Vehicle hud display method, vehicle, and storage medium

Abstract

The embodiment of the application provides a vehicle HUD display method, a vehicle and a storage medium. The method comprises the following steps: collecting the gaze behavior of a target object in a target period and the vehicle driving state in the target period; determining the driving scene in the target period according to the vehicle driving state in the target period, and determining the user portrait of the target object in the driving scene according to the gaze behavior in the target period, wherein the target object portrait is used to represent the preference of the target object for the head-up display (HUD) layout; and determining the layout information of the display interface corresponding to the HUD according to the driving scene in the target period and the target object portrait. The application solves the technical problem of poor intelligent adjustment of the vehicle HUD display layout in the related art.

Description

Technical Field

[0001] This application relates to the field of automotive smart cockpit technology, specifically to a vehicle HUD display method, a vehicle, and a storage medium. Background Technology

[0002] A Head-Up Display (HUD) system projects key information such as vehicle speed, navigation, and driver assistance information onto the windshield, reducing the number of times drivers need to look down at the instrument panel and improving driving safety. There are several HUD layouts, including: Fixed HUD layout: In most mass-produced vehicles, the content, layout (e.g., speedometer on the left, navigation on the right), and information density (e.g., whether speed limits and lane lines are displayed) of the HUD are fixed at the factory. Users can only manually adjust the overall brightness, height, and horizontal offset of the HUD through the settings menu; they cannot customize the layout of individual information modules. Manually customized layout: Some high-end models or aftermarket HUD products allow users to manually adjust the display position and size of different information modules on the HUD by dragging and dropping via the central control screen or a mobile app. However, this setting is one-time and static, and does not dynamically change according to driving scenarios or user habits. HUD switching based on simple rules: A few vehicles automatically switch HUD display modes when certain functions are activated. For example, navigation arrows are automatically displayed when navigation is activated; when sport mode is activated, the display theme switches to one focused on engine speed and gear position. These switching logics are preset and global, and do not involve learning individual user preferences. Environment-based dynamic obstacle avoidance layout: This focuses on real-time obstacle avoidance layout based on environmental perception. By analyzing environmental data, it actively "makes room" for real-world targets on the HUD, allowing virtual information to "give way" to the real world, preventing obstruction and improving safety. However, it does not involve personalized dynamic optimization for users. Real-world information enhancement based on the object being gazed at: When the system detects that the driver is looking at a real object outside the vehicle (such as a building or road sign), it overlays relevant information about that object onto the windshield.

[0003] There is currently no good solution to the above problems. Summary of the Invention

[0004] This application provides a vehicle HUD display method, a vehicle, and a storage medium to at least solve the technical problem of poor intelligence in adjusting the vehicle HUD display layout in related technologies.

[0005] According to one aspect of the embodiments of this application, a vehicle HUD display method is provided, comprising: collecting the gaze behavior of a target object during a target time period and the vehicle driving status during the target time period; determining the driving scenario during the target time period based on the vehicle driving status during the target time period, and determining the user profile of the target object in the driving scenario based on the gaze behavior during the target time period, wherein the user profile of the target object is used to represent the target object's preference for the HUD layout; and determining the layout information of the display interface corresponding to the HUD based on the driving scenario during the target time period and the user profile of the target object.

[0006] Optionally, determining the user profile of the target object in the driving scenario based on the gaze behavior during the target time period includes: extracting features from the gaze behavior to obtain a gaze behavior feature vector; obtaining a Gaussian mixture model in the driving scenario, wherein the Gaussian mixture model contains multiple components, the components are used to represent the attention pattern of the target object, and the attention pattern is used to represent the target object's attention tendency towards the HUD displayed content; obtaining the posterior probability of the gaze behavior feature vector in the multiple components, and determining the component to which the gaze behavior feature vector belongs based on the posterior probability; and determining the user profile of the target object in the driving scenario based on the component to which the gaze behavior feature vector belongs.

[0007] Optionally, determining the component to which the gaze behavior feature vector belongs based on the posterior probability includes: if the maximum posterior probability among multiple components is greater than a preset probability threshold, determining the component with the maximum posterior probability as the component to which the gaze behavior feature vector belongs; if the maximum posterior probability among multiple components is less than a preset probability threshold, creating a new component in the Gaussian mixture model in the driving scenario, and determining the new component as the component to which the gaze behavior feature vector belongs.

[0008] Optionally, the method further includes: after creating a new component in the Gaussian mixture model in the driving scenario, obtaining the total number of components in the Gaussian mixture model in the driving scenario; if the total number of components in the Gaussian mixture model in the driving scenario is greater than a preset threshold, merging the two most similar components in the Gaussian mixture model in the driving scenario.

[0009] Optionally, determining the user profile of the target object in the driving scenario based on the component to which the gaze behavior feature vector belongs includes: determining the component to which the gaze behavior feature vector belongs as the target component and obtaining the gaze behavior features corresponding to the target component; extracting the number of gazes, average single gaze duration, and total gaze duration in different areas of the HUD layout from the gaze behavior features corresponding to the target component; determining the semantic tags corresponding to the target component based on the number of gazes, average single gaze duration, and total gaze duration; and determining the user profile of the target object in the driving scenario based on the semantic tags corresponding to the target component.

[0010] Optionally, the layout information of the display interface corresponding to the HUD is determined based on the driving scenario and target object profile within the target time period, including: obtaining configurable attributes of multiple regions of interest within the layout of the HUD, wherein the configurable attributes include at least: display markers, center coordinates, scaling factor, transparency, and highlight markers; and determining the configurable attributes of each region of interest based on the driving scenario and target object profile within the target time period to obtain the layout information of the display interface corresponding to the HUD.

[0011] Optionally, the method further includes: after obtaining the layout information of the display interface corresponding to the HUD, collecting the behavior of the target object within a preset time period; if the distribution of the target object's attention frequency to each region of interest within the preset time period matches the preset frequency distribution better than the preset matching degree, determining the layout information of the display interface corresponding to the HUD as positive feedback; if the frequency of the target object's gaze switching between multiple regions of interest within the preset time period exceeds a preset frequency threshold or the layout information of the display interface corresponding to the HUD is modified, determining the layout information of the display interface corresponding to the HUD as negative feedback; if the layout information of the display interface corresponding to the HUD is negative feedback, rolling back the layout information of the display interface corresponding to the HUD to the previous layout configuration and generating negative samples.

[0012] Optionally, the driving scenario within the target time period is determined based on the vehicle's driving status within the target time period, including: extracting the vehicle's driving speed, driving route, and driving time from the vehicle's driving status within the target time period; and determining the driving scenario based on the vehicle's driving speed, driving route, and driving time.

