vehicle infotainment systems have become an important carrier for human-machine interaction in vehicles. The smoothness of this interaction directly affects the user's operating experience and perception of the overall vehicle quality. Therefore, a scientific and accurate assessment of the smoothness of vehicle infotainment systems is of great significance for the design optimization, version iteration, and improvement of the overall vehicle experience.

[0003] In related technologies, the evaluation of in-vehicle infotainment system (IVS) smoothness typically relies on subjective human experience, simple operation tests, or single performance indicators. For example, it might involve manually assessing interface responsiveness and animation smoothness, or relying solely on indirect analysis based on system logs and hardware performance parameters. These evaluation methods are not only heavily influenced by the subjective differences of testers, making it difficult to guarantee the consistency and repeatability of evaluation results, but also often fail to accurately quantify key factors highly relevant to user experience. Therefore, existing evaluation methods generally lack a systematic analysis of actual interactive screens, making it difficult to comprehensively characterize the smoothness of the IVS system from the perspective of real-world interaction processes. This results in highly subjective evaluations of IVS smoothness, low quantification, and difficulty in establishing unified evaluation standards. In other words, related technologies present the technical challenge of accurately comparing the smoothness performance of IVS systems across different versions or vehicle models. Summary of the Invention

[0004] The summary section of this application introduces a series of simplified concepts, which will be further explained in detail in the detailed description section. The summary section of this application is not intended to limit the key features and essential technical features of the claimed technical solution, nor is it intended to determine the scope of protection of the claimed technical solution.

[0005] The vehicle infotainment system smoothness evaluation method and related equipment provided in this application can extract visual feature timing data by analyzing the video stream recorded during the vehicle infotainment system interaction process, quantify touch response latency, animation frame dropping and stuttering features, and generate a unified smoothness score, thereby achieving an objective, detailed and comprehensive evaluation of the smoothness of the vehicle infotainment system.

[0006] In a first aspect, this application provides a method for evaluating the smoothness of a vehicle infotainment system, comprising: acquiring a recorded video stream of the vehicle-to-vehicle interaction process of a target vehicle; parsing and processing the recorded video stream to obtain visual feature temporal data of the vehicle infotainment system of the target vehicle during the interaction process; determining quantifiable information of the smoothness of the target vehicle infotainment system based on the visual feature temporal data, wherein the quantifiable information of the smoothness includes touch response latency, animation frame dropping features, and stuttering features; and generating a smoothness score to characterize the smoothness of the vehicle infotainment system based on the touch response latency, the animation frame dropping features, and the stuttering features.

[0007] In some implementations, acquiring the recorded video stream of the vehicle-to-everything (V2X) interaction process for the target vehicle includes: controlling at least one camera device to record the V2X screen of the target vehicle based on preset ambient light compensation, preset shooting position, and preset device image stabilization and fixing processing to obtain an initial video stream, wherein the recording process covers multiple preset interaction scenarios; acquiring operation trigger signal timestamps and hardware response signal timestamps synchronized with the initial video stream; and binding the operation trigger signal timestamps and hardware response signal timestamps with the corresponding video frames in the initial video stream to obtain the recorded video stream.

[0008] In some embodiments, the step of parsing the recorded video stream to obtain the visual feature time-series data of the vehicle's infotainment system during the interaction process includes: performing frame-by-frame preprocessing on the recorded video stream to obtain a target video frame sequence, wherein the frame-by-frame preprocessing includes at least one of image denoising, adaptive image stabilization, color calibration, and frame rate normalization; performing detection and segmentation processing on the target video frame sequence to obtain an interface area image sequence of the vehicle's infotainment screen; extracting visual features from each interface area image in the interface area image sequence to obtain a feature representation of the infotainment interface in each target video frame; determining multiple preset event frames based on the feature representation, the operation trigger signal timestamp, and the hardware response signal timestamp, wherein the preset event frames include touch operation trigger frames, interface response frames, animation start frames, animation end frames, and stuttering exception frames; and constructing the visual feature time-series data based on the interface area image sequence and the timestamp of each preset event frame.

[0009] In some implementations, determining the fluency metrics of the target vehicle system based on the visual feature timing data includes: determining an operation trigger reference time point based on the touch operation trigger frame timestamp and operation trigger signal timestamp in the visual feature timing data; determining an interface feedback time point based on the interface response frame timestamp and hardware response signal timestamp in the visual feature timing data; determining initial delay data based on the time difference between the interface feedback time point and the operation trigger reference time point; and performing anomaly filtering on the initial delay data to obtain the touch response delay.

[0010] In some implementations, the animation frame dropping features include animation frame dropping rate and frame dropping time distribution features; determining the smoothness quantification information of the target vehicle system based on the visual feature time-series data includes: determining the animation playback period based on the animation start frame timestamp and animation end frame timestamp in the visual feature time-series data; determining the theoretical number of playback frames based on the preset animation frame rate and the animation playback period; identifying the actual display frames within the animation playback period from the visual feature time-series data, and counting the number of actual display frames to obtain the actual display frame count; determining the animation frame dropping rate based on the theoretical playback frame count and the actual display frame count; and parsing the timestamp distribution of each actual display frame within the total animation playback period to obtain the frame dropping time distribution features.

[0011] In some implementations, the stuttering features include the number of stuttering events and the duration of stuttering events; determining the smoothness quantification information of the target vehicle system based on the visual feature time-series data includes: obtaining the frame interval time and interface pixel change between adjacent frames in the interface area image sequence based on the visual feature time-series data; obtaining stuttering judgment thresholds corresponding to each interaction scene type in the interface area image sequence based on a preset mapping relationship, wherein the preset mapping relationship defines the correspondence between different interaction scene types and different stuttering judgment thresholds, and the interaction scene types include at least one of touch click scene, list swiping scene, and animation playback scene; determining the stuttering video frame sequence that meets the stuttering conditions based on a first comparison result of the frame interval time and the stuttering judgment threshold, and a second comparison result of the interface pixel change and the preset change threshold; and determining the number of stuttering events and the duration of stuttering events based on the stuttering video frame sequence.

[0012] In some implementations, generating a smoothness score characterizing the smoothness of the vehicle system based on the touch response latency, the animation frame drop features, and the stuttering features includes: standardizing the touch response latency, the animation frame drop features, and the stuttering features respectively to obtain a first standardized value, a second standardized value, and a third standardized value; determining a first weight coefficient corresponding to the first standardized value, a second weight coefficient corresponding to the second standardized value, and a third weight coefficient corresponding to the third standardized value according to a target weight allocation strategy, wherein the target weight allocation strategy is determined based on the vehicle model positioning and / or test target of the target vehicle; and performing a weighted calculation based on the first standardized value and the first weight coefficient, the second standardized value and the second weight coefficient, and the third standardized value and the third weight coefficient to obtain the smoothness score.

[0013] Secondly, this application also provides a vehicle infotainment system smoothness evaluation device, comprising: a video acquisition unit for acquiring a recorded video stream of the vehicle infotainment system interaction process of a target vehicle; a video parsing unit for parsing the recorded video stream to obtain visual feature time-series data of the vehicle infotainment system of the target vehicle during the interaction process; an information quantification unit for determining smoothness quantification information of the target vehicle infotainment system based on the visual feature time-series data, wherein the smoothness quantification information includes touch response latency, animation frame drop features, and stuttering features; and a smoothness scoring unit for generating a smoothness score characterizing the smoothness of the vehicle infotainment system based on the touch response latency, the animation frame drop features, and the stuttering features.

[0014] Thirdly, this application also provides an electronic device, including: a memory and a processor, wherein the processor is configured to implement the steps of the vehicle system smoothness evaluation method described in the first aspect when executing a computer program stored in the memory.

[0015] Fourthly, this application also provides a computer-readable storage medium storing computer-executable instructions or a computer program, wherein when the computer-executable instructions or the computer program are executed by a processor, the steps of the vehicle system smoothness evaluation method described in the first aspect are implemented.

[0016] Fifthly, this application also provides a computer program product, including a computer program or computer executable instructions, wherein when the computer program or computer executable instructions are executed by a processor, the steps of the vehicle system smoothness evaluation method provided in the embodiments of this application are implemented.

[0017] In summary, this application acquires and parses the recorded video stream of the vehicle-machine interaction process, directly extracting visual feature time-series data based on the actual interaction screen. Analyzing the process from an objectively recorded perspective avoids reliance on subjective human perception or simple experience, thus improving the objectivity and consistency of the vehicle-machine smoothness assessment results. The video stream parsing yields visual feature time-series data containing multiple preset event frame timestamps, enabling accurate location of key screen changes during vehicle-machine interaction at specific time points. This provides a reliable time reference for determining subsequent smoothness measurement information, enhancing the analysis... The method offers superior precision and repeatability. Based on temporal data of visual features, it simultaneously determines touch response latency, animation frame drops, and stuttering characteristics, reflecting the operational status of the vehicle system during interaction from multiple aspects, including interactive response, animation playback, and screen continuity. Compared to single-indicator evaluation, this method more comprehensively characterizes the smoothness level of the vehicle system. By generating a smoothness score by integrating touch response latency, animation frame drops, and stuttering characteristics, it unifies multiple different dimensions of smoothness quantification information into a single score to characterize the smoothness of the vehicle system, facilitating an intuitive expression and comprehensive evaluation of the overall smoothness of the vehicle system. In summary, the vehicle system smoothness evaluation method provided in this application extracts temporal data of visual features from the video stream recorded during the vehicle interaction process, quantifies touch response latency, animation frame drops, and stuttering characteristics, and generates a unified smoothness score, enabling an objective, precise, and comprehensive evaluation of the smoothness of the vehicle system. Attached Figure Description

[0018] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit this specification. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A flowchart illustrating a method for evaluating the smoothness of a vehicle infotainment system provided in an embodiment of this application; Figure 2 This is a schematic diagram of the composition structure of a vehicle system smoothness evaluation device provided in an embodiment of this application; Figure 3 This is a schematic diagram of the composition structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0019] The terms used in the specification, claims, and drawings of this application, such as "first," "second," "third," "fourth," etc. (if any), are used to distinguish similar objects and not to describe a specific order or sequence. Therefore, it is to be understood that these terms can be used interchangeably where appropriate, allowing the described embodiments to be used in different orders, unless specifically required by the illustrations or description. Furthermore, the terms "is" and "has," and any variations thereof, are intended to cover, non-exclusively, all possible constituent elements. For example, a process, method, system, product, or apparatus comprising several steps or units is not necessarily limited to the steps or units explicitly listed, but may also include other steps or units not explicitly listed, or steps or units inherent to the process, method, product, or apparatus.

[0020] In this application, a "module" or "unit" refers to a computer program or part of a computer program that has a specific function and works in conjunction with other related parts to achieve a predetermined goal. These modules or units can be implemented by software, hardware (e.g., processing circuitry or memory), or a combination of both. One or more processors or memories can implement one or more modules or units. Furthermore, each module or unit can also be part of a larger module or unit.

[0021] The technical solutions of this application will be described in detail below with reference to the accompanying drawings of the embodiments. It should be noted that the described embodiments are only a part of this application, and not all embodiments. In the following description, the "some embodiments" mentioned are only a subset of all possible embodiments, which may be the same or different subsets, and different embodiments can be combined with each other without conflict.

[0022] Figure 1 This is a schematic flowchart illustrating a method for evaluating the smoothness of a vehicle infotainment system provided in an embodiment of this application. For example, see [link to example]. Figure 1 The vehicle system smoothness evaluation method provided in this application embodiment may include the following steps 101 to 104: Step 101: Obtain a recorded video stream of the vehicle-to-machine interaction process for the target vehicle.

[0023] In some examples, the target vehicle is a specific vehicle requiring an evaluation of its infotainment system smoothness. This target vehicle is equipped with the infotainment system to be evaluated, which can be different models, configurations, or versions of the system software. For example, it can include various vehicle types such as compact family cars, mid-size sport utility vehicles (SUVs), and high-end luxury sedans, or even test vehicles of the same model equipped with different versions of the infotainment system. The infotainment system interaction process refers to the various operations performed by the user when using the target vehicle's infotainment system and the corresponding feedback from the system. This process directly reflects the system's responsiveness and smoothness, and can simulate real user scenarios to design and execute corresponding interactive operations. Examples include user touch-clicks on the infotainment screen, swiping to switch interfaces, triggering pop-up loading, watching animations, triggering multiple commands consecutively, and switching background processes. Recorded video streams are dynamic video data obtained by continuously capturing the vehicle-to-everything (V2X) interaction process of a target vehicle. Recorded video streams are used to fully capture the interface changes and operation scenarios during the V2X interaction process. The V2X interaction process can be recorded in real time by a camera device. For example, a high-definition camera can continuously capture the V2X screen and surrounding operation scenarios to form dynamic video data with a resolution of 1920×1080 pixels and a frame rate of 60 frames per second. This video data can clearly present the moment of operation triggering, the dynamic changes of the V2X interface, and the final feedback result.

[0024] For example, firstly, based on the testing requirements of this vehicle infotainment system smoothness assessment, a target vehicle can be selected from the test vehicle database to ensure that the vehicle's infotainment system is in normal working order. Then, testers can be arranged to perform various vehicle infotainment system interaction operations in sequence according to a pre-set list of user interaction scenarios, comprehensively covering typical operation scenarios in daily use. At the same time, pre-deployed camera equipment can be activated, and the shooting angle and parameters can be adjusted to ensure that the entire vehicle infotainment screen area is fully included in the shooting range. The entire vehicle infotainment system interaction process can be recorded continuously, and finally a recorded video stream containing complete interaction operations and interface feedback can be obtained to provide raw data support for subsequent vehicle infotainment system smoothness assessment.

[0025] By implementing step 101, the interface changes and interactive behaviors of the vehicle system in actual interactive scenarios can be recorded completely and intuitively, providing a real and reproducible raw data foundation for subsequent analysis. This avoids relying solely on subjective feelings for fluency evaluation and improves the objectivity of the evaluation process.

[0026] Step 102: Analyze the recorded video stream to obtain the visual feature time-series data of the vehicle's infotainment system during the interaction process.

[0027] In some examples, parsing processing involves a series of operations to extract, analyze, and transform features from the recorded video stream. This is used to extract relevant information about the vehicle-to-vehicle (V2V) interaction process from the raw video data. Computer data processing techniques can be employed to perform frame-by-frame analysis and feature extraction on the video stream. For example, video parsing algorithms can be used to perform frame decomposition, image feature recognition, and key information extraction on the recorded video stream, transforming dynamic video data into structured feature data. The V2V system is an intelligent system installed in the target vehicle to enable information interaction and function control between the user and the vehicle. It integrates hardware and software programs and is the core object for evaluating the smoothness of the V2V system. It can directly rely on the in-vehicle intelligent system already installed in the target vehicle; for example, in-vehicle infotainment systems developed based on the Android operating system (Android OS) or in-vehicle interactive systems customized based on the Linux operating system (Linux OS). These systems can respond to user touch, voice, and other operation commands and present corresponding interface feedback. Visual feature time-series data refers to feature information that reflects the dynamic changes of the interface during the interaction of the vehicle system. This feature information is linked to the time dimension to form a sequence of data, which contains timestamps of multiple preset event frames. It is key data for quantifying the smoothness of the vehicle system. By parsing and processing the recorded video stream, the interface change features can be extracted and the corresponding time information can be bound. For example, the position change features, color change features, and display status change features of vehicle system interface elements, as well as the timestamp data such as the touch operation trigger time and interface response time corresponding to these features, together constitute ordered visual feature time-series data.

[0028] For example, the recorded video stream obtained in step 101 can first be imported into the processing system, and the video stream can be decomposed frame by frame and preprocessed according to the preset parsing rules. The visual features of the vehicle interface in each frame of the video can be extracted by image recognition technology, and the timestamp information corresponding to each feature can be recorded to ensure that the feature data is accurately associated with the time dimension. Combined with the preset event frame judgment criteria, key feature data related to vehicle interaction can be screened out, and finally integrated to form complete visual feature time series data, providing structured data support for the subsequent calculation of smooth quantification information.

[0029] By implementing step 102, the recorded video stream is parsed and processed to extract the visual features of the vehicle system during the interaction process and form visual feature time series data. This allows the key screen changes during the vehicle-machine interaction process to be accurately located to specific time points, providing a reliable time reference for subsequent smoothness measurement analysis and improving the precision and consistency of the analysis results.

[0030] Step 103: Based on visual feature time-series data, determine the fluency quantification information of the target vehicle system.

[0031] Among them, the smoothness measurement information includes touch response latency, animation frame dropping characteristics, and stuttering characteristics.