[0013] According to another aspect of the embodiments of this application, a vehicle HUD display device is also provided, comprising: a acquisition module, configured to acquire the gaze behavior of a target object during a target time period and the vehicle driving state during the target time period; a determination module, configured to determine the driving scenario during the target time period based on the vehicle driving state during the target time period, and to determine the user profile of the target object in the driving scenario based on the gaze behavior during the target time period, wherein the user profile of the target object is used to represent the target object's preference for the HUD layout; and a layout module, configured to determine the layout information of the display interface corresponding to the HUD based on the driving scenario during the target time period and the user profile of the target object.

[0014] The determination module is also used to extract features from gaze behavior to obtain gaze behavior feature vectors; obtain a Gaussian mixture model in the driving scenario, where the Gaussian mixture model contains multiple components, which are used to represent the attention patterns of the target object, and the attention patterns are used to represent the target object's attention tendency towards the content displayed on the HUD; obtain the posterior probabilities of the gaze behavior feature vector in multiple components, and determine the component to which the gaze behavior feature vector belongs based on the posterior probabilities; and determine the user profile of the target object in the driving scenario based on the component to which the gaze behavior feature vector belongs.

[0015] The determination module is also used to determine the component with the maximum posterior probability as the component to which the gaze behavior feature vector belongs when the maximum posterior probability among multiple components is greater than a preset probability threshold; and to create a new component in the Gaussian mixture model in the driving scenario when the maximum posterior probability among multiple components is less than a preset probability threshold, and to determine the new component as the component to which the gaze behavior feature vector belongs.

[0016] The determination module is also used to obtain the total number of components in the Gaussian mixture model in the driving scenario after creating new components in the Gaussian mixture model in the driving scenario; if the total number of components in the Gaussian mixture model in the driving scenario is greater than a preset threshold, the two most similar components in the Gaussian mixture model in the driving scenario are merged.

[0017] The determination module is also used to identify the component to which the gaze behavior feature vector belongs as the target component and obtain the gaze behavior features corresponding to the target component; extract the number of gazes, average single gaze duration, and total gaze duration of different regions in the HUD layout from the gaze behavior features corresponding to the target component; determine the semantic label corresponding to the target component based on the number of gazes, average single gaze duration, and total gaze duration; and determine the user profile of the target object in the driving scenario based on the semantic label corresponding to the target component.

[0018] The layout module is also used to obtain the configurable attributes of multiple regions of interest within the layout of the HUD. The configurable attributes include at least: display flags, center coordinates, scaling factor, transparency, and highlight flags. The configurable attributes of each region of interest are determined according to the driving scenario and target object profile within the target time period, thereby obtaining the layout information of the display interface corresponding to the HUD.

[0019] The layout module is also used to collect the behavior of the target object within a preset time period after obtaining the layout information of the display interface corresponding to the HUD; if the target object's behavior within the preset time period indicates that the distribution of the target object's attention to each region of interest matches the preset frequency distribution better than the preset matching degree, the layout information of the display interface corresponding to the HUD is determined to be positive feedback; if the target object's behavior within the preset time period indicates that the frequency of the target object's gaze switching between multiple regions of interest is higher than the preset frequency threshold or the layout information of the display interface corresponding to the HUD is modified, the layout information of the display interface corresponding to the HUD is determined to be negative feedback; if the layout information of the display interface corresponding to the HUD is negative feedback, the layout information of the display interface corresponding to the HUD is rolled back to the previous layout configuration and negative samples are generated.

[0020] The determination module is also used to extract the vehicle's speed, route, and time from the vehicle's driving status within the target time period; and to determine the driving scenario based on the vehicle's speed, route, and time.

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

[0022] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored executable program, wherein, when the executable program is running, it controls the device where the computer-readable storage medium is located to perform the methods of various embodiments of this application.

[0023] According to another aspect of the embodiments of this application, a computer program product is also provided, including a computer program that, when executed by a processor, implements the methods of various embodiments of this application.

[0024] According to another aspect of the embodiments of this application, a computer program product is also provided, including a non-volatile computer-readable storage medium storing a computer program, which, when executed by a processor, implements the methods in various embodiments of this application.

[0025] According to another aspect of the embodiments of this application, a computer program is also provided, which, when executed by a processor, implements the methods of the various embodiments of this application.

[0026] In this embodiment, an automated control method is adopted. The target object's gaze behavior and vehicle driving status during the target time period are collected. The driving scenario during the target time period is determined based on the vehicle driving status. The user profile of the target object in the driving scenario is determined based on the gaze behavior during the target time period. The user profile of the target object is used to represent the target object's preference for the HUD layout. The layout information of the display interface corresponding to the HUD is determined based on the driving scenario and the user profile of the target object during the target time period. This achieves the purpose of automatically adjusting the HUD display layout according to the user profile of the target object, thereby improving the technical effect of intelligent HUD display layout and solving the technical problem of poor intelligent adjustment of vehicle HUD display layout in related technologies. Attached Figure Description

[0027] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0028] Figure 1 This is a flowchart of an optional vehicle HUD display method according to an embodiment of this application;

[0029] Figure 2 This is a schematic diagram of an optional vehicle HUD display system according to an embodiment of this application;

[0030] Figure 3 This is a schematic diagram of an optional HUD display layout interface according to an embodiment of this application;

[0031] Figure 4 This is a flowchart of an optional user profile determination method according to an embodiment of this application;

[0032] Figure 5 This is a comparative schematic diagram showing the HUD display layout interface before and after adjustment according to an embodiment of this application;

[0033] Figure 6 This is a structural block diagram of an optional vehicle HUD display device according to an embodiment of this application. Detailed Implementation

[0034] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0035] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0036] In related technologies, HUD layouts lack dynamic personalization: whether fixed or manually customized, they are static. However, drivers' attention and needs for information vary in different driving scenarios (such as highway cruising, urban congestion, nighttime driving, and rainy / snowy weather). For example, in congested traffic, drivers pay more attention to distance indicators and entertainment information; while on highways, they focus more on speed and remaining navigation distance. Existing technologies cannot dynamically adapt to these changes. Furthermore, they ignore implicit user habits: existing systems cannot learn individual user behaviors, such as whether users frequently check instantaneous energy consumption, frequently need more prominent turn prompts due to missed navigation exits, or briefly glance at the multimedia information area while listening to music. These implicit habits could be used to optimize HUD layout to reduce eye movement, but existing technologies cannot capture or utilize them. Simultaneously, manual configuration is cumbersome and unintelligent: requiring users to manually adjust the position and size of each information module involves a long operation path and high learning costs, resulting in most users not using or continuously optimizing it. This leads to a serious waste of the HUD's personalization potential. Existing methods often focus on one-off adjustments, lacking an on-vehicle intelligent agent that continuously learns from the driver's actual behavior and dynamically updates and optimizes the layout. Vehicles cannot become increasingly "understanding" of their owners as driving time increases, thus failing to provide more accurate optimized layouts. While onboard chip computing power has increased, giving vehicles stronger edge computing capabilities, existing systems have not fully utilized this capability to run complex machine learning models, continuously learn user habits locally, and update and improve personalized optimization strategies. The massive amounts of driving data generated by vehicle sensors and controllers have not been systematically used to build owner-specific layout models; the data value has not been deeply explored, missing the opportunity for personalized solutions.