[0032] In some examples, smoothness metrics are data sets that precisely characterize the smoothness performance of the vehicle infotainment system during interaction in numerical form. They objectively reflect the system's performance in aspects such as response speed and screen stability. Based on timestamp information and interface change characteristics in visual feature time-series data, data calculations are performed using preset quantification rules. Smoothness metrics can include touch response latency, animation frame drop characteristics, and stuttering characteristics. Touch response latency is the time interval between a user initiating a touch operation and the first response displayed on the vehicle infotainment system interface. It is an indicator of the system's responsiveness and can be calculated based on the event frame timestamps corresponding to the touch operation and the first interface response in the visual feature time-series data. For example, if a user clicks the navigation icon on the vehicle infotainment system, and the time interval from the click to the navigation interface starting to load is 80 milliseconds, this value represents the touch response latency for that operation. Frame dropping features in in-vehicle infotainment systems refer to the discrepancy between the actual number of video frames displayed and the theoretically expected number during animation playback. This is a crucial dimension reflecting the smoothness of animation playback. It can be determined by analyzing the video frame information corresponding to the animation playback period in visual feature time-series data and comparing the theoretical and actual frame counts. For example, if a vehicle's startup animation has a theoretical playback frame count of 300 frames but only 270 frames are actually displayed, the corresponding frame dropping data and distribution constitute the frame dropping feature of that animation. Lag features refer to the abnormal states such as pauses and delays that occur during dynamic interface changes in the in-vehicle infotainment system during interaction. This directly impacts the user experience. Lag-related information can be identified and extracted based on the interface changes and corresponding timestamps of adjacent frames in visual feature time-series data. For example, if a user scrolls through the in-vehicle infotainment system's music playlist and the interface experiences two noticeable pauses, each lasting 150 milliseconds and 200 milliseconds respectively, these pause counts and durations together constitute the lag feature.

[0033] For example, firstly, feature data and associated timestamps corresponding to interactive scenarios such as touch operation, animation playback, and interface swiping can be filtered from visual feature time series data; for different types of smoothness quantification information, the specific values ​​of touch response delay, animation frame loss related parameters, and stuttering feature data are calculated according to preset algorithms; the calculated quantification information is verified to ensure the validity and accuracy of the data, and finally a complete set of vehicle system smoothness quantification information is formed.

[0034] By implementing step 103, touch response latency, animation frame dropping features, and stuttering features are determined based on visual feature time-series data. This allows for a quantitative description of the vehicle system's operating status from multiple dimensions, such as interaction response speed, animation playback continuity, and screen stability, thereby providing a more comprehensive reflection of the vehicle system's smoothness during interaction.

[0035] Step 104: Based on touch response latency, animation frame dropping features, and stuttering features, a smoothness score is generated to characterize the smoothness of the vehicle infotainment system.

[0036] In some examples, the smoothness score is a single numerical indicator that comprehensively reflects the smoothness performance of the vehicle's infotainment system, obtained through standardized calculations. It quantifies the user experience during vehicle-to-vehicle interaction in an intuitive score format and can serve as a basis for comparing the smoothness of different versions or vehicle models. It can be calculated comprehensively based on quantitative information such as touch response latency, animation frame dropping features, and stuttering features, through a preset scoring model. For example, the smoothness score can be a specific score such as 85 or 92 points generated by a percentage scoring system, with a higher score indicating better smoothness performance of the vehicle's infotainment system.

[0037] The process of generating a smoothness score based on touch response latency, animation frame drop features, and stuttering features involves taking these three core types of smoothness quantification information as input, and transforming them into a single smoothness score through unified calculation rules and weight allocation logic. This achieves a comprehensive unification of multi-dimensional quantitative indicators. A complete scoring model, including data standardization, indicator weight allocation, and weighted calculation, can be pre-built. The three types of indicator data are then substituted into the model for calculation to obtain the score. For example, touch response latency, animation frame drop related parameters, and stuttering statistics can be standardized first, then weights can be allocated according to the degree of influence of each indicator on smoothness, and finally, a smoothness score is obtained through weighted summation.

[0038] By implementing step 104, a smoothness score is generated by integrating touch response latency, animation frame dropping features, and stuttering features. This unifies multi-dimensional smoothness measurement information into an intuitive scoring result, facilitating a comprehensive evaluation and comparative analysis of the overall smoothness of the vehicle system and improving the understandability and application value of the evaluation results.

[0039] In summary, this application embodiment acquires and parses the recorded video stream of the vehicle-machine interaction process, directly extracting visual feature time-series data based on the actual interaction screen. Analysis is performed from the objectively recorded interaction process, avoiding reliance solely on subjective human perception or simple experience, thus improving the objectivity and consistency of the vehicle-machine smoothness evaluation results. After parsing the video stream, visual feature time-series data containing multiple preset event frame timestamps are obtained, enabling accurate location of key screen changes during vehicle-machine interaction at specific time points. This provides a reliable time reference for determining subsequent smoothness measurement information, improving the accuracy of the evaluation. The method improves the precision and repeatability of the analysis; by simultaneously determining touch response latency, animation frame drop features, and stuttering features based on visual feature time-series data, it can reflect the operating status of the vehicle system during the interaction process from multiple aspects such as interaction response, animation playback, and screen continuity. Compared with single-index evaluation, it more comprehensively characterizes the smoothness level of the vehicle system; by generating a smoothness score by comprehensively analyzing touch response latency, animation frame drop features, and stuttering features, it unifies multiple different dimensions of smoothness measurement information into a single score result to characterize the smoothness of the vehicle system, facilitating an intuitive expression and comprehensive evaluation of the overall smoothness of the vehicle system. In summary, the vehicle system smoothness evaluation method provided in this application, by analyzing and extracting visual feature time-series data from the video stream recorded during the vehicle interaction process, quantifies touch response latency, animation frame drop features, and stuttering features, and generates a unified smoothness score, can achieve an objective, precise, and comprehensive evaluation of the smoothness of the vehicle system.

[0040] In some embodiments, step 101 may include: controlling at least one camera device to record the vehicle screen of the target vehicle based on preset ambient light compensation, preset shooting position and preset device anti-shake fixing processing to obtain an initial video stream, wherein the recording process covers multiple preset interactive scenarios; obtaining operation trigger signal timestamps and hardware response signal timestamps synchronized with the initial video stream; binding the operation trigger signal timestamps and hardware response signal timestamps with the corresponding video frames in the initial video stream to obtain a recorded video stream.

[0041] In some examples, preset ambient light compensation is used to eliminate the impact of fluctuations in light intensity in the recording environment on video quality. Pre-set light adjustment strategies and operations can be configured based on the lighting standards of the vehicle infotainment system test scenario, configuring ambient light compensation devices and setting compensation parameters. For example, preset ambient light compensation may include deploying dimmable LED (Light Emitting Diode) fill lights in the test area, with a preset light intensity range of 500-800 lux. When the ambient light is below 500 lux, the fill light is automatically activated; when it is above 800 lux, the fill light intensity is reduced to ensure the vehicle infotainment screen display is clear, without reflections or shadows. Pre-set shooting position is used to fully capture the vehicle infotainment screen interface and interactive operations. Pre-determined camera position and angle parameters can be verified through testing to determine the shooting angle that fully covers the vehicle infotainment screen without distorted viewing angles. For example, preset shooting position may include placing the camera device 30-50 cm directly in front of the vehicle infotainment screen, with the lens at a 90-degree angle to the screen plane, while ensuring the center of the lens and the center of the screen are on the same horizontal line to avoid image stretching or obstruction. Pre-set equipment stabilization is a pre-implemented measure to prevent video blurring caused by slight vibrations of the camera during recording. It involves securing the equipment and enhancing stabilization. This can be achieved by selecting a stabilization device and installing and adjusting it according to pre-defined specifications. For example, pre-set equipment stabilization may include fixing the camera on a professional tripod with anti-slip rubber pads on the bottom, and equipping the camera with an Optical Image Stabilization (OIS) module, providing double protection for the stability of the equipment during recording.

[0042] The camera equipment is the hardware used to record the vehicle-to-everything (V2X) interaction process. Its quantity can be flexibly configured according to testing needs. Camera equipment meeting parameter standards can be selected and deployed based on the recording coverage and accuracy requirements. For example, it can include a single 4K ultra-high-definition camera (3840×2160 pixels resolution, 60 frames per second), or two cameras of the same specifications recording simultaneously from the front and side of the vehicle screen to ensure no interactive scenes are missed. The initial video stream is the raw dynamic video data directly generated by the camera equipment after capturing the V2X interaction process under preset recording conditions. It can control the camera equipment to start recording according to preset parameters, capturing the vehicle screen and surrounding interactive scenes. For example, it can include 10 minutes of raw video data recorded by a single camera equipment, encoded in High Efficiency Video Coding (HEVC). This initial video stream completely records the interface changes corresponding to various interactive operations.

[0043] Preset interaction scenarios are a collection of typical and peripheral operation scenarios that users encounter when using the vehicle's infotainment system in their daily lives. These scenarios can be selected and defined based on user habit surveys and the vehicle's function list. For example, preset interaction scenarios may include touching and clicking the vehicle's navigation icon, swiping to switch multimedia playback interfaces, triggering a Bluetooth connection pop-up, repeatedly clicking the air conditioning temperature adjustment button, and switching between music and navigation applications in the background, comprehensively covering the interaction process of the vehicle's main functions.