[0037] According to an embodiment of this application, a method embodiment for a vehicle HUD display method is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0038] This embodiment provides a vehicle HUD display method. Figure 1 This is a flowchart of an optional vehicle HUD display method according to an embodiment of this application, such as... Figure 1 As shown, the process includes the following steps:

[0039] Step S11: Collect the gaze behavior of the target object during the target time period and the vehicle driving status during the target time period;

[0040] In step S11, the behavior and state of the target object within the target time period are synchronously acquired to establish a basis for the correlation between behavioral data and vehicle operating conditions. Through a synchronous monitoring mechanism, on the one hand, gaze behavior data of the target object within this specific time period is collected, recording visual attention characteristics such as gaze direction, dwell area, and duration; on the other hand, vehicle driving status information within the same target time period is synchronously acquired, covering dynamic driving environment parameters such as vehicle speed, road conditions, and navigation events. Here, gaze behavior and vehicle driving status are collected within the same time dimension, ensuring temporal consistency between behavioral data and environmental background.

[0041] Step S12: Determine the driving scenario within the target time period based on the vehicle driving status within the target time period, and determine the user profile of the target object in the driving scenario based on the gaze behavior within the target time period. The user profile of the target object is used to represent the target object's preference for the head-up display (HUD) layout.

[0042] In step S12, the current driving scenario is defined by monitoring the overall driving status of the vehicle within the target time period, serving as the basic environmental background for subsequent analysis. Subsequently, the system collects and analyzes the target object's gaze behavior data towards the head-up display (HUD) within the target time period, and combines the acquired gaze behavior with the defined driving scenario to derive the target object's user profile in the current driving scenario. The determined user profile is specifically used to characterize the target object's preference features for the HUD layout, achieving a preference mapping based on environmental context and individual behavioral data.

[0043] Step S13: Determine the layout information of the display interface corresponding to the HUD based on the driving scenario and target object profile within the target time period.

[0044] In step S13, the layout information of the display interface corresponding to the HUD is determined based on the identification of the driving scenario within the target time period and the matching of the target object profile. This process aims to combine specific driving environment characteristics with individual user behavior preferences or attention characteristics to generate personalized display configurations. Specifically, the system acquires the current driving scenario (such as highway, city, night, etc.) and the target object profile belonging to that scenario (such as user attention patterns formed based on historical gaze behavior and learning). Through the correlation analysis of these two, the system decides on the most suitable HUD display interface layout for the current situation and user habits, including parameters such as the position, size, and display status of each information module, to achieve optimized information presentation.

[0045] The gaze behavior of the aforementioned target objects and the vehicle's driving status can be collected by the in-vehicle driver monitoring system (DMS) to obtain the gaze direction, gaze point coordinates, gaze duration, and scan path; the current layout configuration can be obtained by the HUD system itself; and vehicle and driving scenario information such as vehicle speed, gear, turn signal status, navigation events (such as turn prompts), and ADAS (Advanced Driver Assistance Systems) warnings can be obtained by the CAN (Controller Area Network) bus.

[0046] It is understood that the process of extracting gaze behavior features in the vehicle HUD display method provided in this application embodiment is as follows: the driver's gaze frequency, average single gaze duration, and saccade review interval for different HUD information areas (e.g., vehicle speed area, navigation area, multimedia area, ADAS warning area) are calculated in real time on the vehicle end.

[0047] At the same time, the system records user control operations on the HUD (such as switching multimedia information via steering wheel buttons or turning off a prompt via voice command) as preference signals, and then extracts preference features.

[0048] The user profile generation process is as follows: A lightweight incremental clustering algorithm (such as online K-means) is used to perform online clustering of the above-mentioned gaze behavior features and preference features, based on driving scenarios (the scenarios can be identified in real time through information such as vehicle speed, navigation status, time, and sensors). Each cluster center represents a typical "attention distribution pattern" of the user in a certain situation, i.e., a user profile sub-label (such as "navigation dependent", "multimedia attention", "minimalist").

[0049] The HUD display layout can be optimized by inputting the current context and the sub-labels corresponding to the identified user profile, and outputting an optimized HUD layout configuration. The optimization objective function is to minimize the user's expected scanning path length for key driving information (such as vehicle speed and warnings) while maximizing the visibility of secondary information preferred by the user (such as multimedia). Output parameters include: display / hide of each information module, display position coordinates (X, Y), scaling ratio, transparency, and highlight status.

[0050] After implementing the new layout, continue to monitor user gaze behavior and manual intervention. If the user quickly shifts their gaze back to the road and the duration of their gaze on the new layout is short (indicating readability), or if no manual adjustments are made, this is considered positive feedback, strengthening the mapping between the current user profile and the layout. If the user frequently makes manual adjustments or exhibits confused scanning (rapidly jumping between areas), this is considered negative feedback, triggering layout rollback and strategy adjustments, and updating the user profile model.

[0051] Based on the above steps S11 to S13, this embodiment of the application adopts an automated control method, which collects the gaze behavior of the target object during the target time period and the vehicle driving status during the target time period; determines the driving scenario during the target time period based on the vehicle driving status during the target time period, and determines the user profile of the target object in the driving scenario based on the gaze behavior during the target time period. The target object profile is used to represent the target object's preference for the HUD layout; and determines the layout information of the display interface corresponding to the HUD based on the driving scenario during the target time period and the target object profile. This achieves the purpose of automatically adjusting the HUD display layout according to the user profile of the target object, thereby realizing the technical effect of improving the intelligence of the HUD display layout and solving the technical problem of poor intelligence in adjusting the vehicle HUD display layout in related technologies.

[0052] In some embodiments of this application, the specific steps for determining the user profile of a target object in a driving scenario based on gaze behavior during a target time period are as follows: feature extraction is performed on the gaze behavior to obtain a gaze behavior feature vector; a Gaussian mixture model is obtained for the driving scenario, wherein the Gaussian mixture model contains multiple components, which represent the attention patterns of the target object, and the attention patterns represent the target object's tendency to focus on the content displayed on the HUD; the posterior probabilities of the gaze behavior feature vector in the multiple components are obtained, and the component to which the gaze behavior feature vector belongs is determined based on the posterior probabilities; the user profile of the target object in the driving scenario is determined based on the component to which the gaze behavior feature vector belongs.

[0053] It should be noted that in the absence of a Gaussian mixture model in the driving scenario, a model containing one Gaussian component can be initialized.

[0054] It should also be noted that gaze behavior can be obtained statistically, or the coordinates of the gaze point in the next second can be directly predicted using sequence models (LSTM or Transformer), rather than statistical region features. The model takes the gaze point sequence of the past 5 seconds and the vehicle state as input and outputs the predicted gaze point. The layout optimization engine dynamically renders the highest priority information (such as warnings) near the predicted location based on the predicted location, achieving "information following the gaze".

[0055] In addition to manual operation, user needs can also be obtained by integrating voice interaction intent: for example, if a user says "how far is the next service area", the system recognizes the temporary strong demand for navigation information, and proactively enlarges and highlights the remaining navigation distance and service area information at the top for a period of time, and then restores it.

[0056] Feature extraction of gaze behavior to obtain gaze behavior feature vectors involves transforming the collected raw gaze data (such as eye movement trajectory, dwell time, etc.) into a high-dimensional numerical representation for machine quantification. "Obtaining a Gaussian mixture model for the driving scenario" refers to calling a pre-set statistical model for the current specific driving environment (such as urban congestion, highway cruising, etc.). This model consists of multiple Gaussian components, where "component" specifically refers to a probability distribution unit representing a particular attention pattern, such as "frequently checking the left rearview mirror" or "focusing on central navigation information." "Obtaining the posterior probability of the gaze behavior feature vector among multiple components and determining the component to which the gaze behavior feature vector belongs based on the posterior probability" involves calculating the probability of the current feature vector being generated by each component using Bayes' theorem and selecting the component with the highest probability as the judgment result, thus discretizing continuous, ambiguous gaze behavior into a clear attention pattern classification. "Determining the user profile of the target object in the driving scenario based on the component to which the gaze behavior feature vector belongs" maps the determined attention pattern to specific user preference tags. By employing the above steps and using Gaussian mixture models to perform probabilistic clustering of gaze features, we can objectively and precisely identify subtle differences in the attention of target objects in specific driving scenarios. This solves the problems of vague user profile construction methods and inability to quantify subtle preference differences in traditional solutions. It achieves attention pattern segmentation based on statistical probability, enabling the generated user profiles to more accurately reflect the target object's specific attention tendency towards the HUD display content, thereby providing a reliable basis for adaptive layout optimization.

[0057] In some embodiments of this application, the specific steps for determining the component to which the gaze behavior feature vector belongs based on the posterior probability are as follows: if the maximum value of the posterior probability among multiple components is greater than a preset probability threshold, the component with the maximum posterior probability is determined to be the component to which the gaze behavior feature vector belongs; if the maximum value of the posterior probability among multiple components is less than the preset probability threshold, a new component is created in the Gaussian mixture model in the driving scenario, and the new component is determined to be the component to which the gaze behavior feature vector belongs.

[0058] When the maximum posterior probability among multiple components is greater than a preset probability threshold, determining the component with the maximum posterior probability as the component to which the gaze behavior feature vector belongs means that when the extracted gaze behavior feature vector highly matches a certain defined attention pattern (component) in the existing Gaussian mixture model, i.e., the calculated attribution probability exceeds the safety judgment line, the system directly uses this known pattern as the basis for the current user's profile. On the other hand, when the maximum posterior probability among multiple components is less than a preset probability threshold, creating a new component in the Gaussian mixture model in the driving scenario and determining the new component as the component to which the gaze behavior feature vector belongs means that when the user's gaze behavior exhibits a new attention tendency that is not covered by the existing model, resulting in a matching degree with all existing components that is lower than the preset standard, the system does not ignore the behavior or force classification, but dynamically instantiates a brand-new component in the Gaussian mixture model of the current driving scenario to specifically represent this new attention pattern. Through this dynamic classification and automatic creation mechanism based on probability thresholds, the system can identify and incorporate users' unpredictable attention habits in real time, avoiding profile matching errors caused by changes in user focus. This enables incremental updates and precise adaptation of user profiles as driving scenarios and user habits evolve.

[0059] After creating a new component in the Gaussian mixture model for the driving scenario, the total number of components in the Gaussian mixture model for the driving scenario is obtained; if the total number of components in the Gaussian mixture model for the driving scenario is greater than a preset threshold, the two most similar components in the Gaussian mixture model for the driving scenario are merged.

[0060] Obtaining the total number of components in the Gaussian mixture model in a driving scenario refers to the real-time counting of the number of Gaussian distribution components used to represent the attention pattern of the target object during algorithm execution. "Merging the two most similar components in the Gaussian mixture model in a driving scenario" means that when the total number of components exceeds a preset threshold, the two Gaussian components with the closest mathematical features are identified by calculating the distance or covariance difference between the mean vectors of each component. Their data parameters are then weighted and averaged or refitted to generate a new, broader single component to replace the original two components. This process dynamically monitors model complexity and automatically triggers a simplification mechanism when the creation of new components causes the model structure to expand beyond the preset load limit. The merging operation eliminates redundant details caused by data accumulation, avoiding the waste of computational resources and inference delays caused by the infinite accumulation of components, while maintaining the salience of the core attention pattern. This significantly improves the real-time processing efficiency and system stability of the vehicle in complex driving scenarios while ensuring the accuracy of personalized profiling.

[0061] In some embodiments of this application, the specific steps for determining the user profile of the target object in a driving scenario based on the component to which the gaze behavior feature vector belongs are as follows: the component to which the gaze behavior feature vector belongs is determined as the target component, and the gaze behavior features corresponding to the target component are obtained; the number of gazes, the average single gaze duration, and the total gaze duration in different areas of the HUD layout are extracted from the gaze behavior features corresponding to the target component; the semantic tags corresponding to the target component are determined based on the number of gazes, the average single gaze duration, and the total gaze duration; and the user profile of the target object in a driving scenario is determined based on the semantic tags corresponding to the target component.

[0062] It is understandable that the semantic tags corresponding to the target components can be used as tags for user profiles.