[0044] The operation trigger signal timestamp and hardware response signal timestamp, synchronized with the initial video stream, are time stamps that record the moment the user operation is initiated and the moment the vehicle's hardware responds. These timestamps are consistent with the recording timeline of the initial video stream. Operation trigger signals and hardware response signals can be synchronously captured by signal acquisition equipment, and corresponding timestamps can be generated using time synchronization technology. The process of binding the operation trigger signal timestamp and hardware response signal timestamp with the corresponding video frames in the initial video stream to obtain the recorded video stream involves establishing a correlation between timestamps and video frames using a time synchronization algorithm. This ensures that each frame in the initial video stream corresponds to the relevant operation trigger and hardware response time information. Timeline alignment technology can be used to match and associate the two types of timestamps with the frame times of the initial video stream. For example, video processing software can identify the frame (frame number 1200) in the initial video stream where the user clicks the screen, embedding the operation trigger signal timestamp and hardware response signal timestamp into the metadata of that frame and subsequent associated frames, forming a complete recorded video stream with timestamp associations.

[0045] By implementing the above embodiments, the vehicle screen is recorded under preset ambient light, shooting angle, and fixed device stabilization conditions. The operation trigger signal timestamp and hardware response signal timestamp are synchronously bound to the video frame. This can effectively reduce the impact of external environmental differences on the video acquisition results, while ensuring that the video data acquired under different test vehicles and different test batches have a consistent time base. This provides a stable and unified data foundation for the subsequent comparative analysis of smoothness indicators and improves the comparability between test results of different models or versions.

[0046] In some embodiments, step 102 may include: performing frame-by-frame preprocessing on the recorded video stream to obtain a target video frame sequence, wherein the frame-by-frame preprocessing may include at least one of image denoising, adaptive image stabilization, color calibration, and frame rate normalization; performing detection and segmentation processing on the target video frame sequence to obtain an interface area image sequence of the vehicle's infotainment screen; extracting visual features from each interface area image in the interface area image sequence to obtain a feature representation of the vehicle's infotainment interface in each target video frame; determining multiple preset event frames based on the feature representation, operation trigger signal timestamps, and hardware response signal timestamps, wherein the preset event frames may include touch operation trigger frames, interface response frames, animation start frames, animation end frames, and stuttering exception frames; and constructing visual feature time-series data based on the interface area image sequence and the timestamps of each preset event frame.

[0047] In some examples, frame-by-frame preprocessing is a set of operations that sequentially optimize each video frame in the recorded video stream to improve the image quality, eliminate interference, and provide a reliable image foundation for subsequent analysis. Image processing algorithms can be used to perform a series of optimization operations on the video stream frame by frame. Frame-by-frame preprocessing may include at least one of image denoising, adaptive image stabilization, color calibration, and frame rate normalization. Image denoising is the process of eliminating noise, graininess, and other interference information in video frames caused by shooting environment or equipment factors. Image denoising algorithms can be used to process single frames; for example, a Gaussian filtering algorithm can be used to smooth video frames with slight graininess, remove meaningless interference pixels, and retain the effective information of the vehicle interface. Adaptive image stabilization automatically adjusts the image position and compensates for slight jitter in video frames. It uses image registration algorithms to analyze the positional offset of adjacent frames and automatically corrects the image. For example, when a horizontal jitter offset of 5 pixels is detected in consecutive video frames, the algorithm automatically performs reverse offset compensation on subsequent frames to ensure stable interface element positions. Color calibration adjusts the color parameters of video frames to match the actual colors displayed on the vehicle's infotainment screen. Based on a standard color template, it uses color correction algorithms to adjust parameters such as brightness, contrast, and color temperature. For example, it can adjust the brightness of a darker video frame from 50 to the standard value of 80, and calibrate the color temperature of a cooler frame from 5000K to 6500K, restoring the true colors of the vehicle's infotainment screen. Frame rate normalization is a process that adjusts the actual frame rate of a video stream to a preset standard frame rate. Frame rate conversion algorithms can be used to perform frame interpolation or frame extraction on the video stream. For example, a video stream with an actual frame rate of 50 frames per second can be normalized to a standard frame rate of 60 frames per second by adding frame data through interpolation, ensuring the consistency of the temporal dimension in subsequent analysis. The target video frame sequence is a structured image set formed by preprocessing each frame and arranging them in chronological order. It retains complete image information of the vehicle-machine interaction process and optimizes its quality. It can be obtained by preprocessing each frame of a recorded video stream and then organizing it according to the original video stream's chronological order. For example, a 10-minute recorded video stream with a frame rate of 60 frames per second can be processed frame by frame to form a target video frame sequence of 36,000 frames ordered chronologically.

[0048] Detection and segmentation processing involves identifying the effective area of ​​the vehicle's infotainment screen within a video frame and extracting that area from the image to filter out background interference and focus on the core interactive interface. Image detection and segmentation algorithms can be used to process the target video frame sequence. For example, object detection algorithms can be used to locate the boundary coordinates of the vehicle's infotainment screen, and then image segmentation algorithms can be used to separate the screen area from the background, which includes interior trim and dashboard reflections. The interface area image sequence is a collection of images arranged in chronological order, containing only the effective area of ​​the vehicle's infotainment screen after detection and segmentation processing. It can be obtained by performing detection and segmentation processing on the target video frame sequence and then reorganizing it in its original chronological order. For example, by extracting the vehicle's infotainment screen area images one by one from a target video frame sequence of 36,000 frames, a structured image sequence containing only the interface content can be formed.

[0049] Visual feature extraction is the process of extracting key information that characterizes the state and element changes of the vehicle's infotainment interface from an image of the interface region. Image feature extraction algorithms can be used to analyze a single frame of the interface region image; for example, edge detection algorithms can extract the contour features of interface elements, and texture feature algorithms can extract the surface texture information of components such as buttons and icons. The feature representation of the vehicle's infotainment interface in each target video frame is obtained through visual feature extraction. This provides quantifiable feature data describing the state of a single frame of the vehicle's infotainment interface. This data can be in the form of structured data such as vectors or matrices. After performing visual feature extraction on each frame of the interface region image, the extracted feature information can be transformed into a standardized data form. For example, the contour features, icon position features, and text display features of a navigation interface frame can be integrated into a 128-dimensional feature vector as the feature representation of that frame's interface.

[0050] Multiple preset event frames are predefined, specific video frames that reflect key nodes in vehicle-machine interaction. These frames correspond to important state changes during the interaction process and can be selected from the image sequence of the interface area based on feature representation and relevant timestamps using an event recognition algorithm. The process of determining multiple preset event frames based on feature representation, operation trigger signal timestamps, and hardware response signal timestamps involves combining the feature data of a single-frame interface, the time stamp of the operation initiation, and the time stamp of the hardware response. An algorithm identifies and filters the video frames corresponding to key interaction nodes, allowing the construction of an event frame recognition model. Input feature representation and two types of timestamp data are used for matching and judgment. For example, matching the trigger signal timestamp of a certain operation with the "screen touch point appearance" feature in the interface feature representation locates the touch operation trigger frame corresponding to that operation. A touch operation trigger frame is a video frame that records the moment a user initiates a touch operation. This frame contains the interface state at the time the touch operation occurs and can be determined by matching the timestamp of the operation trigger signal with features such as "touch point appearance" and "initial state before interface unresponsiveness" in the interface characteristics. For example, when a user clicks the music play button, the video frame that records the first appearance of the touch point at the button position is the touch operation trigger frame for that operation.

[0051] The interface response frame is a video frame that records the first time the vehicle's infotainment system responds to a user's action. This frame marks the beginning of the interface's responsiveness to the action and can be determined by comparing changes in interface features after the touch operation trigger frame, combined with the hardware response signal timestamp. For example, the video frame in which the navigation loading animation first appears after the user clicks the navigation icon is the interface response frame for that operation. The animation start frame is a video frame that records the beginning of the animation playback on the vehicle's infotainment system. This frame marks the starting point of the animation playback process and can be determined by identifying features such as "the first appearance of animated elements" and "the static interface beginning to change dynamically." For example, the first frame of the welcome animation when the vehicle starts up is the animation start frame. The animation end frame is a video frame that records the completion of the animation playback and the return of the interface to a stable state. This frame marks the end point of the animation playback process and can be determined by identifying features such as "animated elements stopping changing" and "the interface entering a stable display state." For example, the video frame in which the main menu is displayed after the welcome animation finishes playing is the animation end frame. A stuttering frame is a video frame that records a pause or abnormal delay in the dynamic changes of the vehicle's infotainment system interface. This frame reflects a problem with the smoothness of the interface. By analyzing the characteristic changes and time intervals of adjacent frames, video frames with abnormally small changes and intervals exceeding the normal range can be identified. For example, when a user scrolls through a list, if the positions of interface elements do not change significantly for three consecutive frames and the frame intervals exceed the normal threshold, these three frames are stuttering frames.