[0063] Determining the component to which the gaze behavior feature vector belongs as the target component and obtaining the gaze behavior features corresponding to the target component means that the system locks the most representative attention pattern category by calculating the posterior probability and retrieves the preset behavioral statistical benchmark of that category. Extracting the number of times the user gazes, the average single gaze duration, and the total gaze duration in different areas of the HUD layout from the gaze behavior features corresponding to the target component means that specific indicators reflecting the distribution of the user's visual focus are quantitatively extracted from the above benchmarks. For example, the more times a certain area is gazed at and the longer the single dwell time, the higher the attractiveness or information acquisition demand of that area for the user. Determining the semantic label corresponding to the target component based on the number of times the user gazes, the average single gaze duration, and the total gaze duration means that the above quantitative data is mapped to intuitive business meanings, such as "high-frequency attention area" or "ignore area", thereby giving the abstract statistical component an interpretable physical meaning. Determining the user profile of the target object in the driving scenario based on the semantic label corresponding to the target component means that a personalized model describing the user's HUD layout preferences is constructed based on these semantic labels. By introducing the component classification mechanism of Gaussian mixture model and specific behavioral feature extraction and semantic mapping steps, the originally black-box gaze behavior data is transformed into user preference labels with clear physical meaning and interpretability. This enables the system to accurately identify the user's attention tendency towards different display areas of the HUD based on intuitive indicators such as "number of gazes" and "average single gaze duration". This provides a quantitative and reliable basis for generating personalized HUD layout information, effectively improving the targeting of layout optimization and the convenience of user use.

[0064] In some embodiments of this application, the specific steps for determining the layout information of the display interface corresponding to the HUD based on the driving scenario and target object profile within the target time period are as follows: obtaining configurable attributes of multiple regions of interest within the layout of the HUD, wherein the configurable attributes include at least: display flag, center coordinates, scaling factor, transparency, and highlight flag; determining the configurable attributes of each region of interest based on the driving scenario and target object profile within the target time period to obtain the layout information of the display interface corresponding to the HUD.

[0065] Obtaining configurable attributes for multiple regions of interest within the HUD layout refers to dividing the HUD display interface into several independent visual function modules (such as navigation arrows, vehicle speed, warning lights, etc.) and extracting specific parameters for each module, including display indicators, center coordinates, scaling factor, transparency, and highlight indicators. Specifically, "display indicators" control the visibility of information elements; "center coordinates" determine their specific position on the windshield projection surface; "scaling factor" adjusts the size of the information content to fit the field of view; "transparency" adjusts the degree of background occlusion to ensure adequate observation of the road ahead; and "highlight indicators" enhance the visual weight of key information. Furthermore, based on the driving scenario and target object profile within the target time period, specific parameters are determined... Defining configurable attributes for each region of interest means that the system does not uniformly adjust the entire interface. Instead, it independently calculates the optimal parameter combination for each module based on the current driving conditions (such as highway cruising or congested following) and the user's personal visual preferences. For example, in nighttime driving scenarios, it automatically reduces the transparency of non-critical information and adjusts the center coordinates to avoid blind spots. By configuring multi-dimensional attributes independently for each region of interest, the system achieves dynamic and refined adjustments to the HUD display interface layout. This ensures that the information display meets both the safety requirements of the current driving scenario and the user's personalized habits, effectively solving the problems of rigid traditional HUD layouts and the inability to improve the driving experience through collaborative optimization of multi-dimensional attributes such as position, size, and visibility.

[0066] After obtaining the layout information of the HUD's corresponding display interface, the behavior of the target object is collected within a preset time period. If the distribution of the target object's attention frequency in each region of interest within the preset time period matches the preset frequency distribution better than the preset matching degree, the layout information of the HUD's corresponding display interface is determined to be positive feedback. If the frequency of the target object's gaze switching between multiple regions of interest within the preset time period exceeds the preset frequency threshold or the layout information of the HUD's corresponding display interface is modified, the layout information of the HUD's corresponding display interface is determined to be negative feedback. If the layout information of the HUD's corresponding display interface is negative feedback, the layout information of the HUD's corresponding display interface is rolled back to the previous layout configuration, and negative samples are generated.

[0067] In addition to the methods mentioned above, HUD display layouts can also utilize augmented reality HUDs (AR-HUDs), optimizing the anchoring position and animation effects of 3D virtual information instead of 2D layouts. For example, for "navigation-dependent" users, turn arrows can be dynamically "drawn" on the real road surface instead of in a fixed area. When significant changes occur in the HUD layout, a voice prompt can inform the user: "We have detected that you are using navigation. The navigation prompt has been moved to the center. Are you comfortable with this?" Allowing the user to confirm / decline via nodding (recognized by the DMS) or voice, forming a more natural interactive loop. Optimization goals can be extended to visual comfort and cognitive load. For example, when it is detected that the driver is staring at a complex information area for an extended period, the displayed content in that area can be automatically simplified, reducing information density; or the overall brightness and contrast of the HUD can be automatically adjusted based on ambient light and the driver's pupil size, working in conjunction with layout optimization.

[0068] The applicability of the current layout is verified by monitoring actual interaction data during driving. The "matching degree between the gaze frequency distribution and the preset frequency distribution" refers to comparing the statistical results of the frequency of the target object's natural gaze at various regions of interest (such as vehicle speed, navigation arrows, warning information, etc.) within a preset time period with the gaze frequency model under standard or ideal conditions in that scenario. If the comparison result is higher than the preset matching degree, it indicates that the target object's gaze flow conforms to expected habits. At this time, the layout information of the HUD's corresponding display interface is determined to be positive feedback, indicating that the layout does not need adjustment. If the matching degree is insufficient, it is further determined whether the target object's gaze frequently switches between multiple regions of interest (i.e., the gaze switching frequency is higher than the preset frequency threshold, reflecting that the target object's gaze is not properly adjusted due to the frequency of gaze switching). When information is scattered or poorly laid out, making it difficult to quickly locate key information (a state of anxiety or searching), or when a manual layout modification operation is triggered, the layout information of the HUD's corresponding display interface is determined to be negative feedback. When negative feedback is determined, the layout information of the HUD's corresponding display interface is rolled back to the previous layout configuration, that is, restored to the layout state that was adopted by the system or confirmed by the user before the current layout was generated. This immediately eliminates the occupation of the driver's attention by the current bad layout, avoids the continuous interference of incorrect configuration with driving safety, and generates a negative sample to record this failure case, providing data support for subsequent algorithm optimization. This realizes dynamic closed-loop optimization of the layout strategy, improves the personalization adaptation accuracy of HUD display information and driving safety.

[0069] In some embodiments of this application, the specific steps for determining the driving scenario within a target time period based on the vehicle's driving status within that target time period are as follows: extracting the vehicle's driving speed, driving route, and driving time from the vehicle's driving status within the target time period; and determining the driving scenario based on the vehicle's driving speed, driving route, and driving time.

[0070] Extracting vehicle speed, route, and time from the vehicle's driving status within a target time period involves concretizing the originally broad vehicle operation data into three quantifiable key dimensions: speed reflects the vehicle's dynamic speed, route identifies the physical path the vehicle is on (such as highways, urban roads, or specific navigation sections), and time records the duration of that path or state. Determining the driving scenario based on vehicle speed, route, and time involves using these three parameters as input to comprehensively determine the specific driving situation. For example, when the speed is higher than a set threshold and the route is a highway, it is determined to be a highway cruising scenario, while when the speed is lower and the location is a complex urban intersection, it is determined to be a congested traffic scenario. By introducing these three specific features, the definition of driving scenarios is no longer ambiguous but is accurately divided based on objective and quantifiable key driving elements. This ensures that the subsequently generated user profile can be more accurately mapped to a specific objective driving environment, achieving a close fit between the HUD display layout and the real driving situation, improving the accuracy of personalized configuration and driving safety.

[0071] This application also provides a vehicle HUD display system, which can be deployed in a car equipped with a domain controller and a driver monitoring camera (DMS), such as... Figure 2 As shown, it includes: a data acquisition layer, a feature extraction and context recognition layer, a user profiling layer, a decision-making layer, an execution layer, and a feedback layer. The data acquisition layer includes a 1080p resolution, 60Hz near-infrared DMS camera to capture the driver's eye movements; it collects CAN bus data through a data acquisition module to obtain information such as vehicle speed, navigation information, ADAS warning status, gear position, ambient light sensor values, time, and location; and it obtains the current HUD display layout.

[0072] Feature extraction and context recognition layer: Includes: Gaze mapping module: Utilizes a trained gaze estimation algorithm to calculate the driver's two-dimensional gaze coordinates on the windshield. These coordinates are then matched with the HUD projection area, such as... Figure 3 As shown, the ROIs include ROI-A to ROI-E, with different regions of interest corresponding to different information display areas. The task is to determine which ROI from ROI-A to ROI-E it belongs to. Using a time window of T=10 seconds, the characteristics of each ROI are calculated: number of gazes, total gaze duration, and average single gaze duration.

[0073] The preference analysis module monitors user actions, such as switching multimedia tracks using the steering wheel scroll wheel (indicating attention to multimedia information) or using the voice command "turn off navigation prompts" (indicating a reduced need for navigation information). These events are quantified as preference increment signals.

[0074] The context recognition module outputs context labels based on factors such as vehicle speed, navigation status, and time. For example: vehicle speed > 80 km / h and route is highway → "Highway Cruise"; vehicle speed < 30 km / h and navigation is valid → "City Navigation"; ambient light is dim and time is later than sunset → "Nighttime".

[0075] The user attention profile generation unit is used to generate user profiles. The specific process is as follows: Figure 4 As shown, this unit maintains an independent incremental Gaussian Mixture Model (GMM) for each context to cluster gaze behavior feature vectors: Step 1: Receive gaze behavior feature vectors and context. Step 2: Determine if a GMM model already exists for the context; if not, initialize a model containing one Gaussian component. Step 3: Calculate the posterior probability of each component of the gaze behavior feature vector under the current model. If the maximum posterior probability max(P) is greater than the threshold α (e.g., 0.6), it is assigned to that component (i.e., attention pattern subclass), and the mean, covariance, and mixture weights of that component are updated (using the online EM algorithm with a step size of 0.05). Step 4: If max(P) < α, a new Gaussian component is created for the gaze behavior feature vector, representing a new attention pattern (e.g., "extreme multimedia attention"). If the total number of components exceeds the upper limit (e.g., 8), the two most similar components are merged. Step 5: Each component is assigned a semantic label. The rule engine analyzes the centroid of this component in high-dimensional features: if "the number of times the navigation area is viewed and the average single gaze duration in the navigation area are significantly higher than in other areas", it is labeled as "navigation-dependent"; if "the number of times the central area is viewed is low and the number of times the multimedia area is viewed is high", it is labeled as "entertainment-preference", etc. These labels are stored in the profile database along with the component parameters.

[0076] The HUD layout optimization decision engine outputs a display layout configuration based on the current context and current attention pattern label (user profile label, such as "navigation dependent"): it has a built-in layout parameterized template library. Each template defines configurable attributes for information modules within each ROI: display flag, center coordinates, scaling factor s (0.5-2), transparency a (0-1, 0 for complete transparency), and highlight flag.

[0077] It should be noted that the decision-making logic is as follows: If the current driving scenario is "city navigation" and the user profile label is "navigation-dependent," then the scaling factor of the navigation area ROI-B is set to 1.5, the center coordinate is moved to (0.5, 0.3) (slightly above the center), and the highlight is set to true; at the same time, the transparency of the multimedia area ROI-A is set to 0.7, and non-critical information (such as energy consumption curves) is hidden. If the current driving scenario is "highway cruising" and the user profile label is "entertainment-oriented," then the navigation area ROI-B is hidden (because there is no navigation task), the scaling factor of the multimedia area ROI-A is set to 1.2, and it is moved to (0.7, 0.2) (right secondary view), while keeping the vehicle speed display area ROI-C centered and opaque. The specific layout parameters are then output to the HUD controller.

[0078] The feedback and continuous learning unit monitors driver behavior within a 60-second window after each layout change. Positive feedback conditions: No manual HUD adjustments, and the gaze distribution of each ROI (Region of Interest) matches the expected pattern with a correlation coefficient > 0.8 (e.g., in navigation mode, gaze at navigation ROI-B does indeed increase). Negative feedback conditions: The user accesses the HUD settings menu via steering wheel buttons and modifies the layout, or an abnormal saccade pattern appears (the gaze point jumps frequently between ROIs, variance > threshold). Negative feedback triggers: ① The decision engine is notified to roll back to the previous stable layout; ② The failed scenario (driving scene, user profile label, display layout configuration parameters) is combined as a negative sample and fed back to the profile generation unit, causing it to reduce the weight of the corresponding user profile during subsequent clustering, or guiding the model to split into more refined user profile labels. Figure 5 The diagram shows a comparison of the HUD display layout before and after adjustment. Figure 5 As shown, the figure contains two sub-figures, left and right, illustrating the differences in HUD display status under the same driving scenario: before personalized optimization (default layout) and after adjustment based on the vehicle HUD layout display method in this application (dynamic layout). Specific details are as follows: Figure 5 Left image: The HUD projection area is divided into several fixed regions of interest (ROIs), including the speed display area (usually located on the left or in the center), the navigation guidance area (usually on the right or above), and the multimedia / entertainment information area (usually on the side). The size and position of all information modules are fixed and are not adjusted according to the user's current attention distribution.