[0052] The process of constructing visual feature time-series data based on the sequence of interface area images and the timestamps of each preset event frame involves associating the time-sorted sequence of interface area images with the timestamps corresponding to each preset event frame, integrating them into a structured data set containing both time and visual feature dimensions. Data association algorithms can be used to bind the feature representations of the interface area images, the identifiers of the preset event frames, and the corresponding timestamps, organizing and storing them in chronological order. For example, associating the feature vector of each frame in the interface area image sequence, the identifier of whether it is a preset event frame, with the timestamp corresponding to that frame forms a feature data chain arranged along the time axis, which is the visual feature time-series data.

[0053] Through the implementation of the above embodiments, the recorded video stream is preprocessed frame by frame, the interface area is detected and segmented, and visual features are extracted. Combined with operation trigger signals and hardware response signals, multiple types of preset event frames are identified to construct structured visual feature time-series data. This enables key events in the vehicle-machine interaction process to be accurately located on a unified timeline, providing a fine and consistent analytical basis for the smoothness performance of different vehicles and different versions under the same interaction events, and avoiding evaluation deviations caused by inconsistent event recognition.

[0054] In some embodiments, step 103 may include: determining an operation trigger reference time point based on the touch operation trigger frame timestamp and operation trigger signal timestamp in the visual feature timing data; determining an interface feedback time point based on the interface response frame timestamp and hardware response signal timestamp in the visual feature timing data; determining initial delay data based on the time difference between the interface feedback time point and the operation trigger reference time point; and performing anomaly filtering on the initial delay data to obtain the touch response delay.

[0055] In some examples, the touch operation trigger frame timestamp is a timestamp corresponding to the touch operation trigger frame in the visual feature time-series data, precisely recording the instant the user initiates the touch operation. The operation trigger signal timestamp is synchronously bound to the recorded video stream; it is the timestamp corresponding to the operation trigger signal generated when the user initiates the touch operation, reflecting the precise moment the operation command was issued. The operation trigger reference time point is the starting reference time used to calculate the touch response delay. It is determined based on both the touch operation trigger frame timestamp and the operation trigger signal timestamp, ensuring the accuracy of the delay calculation starting point. A unified delay calculation starting time can be determined by fusing the two types of timestamps using preset rules; for example, the average of the two types of timestamps can be used as the operation trigger reference time point.

[0056] The interface response frame timestamp is the time stamp corresponding to the interface response frame in the visual feature time-series data. This timestamp records the moment when the vehicle's infotainment interface first provides visual feedback to a touch operation. The hardware response signal timestamp is synchronously bound to the recorded video stream and is the time stamp generated by the vehicle's hardware after receiving the operation command. The interface feedback time point is the termination reference time for calculating the touch response delay. It can be determined based on the interface response frame timestamp and the hardware response signal timestamp, comprehensively reflecting the actual effective moment of the interface feedback. The two types of timestamps can be integrated and processed using preset rules to determine the termination time of the delay calculation. For example, the later of the interface response frame timestamp and the hardware response signal timestamp can be selected as the interface feedback time point. That is, when the interface response frame timestamp is "1698765432887.5 milliseconds" and the hardware response signal timestamp is "1698765432827.5 milliseconds", the larger value is used. The timeframe of "1698765432887.5 milliseconds" is used as the interface feedback time point to ensure that a complete feedback node is only considered when the hardware completes its response and the interface displays visual feedback. This avoids delay calculation deviations caused by the hardware responding early but the interface not responding synchronously, or, in another case, avoids delay data distortion caused by false feedback frames appearing on the interface before the hardware has completed an effective response. This ensures that the calculation of touch response delay can truly reflect the complete process of the hardware receiving the instruction, completing the response, and the interface displaying effective feedback, which is consistent with the core experience logic of the user's actual perception from initiating an operation to seeing effective feedback on the interface.

[0057] Initial latency data is raw touch response latency data obtained directly from the time difference without data verification. It reflects the initial time interval from operation initiation to interface feedback, and the calculated result of the time difference can be directly used as the initial latency data. Anomaly filtering is a process that filters and removes abnormal values ​​in the initial latency data caused by non-realistic factors such as occasional system failures or extreme environmental interference, ensuring the authenticity and reliability of the touch response latency data. Pre-set anomaly judgment rules and filtering algorithms can be used to verify the initial latency data; for example, a normal latency range of 30-200 milliseconds can be set using statistical methods, and initial latency data exceeding this range will be judged as abnormal and removed. The process of obtaining the touch response latency by anomaly filtering of the initial latency data involves removing outliers from the initial latency data, retaining the effective data that truly reflects the vehicle's touch response performance. The average of multiple initial latency data after anomaly filtering can be used as the final touch response latency.

[0058] By implementing the above embodiments, the touch response delay is calculated based on the timestamp relationship between the touch operation trigger frame, hardware response signal and interface response frame, and the initial delay data is filtered for anomalies. This effectively eliminates the influence of occasional interference on the delay result, making the key indicator of touch response delay more realistic and stable. It can form repeatable response performance comparison results between different car models or different car system versions.

[0059] In some embodiments, the aforementioned animation frame dropping features may include animation frame dropping rate and frame dropping time distribution features; the aforementioned step 103 may include: determining the animation playback period based on the animation start frame timestamp and animation end frame timestamp in the visual feature time series data; determining the theoretical number of playback frames based on the preset animation frame rate and animation playback period; identifying the actual display frames within the animation playback period from the visual feature time series data, and counting the number of actual display frames to obtain the actual display frame count; determining the animation frame dropping rate based on the theoretical playback frame count and the actual display frame count; and parsing the timestamp distribution of each actual display frame within the total animation playback period to obtain the frame dropping time distribution features.

[0060] In some examples, the animation start frame timestamp is the timestamp corresponding to the animation start frame in the visual feature time series data, accurately recording the instant the in-vehicle animation begins playback; for example, the start frame timestamp of a certain in-vehicle navigation startup animation is "1698765450100 milliseconds," which is the recorded start time of the animation playback. The animation end frame timestamp is the timestamp corresponding to the animation end frame in the visual feature time series data, recording the instant the in-vehicle animation finishes playback and the interface enters a stable state; for example, the end frame timestamp of the aforementioned navigation startup animation is "1698765455100 milliseconds," marking the official end of the animation playback. The animation playback period is the complete time interval from the start to the end of the animation, reflecting the total duration of the animation playback, and can be calculated by subtracting the animation start frame timestamp from the animation end frame timestamp; for example, subtracting the start frame timestamp "1698765455100 milliseconds" from the animation end frame timestamp "1698765450100 milliseconds" yields an animation playback period of 5000 milliseconds (i.e., 5 seconds).

[0061] The preset animation frame rate is the standard playback frame rate set during the design of the vehicle's infotainment system animation. It serves as the benchmark parameter for calculating the theoretical playback frame rate and can be extracted from the animation design parameters or relevant configuration files of the vehicle's infotainment system. For example, the preset animation frame rate for the navigation startup animation mentioned above is 60 frames per second (FPS), representing the expected number of frames per second during animation design. The theoretical playback frame rate is calculated based on the preset animation frame rate and the animation playback period. It represents the ideal number of frames the animation should play completely and can be calculated by multiplying the preset animation frame rate by the animation playback period (unit converted to seconds). For example, 60 frames per second multiplied by 5 seconds gives a theoretical playback frame rate of 300 frames for the navigation startup animation. Actual display frames are the video frames that actually appear on the vehicle's infotainment interface during the animation playback period and exhibit effective visual changes. Invalid duplicate frames or frames that are not displayed correctly are excluded. These can be selected from visual feature timing data to filter video frames that meet the effective display standard during the animation playback period. For example, in the aforementioned 5-second playback period, after removing 3 duplicate frames without visual changes, the remaining valid video frames are all actual display frames. The actual number of frames displayed is the total number of frames actually displayed, reflecting the effective number of frames actually played in the animation; for example, according to statistics, the actual number of frames displayed in the above navigation startup animation is 285 frames.

[0062] The frame drop rate is the proportion of frames lost during animation playback relative to the theoretical number of frames to be played. It is a core indicator for measuring the smoothness of animation and can be calculated by dividing (theoretical number of frames played - actual number of frames displayed) by the theoretical number of frames played and then multiplying by 100%. For example, (300 frames - 285 frames) ÷ 300 frames × 100% gives a frame drop rate of 5% for this animation. The frame drop time distribution characteristic refers to the temporal location, concentration, and other related characteristics of frame drop events within the animation playback period. It reflects the spatiotemporal distribution pattern of frame drops. This can be obtained by analyzing the timestamp sequence of actual displayed frames to locate the time intervals corresponding to frame drop gaps, and then summarizing the distribution pattern of frame drops to obtain the frame drop time distribution characteristic. For example, analysis revealed that of the 25 frames of frame drops in the navigation startup animation, 20 frames were concentrated between the 3rd and 4th seconds of the animation playback, while 5 frames were scattered in other time periods. This concentrated distribution is the frame drop time distribution characteristic.