[0079] Figure 5The right image shows the following: Information modules that the user is currently most concerned about or that the system deems to have the highest priority (such as navigation arrows or key ADAS warnings) are moved to the driver's main line-of-sight area (i.e., the area where the line of sight deviates least from the road, usually the center or slightly above). Key information modules (such as navigation guidance) are magnified. Secondary information modules (such as multimedia details and non-critical energy consumption data) are minimized. Key information modules are highlighted or have increased opacity to make them more visually prominent; background or non-critical information has reduced brightness or transparency to minimize visual interference.

[0080] The arrows or dashed lines can schematically represent the driver's scattered gaze in the left image (requiring multiple scans of different areas) and the focused gaze in the right image (with the gaze concentrated on the optimized core information area), to demonstrate the technical effect of this application in improving information acquisition efficiency and reducing driving cognitive load.

[0081] Based on the vehicle HUD display method proposed in this application, taking a target user whose daily commuting route is 50% highway and 50% city as an example, on highways, the system identifies that the target user is accustomed to focusing on driving range and multimedia information ("long-distance cruise type"). It automatically enlarges the energy consumption card and song information, shrinks the navigation information to only display lane keeping status, and only enlarges the navigation information 2 km before the highway exit to prompt the user to exit the highway. Upon entering the city, the system switches to "city navigation type," enlarging the navigation arrow and placing it in the center, hiding the multimedia information, and increasing the brightness of the forward collision warning flashing. One day, the target user manually switches the HUD to minimalist mode (displaying only vehicle speed). The system records this operation as strong negative feedback, and subsequently, during similar time periods (such as evening rush hour), it proactively shifts the layout towards a minimalist style.

[0082] Throughout the process, the target never actively set any layout details, but the HUD display increasingly met his expectations, and the time spent scanning his eyes decreased significantly.

[0083] The vehicle HUD display method provided in this application significantly improves driving safety. By dynamically placing the information that the user is currently most concerned about in the optimal field of vision (such as the main line of sight area), it can reduce the cumulative time that the driver's gaze deviates from the road. All user gaze data and operation data are processed in a closed loop on the vehicle end and are not uploaded to the cloud, completely solving the privacy and compliance issues of biometric data. Users do not need to make any manual settings. The system automatically learns and optimizes the layout through daily driving, achieving an evolutionary effect of "the more you use it, the better it becomes." The system can automatically switch to the most suitable layout scheme according to different scenarios such as highways, city streets, nighttime, and rainy days. For example, when driving at night, it automatically reduces the brightness and size of non-critical information to reduce visual interference to the driver. The closed-loop feedback mechanism ensures that when the system's optimization does not conform to the user's current intention (for example, the user wants to focus on music rather than navigation), the system can quickly "correct itself" and learn the user's real-time intention, avoiding a "confrontation" between the system and the user.

[0084] To better illustrate the vehicle HUD display method in the embodiments of this application, further explanation is provided below.

[0085] Step 1: Real-time acquisition of driver gaze behavior data, vehicle driving context data, and current HUD layout configuration: The camera captures images of the driver's face and eyes at a frame rate of 60 frames per second; the CAN bus data receiving module acquires vehicle speed, gear position, turn signal status, navigation events, Advanced Driver Assistance Systems (ADAS) warning status, ambient light sensor values, and timestamps; the HUD configuration module reads the two-dimensional coordinates, scaling ratio, transparency, and display status of each information module on the windshield; the above multi-source data are aligned through a time synchronization mechanism to form a raw data frame containing multimodal information.

[0086] Step 2: Extract driver gaze features, operation preference signals, and driving context labels based on the raw data stream: Convert the eye features in the video frames into two-dimensional gaze point coordinates on the windshield; Spatially match these coordinates with the Region of Interest (ROI) defined by the HUD projection area (including the speed zone, navigation zone, multimedia zone, ADAS warning zone, etc.) to determine the ROI to which the current gaze belongs; Using a 10-second time window, count the number of gazes, total gaze duration, and average single gaze duration for each ROI to form a gaze behavior feature vector; The preference analysis module monitors the operation of switching multimedia tracks using steering wheel buttons or turning off specific prompts using voice commands, quantifying these into incremental signals of multimedia attention preference and navigation demand preference; The context recognition module determines the current driving context label, such as "highway cruise," "city navigation," or "night driving," based on the vehicle speed threshold, navigation activation status, ambient light intensity, and time period, using a rule engine.

[0087] Step 3: Generate user attention profile labels for the current driving context based on incremental Gaussian mixture models: The user attention profile generation unit first checks whether a Gaussian mixture model matching the current driving context label exists in the profile library. If a Gaussian Model (GMM) does not exist, a GMM model containing a single Gaussian component is initialized. If a GMM model exists, the current gaze behavior feature vector is used as input to calculate its posterior probability under each Gaussian component of the current GMM model. The component with the highest posterior probability is selected. If its probability value is greater than a preset threshold (e.g., 0.6), the feature vector is assigned to that component, and the mean, covariance matrix, and mixing weights of that component are updated using an incremental version of the online Expectation-Maximization (EM) algorithm. If the maximum posterior probability is less than the threshold, it is determined to be a new attention pattern, and a new Gaussian component is created. If the total number of components exceeds a preset upper limit (e.g., 8), the two components with the highest similarity are merged. Finally, the rule engine analyzes the feature centroid of the currently active component. If the number of views in the navigation area is significantly higher than in other areas, the semantic label "navigation dependent" is assigned. If the attention in the multimedia area is significantly higher than in other areas, the semantic label "entertainment preference" is assigned, etc., thereby obtaining the current user attention profile label.

[0088] Step 4: Based on driving scenario labels and user attention profile sub-labels, the HUD layout configuration is determined and optimized. The HUD layout optimization decision engine receives the current driving scenario and user attention profile labels as input. The engine first retrieves the basic layout template corresponding to the current scenario from the layout parameterization template library. This template defines the configurable attributes of information modules within each ROI, including display indicators, center coordinates, scaling factor, transparency, and highlight status. Subsequently, the engine dynamically adjusts the parameters in the basic template according to the user attention profile labels. If the label is "navigation dependent" and the scenario is "city navigation," the calculation engine adjusts the scaling factor of the navigation area to 1.5, moves the center coordinate to the upper center of the field of view, enables the highlight status, and reduces the transparency of the multimedia area to 0.7 to reduce visual interference. If the label is "entertainment preference" and the scenario is "high-speed cruising," the calculation engine hides the display indicator of the navigation area, adjusts the scaling factor of the multimedia area to 1.2 and moves it to the right secondary field of view, while keeping the speed area centered and opaque. Finally, the engine calculates a set of layout parameters containing the precise attributes of each information module.