[0063] By implementing the above embodiments, the animation playback period is determined based on the animation start frame and end frame, and the animation frame drop rate is calculated by combining the theoretical playback frame number and the actual display frame number. At the same time, the time distribution characteristics of frame drop are analyzed, which can not only reflect whether there is a frame drop problem in the animation, but also depict the distribution of frame drop in time. This can more accurately evaluate the difference in smoothness of different vehicle system in animation playback scenarios and improve the accuracy of animation performance comparison analysis between different versions or vehicle models.

[0064] In some embodiments, the aforementioned stuttering features may include the number of stuttering events and the duration of stuttering events; the aforementioned step 103 may include: obtaining the frame interval time and the amount of interface pixel change between adjacent frames in the interface area image sequence based on visual feature time-series data; obtaining stuttering judgment thresholds corresponding to each interaction scene type in the interface area image sequence based on a preset mapping relationship, wherein the preset mapping relationship defines the correspondence between different interaction scene types and different stuttering judgment thresholds, and the interaction scene types may include at least one of touch click scene, list sliding scene, and animation playback scene; determining the stuttering video frame sequence that meets the stuttering conditions based on the first comparison result of the frame interval time and the stuttering judgment threshold, and the second comparison result of the amount of interface pixel change and the preset change threshold; and determining the number of stuttering events and the duration of stuttering events based on the stuttering video frame sequence.

[0065] In some examples, the frame interval time is the difference in timestamps between two adjacent video frames in the interface area image sequence. It reflects the time span between adjacent frames and is an important basis for determining whether the image is frozen. The timestamps of adjacent interface area images can be extracted from the visual feature time series data and calculated by subtracting the timestamp of the previous frame from the timestamp of the next frame. For example, the timestamp of the 100th frame in the interface area image sequence is "1698765460100 milliseconds" and the timestamp of the 101st frame is "1698765460150 milliseconds". The difference of 50 milliseconds is the frame interval time of this group of adjacent frames. Interface pixel change refers to the number of pixels whose grayscale or color values ​​change in two adjacent frames of the interface area image. It reflects the degree of dynamic change in the interface content. The total pixel difference between two adjacent frames of the interface area image can be calculated using image pixel comparison algorithms. For example, the Sum of Absolute Differences (SAD) algorithm can be used to compare two adjacent frames of images, and it can be found that 1200 pixels have changed. This value is the interface pixel change.

[0066] Interaction scenario types are categories of vehicle-mounted system interactions categorized based on user actions and the vehicle's infotainment interface response patterns. Different scenarios have varying requirements for smoothness and standards for perceiving lag. Touch-click scenarios involve users interacting by tapping icons, buttons, or other controls on the vehicle's screen. In this scenario, the interface primarily changes through partial responses or pop-up loading. This can be identified by recognizing single-point touch signals and changes triggered by local elements in visual feature time-series data; for example, it could include interactions triggered by clicking the navigation destination input box or pressing a volume control button. List scrolling scenarios involve users scrolling up, down, left, or right on the vehicle's screen to display list-type interfaces (such as contact lists or music lists). This can be identified by recognizing continuous scrolling signals and the overall displacement of interface elements in visual feature time-series data; for example, it could include scrolling the vehicle's system settings menu or scrolling the multimedia playback history list. Animation playback scenarios are interactive scenarios where the vehicle infotainment system automatically plays animation effects (such as startup animations, transition animations, and function loading animations). In this scenario, the interface is centered on dynamic visual display, which can be determined by identifying the characteristics of no active touch signals and continuous dynamic changes of interface elements in the visual feature time sequence data. For example, it can include scenarios such as playing the brand logo animation when the vehicle starts up and playing the interface transition animation when switching applications.

[0067] The preset mapping relationship is a predefined set of rules used to associate different interaction scenario types with corresponding stuttering thresholds. It can match appropriate stuttering judgment standards for different scenarios. Based on massive user interaction test data and experience survey results, a table of correspondences between scenario types and thresholds can be established and stored. For example, the preset mapping relationship table explicitly states that the stuttering threshold for touch click scenarios is 80 milliseconds, and the threshold for list scrolling scenarios is 100 milliseconds. The stuttering threshold is a time threshold used to determine whether the interval between adjacent frames in a certain interaction scenario constitutes stuttering. When the frame interval exceeds this threshold, stuttering is highly likely to occur. Based on the preset mapping relationship, the corresponding threshold can be matched according to the currently identified interaction scenario type. For example, if the current interaction scenario is a list scrolling scenario, the corresponding stuttering threshold matched from the preset mapping relationship is 100 milliseconds. The first comparison result is the conclusion obtained by comparing the frame interval time of adjacent frames with the stuttering judgment threshold corresponding to the current scene. It is used to initially determine whether there is a possibility of stuttering. The frame interval time can be compared with the stuttering judgment threshold through a numerical comparison algorithm, and the result is output as "greater than", "equal to", or "less than". For example, if the frame interval time is 120 milliseconds and the stuttering judgment threshold is 100 milliseconds, the first comparison result is "the frame interval time is greater than the stuttering judgment threshold".

[0068] The preset change threshold is a pre-defined threshold value for pixel change used to determine whether there is effective dynamic change in the interface. When the pixel change is below this threshold, the interface is considered to have no substantial change. A unified or scenario-specific threshold can be set based on statistical data of normal interface changes in different interaction scenarios. For example, a preset general interface pixel change threshold of 500 pixels is used; anything below this value is considered to have no effective change. The second comparison result is the conclusion obtained by comparing the pixel change of adjacent frames with the preset change threshold. This helps determine whether the interface is in a stagnant state. A numerical comparison algorithm can be used to compare the pixel change with the preset change threshold and output a result of "greater than," "equal to," or "less than." For example, if the pixel change is 300 pixels and the preset change threshold is 500 pixels, the second comparison result is "the pixel change is less than the preset change threshold."

[0069] The stuttering condition is a set of combined rules for determining whether a sequence of video frames constitutes a stuttering event. It's a judgment logic formed by combining the results of a first comparison and a second comparison. Predefined conditions include a frame interval exceeding a stuttering threshold and an interface pixel change below a preset threshold. For example, a set of adjacent frames meets the condition of "frame interval > 100 milliseconds" and "interface pixel change < 500". A stuttering video frame sequence is a set of video frames that continuously meet the stuttering condition. This sequence reflects the continuous process of a stuttering event. Frames that continuously meet the stuttering condition can be selected from the interface area image sequence and arranged in chronological order. For example, if frames 200-203 are four consecutive frames that meet the stuttering condition, the set of these four frames arranged in order constitutes the stuttering video frame sequence.

[0070] The number of stuttering events is the total number of times a stuttering video frame sequence occurs during the entire interaction process, reflecting the frequency of stuttering events. It can be counted and statistically analyzed by identifying stuttering video frame sequences that meet certain criteria within the interface area image sequence. For example, in a 10-minute interaction test, if 3 independent stuttering video frame sequences are identified, the number of stuttering events is 3. The duration of a stuttering event is the duration corresponding to a single stuttering event, reflecting the degree of impact of the stuttering event on the user experience. It can be calculated by subtracting the timestamp of the first frame from the timestamp of the last frame in the stuttering video frame sequence. For example, if the timestamp of the first frame of a stuttering video frame sequence is "1698765470200 milliseconds" and the timestamp of the last frame is "1698765470450 milliseconds", the difference of 250 milliseconds is the duration of the stuttering event.

[0071] The process of determining the number and duration of stuttering events based on stuttering video frame sequences involves statistically analyzing the number of stuttering video frame sequences and calculating the time span of each sequence to obtain core quantitative data on stuttering characteristics. This allows for deduplication and classification of the identified stuttering video frame sequences, as well as the calculation of the time span of each sequence. For example, all consecutive frame sequences that meet the stuttering conditions can be independently identified, and the number of non-overlapping sequences can be counted as the number of stuttering events. The time difference between each sequence can then be calculated as the corresponding stuttering event duration.

[0072] By implementing the above embodiments, combining the interval time between adjacent frames, the amount of change in interface pixels, and the stuttering judgment thresholds corresponding to different interaction scenario types, stuttering events can be identified and the number and duration of stuttering events can be counted. Different judgment standards can be adopted for different interaction scenarios, avoiding errors caused by a one-size-fits-all judgment, thereby more accurately reflecting the stuttering performance of different models or versions in actual use scenarios, and enhancing the consistency and comparability of stuttering indicators in cross-model and cross-version evaluations.

[0073] In some embodiments, step 104 may include: standardizing the touch response latency, animation frame dropping features, and stuttering features respectively to obtain a first standardized value, a second standardized value, and a third standardized value; determining a first weight coefficient corresponding to the first standardized value, a second weight coefficient corresponding to the second standardized value, and a third weight coefficient corresponding to the third standardized value according to a target weight allocation strategy, wherein the target weight allocation strategy is determined based on the vehicle model positioning and / or test target; and performing a weighted calculation based on the first standardized value and the first weight coefficient, the second standardized value and the second weight coefficient, and the third standardized value and the third weight coefficient to obtain a smoothness score.

[0074] In some examples, standardization is a process of converting smoothness metrics with different dimensions and numerical ranges into standardized data on a unified scale. This eliminates dimensional differences between metrics, ensuring comparability and assimilation. Pre-defined standardization algorithms can be used to transform the values ​​of various quantitative metrics. For example, the Min-Max Normalization Algorithm can be used to map different ranges of touch response latency, animation frame drop rates, and other metrics to a uniform 0-1 value range. The first standardized value is the standardized data obtained after standardizing touch response latency, representing the quantitative performance of the touch response latency metric within the unified evaluation system. The second standardized value is the standardized data obtained after standardizing animation frame drop features, representing the quantitative performance of animation frame drop features within the unified evaluation system. The third standardized value is the standardized data obtained after standardizing stuttering features, representing the quantitative performance of stuttering features within the unified evaluation system.

[0075] Vehicle positioning refers to the market segmentation of a target vehicle, reflecting its product positioning and the needs of its target user group. It is a crucial basis for weight allocation and can be determined based on the vehicle's official configuration parameters, market pricing, and product promotional positioning. Examples include vehicles positioned as high-end luxury models, mid-range family cars, and entry-level economy cars. The testing objective is the purpose of evaluating the smoothness of the vehicle's infotainment system. Clearly defining the evaluation focus and application scenarios is a key reference factor for weight allocation and can be based on the specific needs of the test initiator. Examples include testing objectives prioritizing user experience, performance stability, or functional compatibility.

[0076] The target weight allocation strategy is a rule system that assigns weight coefficients to three standardized values ​​based on the target vehicle's model positioning and / or testing objectives. This system reflects the differentiated impact of different indicators on the smoothness score. Weight allocation rule tables can be pre-defined based on vehicle positioning requirements and testing objective orientation. For example, for high-end luxury vehicles where the testing objective prioritizes user experience, an allocation strategy of "first weight coefficient 0.4, second weight coefficient 0.3, third weight coefficient 0.3" can be developed. Similarly, for mid-range family vehicles where the testing objective prioritizes performance stability, an allocation strategy of "first weight coefficient 0.3, second weight coefficient 0.3, third weight coefficient 0.4" can be developed, emphasizing the impact of lag on daily usage stability and aligning with the core need of family users for continuous and stable operation of the vehicle's infotainment system. Furthermore, for entry-level economy vehicles where the testing objective prioritizes the smoothness of basic functions, an allocation strategy of "first weight coefficient 0.45, second weight coefficient 0.25, third weight coefficient 0.3" can be developed, strengthening the weight of touch response latency, a fundamental interactive experience indicator, to match the core demand of entry-level vehicle users for smooth core operations.

[0077] The first weighting coefficient is the weighting parameter corresponding to the first standardized value, reflecting the importance of touch response latency in the smoothness score. The corresponding coefficient can be determined according to the target weighting allocation strategy. The second weighting coefficient is the weighting parameter corresponding to the second standardized value, reflecting the importance of animation frame dropping features in the smoothness score. The corresponding coefficient can be determined according to the target weighting allocation strategy. The third weighting coefficient is the weighting parameter corresponding to the third standardized value, reflecting the importance of stuttering features in the smoothness score. The corresponding coefficient can be determined according to the target weighting allocation strategy.

[0078] The process of obtaining a smoothness score by weighting the first standardized value with the first weight coefficient, the second standardized value with the second weight coefficient, and the third standardized value with the third weight coefficient is to multiply each standardized value with its corresponding weight coefficient and then sum them to obtain a single score that comprehensively reflects the smoothness of the vehicle system. A weighted summation algorithm can be used to calculate the three types of standardized values ​​and their corresponding weight coefficients. For example, the first standardized value 0.8 × 0.4 + the second standardized value 0.75 × 0.3 + the third standardized value 0.9 × 0.3 = 0.81. This result is then mapped to a 0-100 score scale to obtain a smoothness score of 81.

[0079] Through the implementation of the above embodiments, touch response latency, animation frame dropping features, and stuttering features are standardized and processed. After assigning weights according to vehicle model positioning and test objectives, a smoothness score is generated. While preserving the differences of each indicator, the multi-dimensional smoothness performance is uniformly mapped to a single score result, so that different vehicle models or different versions can still be compared intuitively and quantitatively under different test objectives. This effectively solves the problem of lack of a unified evaluation standard in the evaluation of vehicle system smoothness.

[0080] Furthermore, as an implementation of the aforementioned method embodiments, this application also provides a vehicle infotainment system smoothness evaluation device for implementing the aforementioned method embodiments. This device embodiment corresponds to the aforementioned method embodiments. For ease of reading, this vehicle infotainment system smoothness evaluation device embodiment will not repeat the details of the aforementioned method embodiments one by one, but it should be clear that the device in this application embodiment can correspondingly implement all the contents of the aforementioned method embodiments. For example... Figure 2 As shown, the vehicle infotainment system smoothness evaluation device 20 includes: a video acquisition unit 201, a video parsing unit 202, an information quantification unit 203, and a smoothness scoring unit 204. The video acquisition unit 201 is used to acquire a recorded video stream of the vehicle infotainment system interaction process for a target vehicle. The video parsing unit 202 is used to parse and process the recorded video stream to obtain the visual feature time-series data of the vehicle infotainment system during the interaction process. The information quantification unit 203 is used to determine the smoothness quantification information of the target vehicle infotainment system based on the aforementioned visual feature time-series data. The smoothness quantification information may include touch response latency, animation frame drop features, and stuttering features. The smoothness scoring unit 204 is used to generate a smoothness score characterizing the smoothness of the vehicle infotainment system based on the touch response latency, animation frame drop features, and stuttering features.

[0081] In some embodiments, the video acquisition unit 201 is further configured to control at least one camera device to record the vehicle screen of the target vehicle based on preset ambient light compensation, preset shooting position and preset device anti-shake fixing processing to obtain an initial video stream, wherein the recording process covers multiple preset interactive scenarios; acquire operation trigger signal timestamps and hardware response signal timestamps synchronized with the initial video stream; and bind the operation trigger signal timestamps and hardware response signal timestamps with the corresponding video frames in the initial video stream to obtain a recorded video stream.

[0082] In some embodiments, the video parsing unit 202 is further configured to perform frame-by-frame preprocessing on the recorded video stream to obtain a target video frame sequence, wherein the frame-by-frame preprocessing includes at least one of image denoising, adaptive image stabilization, color calibration, and frame rate normalization; perform detection and segmentation processing on the target video frame sequence to obtain an interface area image sequence of the vehicle's infotainment screen; extract visual features from each interface area image in the interface area image sequence to obtain a feature representation of the vehicle's infotainment interface in each target video frame; determine multiple preset event frames based on the feature representation, operation trigger signal timestamp, and hardware response signal timestamp, wherein the preset event frames include touch operation trigger frames, interface response frames, animation start frames, animation end frames, and stuttering exception frames; and construct visual feature time-series data based on the interface area image sequence and the timestamp of each preset event frame.

[0083] In some embodiments, the information quantization unit 203 is further configured to determine the operation trigger reference time point based on the touch operation trigger frame timestamp and operation trigger signal timestamp in the visual feature timing data; determine the interface feedback time point based on the interface response frame timestamp and hardware response signal timestamp in the visual feature timing data; determine the initial delay data based on the time difference between the interface feedback time point and the operation trigger reference time point; and perform anomaly filtering on the initial delay data to obtain the touch response delay.

[0084] In some embodiments, the animation frame dropping features include the animation frame dropping rate and the frame dropping time distribution features; the information quantization unit 203 is also used to determine the animation playback period based on the animation start frame timestamp and animation end frame timestamp in the visual feature time series data; determine the theoretical number of playback frames based on the preset animation frame rate and the animation playback period; identify the actual display frames in the animation playback period from the visual feature time series data, and count the number of actual display frames to obtain the actual display frame count; determine the animation frame dropping rate based on the theoretical playback frame count and the actual display frame count; and analyze the timestamp distribution of each actual display frame in the total animation playback period to obtain the frame dropping time distribution features.

[0085] In some embodiments, the stuttering features include the number of stuttering events and the duration of stuttering events; the information quantization unit 203 is further configured to obtain the frame interval time and the amount of interface pixel change between adjacent frames in the interface area image sequence based on visual feature time-series data; obtain stuttering judgment thresholds corresponding to each interaction scene type in the interface area image sequence based on a preset mapping relationship, wherein the preset mapping relationship defines the correspondence between different interaction scene types and different stuttering judgment thresholds, and the interaction scene types include at least one of touch click scene, list sliding scene, and animation playback scene; determine the stuttering video frame sequence that meets the stuttering conditions based on the first comparison result of the frame interval time and the stuttering judgment threshold, and the second comparison result of the amount of interface pixel change and the preset change threshold; and determine the number of stuttering events and the duration of stuttering events based on the stuttering video frame sequence.

[0086] In some embodiments, the smoothness scoring unit 204 is further configured to standardize the touch response latency, animation frame dropping features, and stuttering features respectively to obtain a first standardized value, a second standardized value, and a third standardized value; determine a first weight coefficient corresponding to the first standardized value, a second weight coefficient corresponding to the second standardized value, and a third weight coefficient corresponding to the third standardized value according to a target weight allocation strategy, wherein the target weight allocation strategy is determined based on the vehicle model positioning and / or test target of the target vehicle; and perform a weighted calculation based on the first standardized value and the first weight coefficient, the second standardized value and the second weight coefficient, and the third standardized value and the third weight coefficient to obtain a smoothness score.

[0087] This application also provides a computer-readable storage medium storing computer-executable instructions or computer programs, which, when executed by a processor, will cause the processor to perform any step of the vehicle system smoothness evaluation method provided in this application.

[0088] In some embodiments, the computer-readable storage medium may be a random access memory (RAM), a read-only memory (ROM), flash memory, a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM); or it may be a variety of devices that include one or any combination of the above-mentioned memories.

[0089] In some embodiments, computer-executable instructions may take the form of programs, software, software modules, scripts, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as stand-alone programs or as modules, components, subroutines, or other units suitable for use in a computing environment.

[0090] In some embodiments, computer-executable instructions may, but do not necessarily, correspond to files in a file system, and may be stored as part of a file that holds other programs or data, for example, in one or more scripts in a HyperText Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple co-located files (e.g., files that store one or more modules, subroutines, or code sections).

[0091] In some embodiments, computer-executable instructions may be deployed to execute on an electronic device, or on multiple electronic devices located at one location, or on multiple electronic devices distributed across multiple locations and interconnected via a communication network.

[0092] like Figure 3 As shown, this application also provides an electronic device 30, including a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor. When the processor 320 executes the computer program 311, it implements any step of the above-described vehicle system smoothness evaluation method.

[0093] This application also provides a computer program product comprising a computer program or computer-executable instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer program or computer-executable instructions from the computer-readable storage medium and executes the computer program or computer-executable instructions, causing the electronic device to perform any step of the vehicle system smoothness evaluation method described above.

[0094] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for evaluating the smoothness of a vehicle infotainment system, characterized in that, include: Acquire a recorded video stream of the vehicle-to-infrastructure (V2I) interaction process for the target vehicle; The recorded video stream is parsed and processed to obtain the visual feature time-series data of the vehicle's infotainment system during the interaction process. Based on the visual feature time-series data, the smoothness metrics of the target vehicle system are determined, wherein the smoothness metrics include touch response latency, animation frame drop features, and stuttering features. Based on the touch response latency, the animation frame dropping features, and the stuttering features, a smoothness score is generated to characterize the smoothness of the vehicle system.

2. The vehicle infotainment system smoothness evaluation method according to claim 1, characterized in that, The acquisition of the recorded video stream of the vehicle-to-machine interaction process for the target vehicle includes: Based on preset ambient light compensation, preset shooting position and preset device anti-shake fixing processing, at least one camera device is controlled to record the vehicle screen of the target vehicle to obtain an initial video stream, wherein the recording process covers multiple preset interactive scenarios; Acquire the operation trigger signal timestamp and hardware response signal timestamp that are synchronized with the initial video stream; The operation trigger signal timestamp and the hardware response signal timestamp are bound to the corresponding video frames in the initial video stream to obtain the recorded video stream.

3. The vehicle infotainment system smoothness evaluation method according to claim 2, characterized in that, The step of parsing and processing the recorded video stream to obtain the visual feature time-series data of the target vehicle's in-vehicle infotainment system during the interaction process includes: The recorded video stream is preprocessed frame by frame to obtain a target video frame sequence, wherein the frame-by-frame preprocessing includes at least one of image denoising, adaptive image stabilization, color calibration and frame rate normalization. The target video frame sequence is detected and segmented to obtain the interface area image sequence of the vehicle's infotainment screen. Visual features are extracted from each interface region image in the interface region image sequence to obtain the feature representation of the vehicle interface in each target video frame. Based on the feature representation, the operation trigger signal timestamp, and the hardware response signal timestamp, multiple preset event frames are determined, wherein the preset event frames include touch operation trigger frames, interface response frames, animation start frames, animation end frames, and stuttering exception frames. The visual feature time-series data is constructed based on the image sequence of the interface area and the timestamp of each preset event frame.

4. The vehicle infotainment system smoothness evaluation method according to claim 3, characterized in that, The step of determining the fluency quantification information of the target vehicle system based on the visual feature time-series data includes: Based on the touch operation trigger frame timestamp and operation trigger signal timestamp in the visual feature timing data, the operation trigger reference time point is determined; Based on the interface response frame timestamp and hardware response signal timestamp in the visual feature time series data, the interface feedback time point is determined. The initial delay data is determined based on the time difference between the interface feedback time point and the operation trigger reference time point; The initial delay data is subjected to anomaly filtering to obtain the touch response delay.

5. The vehicle infotainment system smoothness evaluation method according to claim 3, characterized in that, The animation frame drop features include the animation frame drop rate and frame drop time distribution features; the determination of the smoothness quantification information of the target vehicle system based on the visual feature time-series data includes: Based on the animation start frame timestamp and animation end frame timestamp in the visual feature time series data, the animation playback period is determined; The theoretical number of frames to be played is determined based on the preset animation frame rate and the animation playback period. From the visual feature time-series data, identify the actual display frames within the animation playback period, and count the number of actual display frames to obtain the actual display frame count; The animation frame drop rate is determined based on the theoretical playback frame count and the actual display frame count. The timestamp distribution of each actual displayed frame within the total animation playback time period is analyzed to obtain the frame loss time distribution characteristics.

6. The vehicle infotainment system smoothness evaluation method according to claim 3, characterized in that, The lag characteristics include the number of lag events and the duration of each lag event; The step of determining the fluency quantification information of the target vehicle system based on the visual feature time-series data includes: Based on the visual feature time-series data, the frame interval time and interface pixel change amount between adjacent frames in the interface region image sequence are obtained. Based on a preset mapping relationship, a stuttering judgment threshold corresponding to each interaction scene type in the image sequence of the interface area is obtained. The preset mapping relationship defines the correspondence between different interaction scene types and different stuttering judgment thresholds. The interaction scene types include at least one of touch click scene, list swiping scene, and animation playback scene. Based on the first comparison result between the frame interval time and the stuttering determination threshold, and the second comparison result between the interface pixel change amount and the preset change amount threshold, a stuttering video frame sequence that meets the stuttering condition is determined. Based on the stuttering video frame sequence, the number of stuttering events and the duration of each stuttering event are determined.

7. The vehicle infotainment system smoothness evaluation method according to claim 1, characterized in that, The process of generating a smoothness score to characterize the smoothness of the vehicle infotainment system based on the touch response latency, animation frame drop features, and stuttering features includes: The touch response delay, the animation frame dropping feature, and the stuttering feature are standardized respectively to obtain a first standardized value, a second standardized value, and a third standardized value. According to the target weight allocation strategy, a first weight coefficient corresponding to the first standardized value, a second weight coefficient corresponding to the second standardized value, and a third weight coefficient corresponding to the third standardized value are determined, wherein the target weight allocation strategy is determined based on the vehicle model positioning and / or test target of the target vehicle; The fluency score is obtained by weighting the first standardized value with the first weight coefficient, the second standardized value with the second weight coefficient, and the third standardized value with the third weight coefficient.

8. A vehicle infotainment system smoothness evaluation device, characterized in that, include: The video acquisition unit is used to acquire a recorded video stream of the vehicle-to-machine interaction process for the target vehicle. The video parsing unit is used to parse and process the recorded video stream to obtain the visual feature time-series data of the vehicle's infotainment system during the interaction process. The information quantification unit is used to determine the smoothness quantification information of the target vehicle system based on the visual feature time series data, wherein the smoothness quantification information includes touch response latency, animation frame drop features, and stuttering features. The smoothness scoring unit is used to generate a smoothness score that characterizes the smoothness of the vehicle system based on the touch response latency, the animation frame dropping features, and the stuttering features.

9. An electronic device, comprising: The memory and processor are characterized in that the processor, when executing a computer program stored in the memory, implements the steps of the vehicle system smoothness evaluation method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon computer-executable instructions or a computer program, characterized in that, When the computer-executable instructions or the computer program are executed by the processor, the steps of the vehicle system smoothness evaluation method as described in any one of claims 1 to 7 are implemented.

Images