[0089] Step 5: Execute layout configuration change and monitor closed-loop feedback signals: The HUD controller receives the layout configuration parameter set and immediately updates the display content on the windshield to achieve layout switching; within a 60-second monitoring window after the layout change, the system continuously collects the driver's gaze behavior data and operation signals; the feedback and continuous learning unit analyzes the data in this window: if the driver does not manually adjust the HUD settings through steering wheel buttons or voice commands, and the gaze distribution of key information areas (such as the navigation area or vehicle speed area) matches the attention pattern expected by the profile label in Step 3 with a degree greater than 0.8, it is judged as positive feedback; if the driver performs manual adjustment operations, or the gaze point jumps frequently between multiple ROIs (variance exceeds the threshold), exhibiting confused saccades, it is judged as negative feedback.

[0090] Step 6: Update the user attention profile model or perform a layout rollback based on feedback signals:

[0091] If the feedback signal is positive, the system marks the current layout configuration as valid and strengthens the mapping relationship between the current user attention profile sub-label and the layout strategy, without modifying the model parameters. If the feedback signal is negative, the system first notifies the HUD controller to roll back to the previous stable layout configuration before the change to restore the driver's visual comfort. Subsequently, the system uses the failed combination of (driving context label, user attention profile sub-label, and layout configuration) as a negative sample and feeds it back to the user attention profile generation unit in step 3. In subsequent iterations, the user attention profile generation unit reduces the Gaussian component weight of the pattern corresponding to the negative sample or triggers the model splitting mechanism to generate more refined sub-patterns, thereby correcting the user attention profile.

[0092] Figure 6 This is a structural block diagram of an optional vehicle HUD display device according to an embodiment of this application, such as... Figure 6 As shown. It should be noted that this device can be used to execute the aforementioned vehicle HUD display method. The device includes: a data acquisition module 601, used to acquire the gaze behavior of the target object during a target time period and the vehicle's driving state during the target time period; a determination module 602, used to determine the driving scenario during the target time period based on the vehicle's driving state, and to determine the user profile of the target object in the driving scenario based on the gaze behavior during the target time period, wherein the user profile represents the target object's preference for the HUD layout; and a layout module 603, used to determine the layout information of the display interface corresponding to the HUD based on the driving scenario during the target time period and the user profile of the target object.

[0093] The determination module 602 is also used to extract features from the gaze behavior to obtain a gaze behavior feature vector; obtain a Gaussian mixture model in the driving scenario, wherein the Gaussian mixture model contains multiple components, the components are used to represent the attention pattern of the target object, and the attention pattern is used to represent the target object's attention tendency to the HUD display content; obtain the posterior probability of the gaze behavior feature vector in multiple components, and determine the component to which the gaze behavior feature vector belongs based on the posterior probability; determine the user profile of the target object in the driving scenario based on the component to which the gaze behavior feature vector belongs.

[0094] The determination module 602 is further configured to determine the component with the maximum posterior probability as the component to which the gaze behavior feature vector belongs when the maximum posterior probability among multiple components is greater than a preset probability threshold; and to create a new component in the Gaussian mixture model in the driving scenario when the maximum posterior probability among multiple components is less than a preset probability threshold, and to determine the new component as the component to which the gaze behavior feature vector belongs.

[0095] The determination module 602 is also used to obtain the total number of components in the Gaussian mixture model in the driving scenario after creating a new component in the Gaussian mixture model in the driving scenario; and to merge the two most similar components in the Gaussian mixture model in the driving scenario if the total number of components in the Gaussian mixture model in the driving scenario is greater than a preset number threshold.

[0096] The determination module 602 is further configured to determine the component to which the gaze behavior feature vector belongs as the target component, and obtain the gaze behavior features corresponding to the target component; extract the number of gazes, average single gaze duration, and total gaze duration of different regions in the HUD layout from the gaze behavior features corresponding to the target component; determine the semantic label corresponding to the target component based on the number of gazes, average single gaze duration, and total gaze duration; and determine the user profile of the target object in the driving scenario based on the semantic label corresponding to the target component.

[0097] The layout module 603 is also used to obtain configurable attributes of multiple regions of interest within the layout of the HUD. The configurable attributes include at least: display flags, center coordinates, scaling factor, transparency, and highlight flags. The configurable attributes of each region of interest are determined according to the driving scenario and target object profile within the target time period to obtain the layout information of the display interface corresponding to the HUD.

[0098] The layout module 603 is further configured to, after obtaining the layout information of the display interface corresponding to the HUD, collect the behavior of the target object within a preset time period; if the behavior of the target object within the preset time period indicates that the matching degree between the target object's attention frequency distribution to each region of interest and the preset frequency distribution is higher than the preset matching degree, determine that the layout information of the display interface corresponding to the HUD is positive feedback; if the behavior of the target object within the preset time period indicates that the frequency of the target object's gaze switching between multiple regions of interest is higher than the preset frequency threshold or that the layout information of the display interface corresponding to the HUD is modified, determine that the layout information of the display interface corresponding to the HUD is negative feedback; if the layout information of the display interface corresponding to the HUD is negative feedback, roll back the layout information of the display interface corresponding to the HUD to the previous layout configuration and generate negative samples.

[0099] The determination module 602 is also used to extract the vehicle's driving speed, driving route, and driving time from the vehicle's driving status within the target time period; and to determine the driving scenario based on the vehicle's driving speed, driving route, and driving time.

[0100] Embodiments of this application also provide a vehicle, including: a memory storing an executable program; and a processor for running the program, wherein the program executes the methods described in various embodiments of this application when it runs.

[0101] Embodiments of this application also provide a computer-readable storage medium including a stored executable program, wherein, when the executable program is running, it controls the device where the computer-readable storage medium is located to perform the methods of various embodiments of this application.

[0102] Embodiments of this application also provide a computer program product, including a computer program that, when executed by a processor, implements the methods of various embodiments of this application.

[0103] Embodiments of this application also provide a computer program product, including a non-volatile computer-readable storage medium for storing a computer program that, when executed by a processor, implements the methods in various embodiments of this application.

[0104] Embodiments of this application also provide a computer program that, when executed by a processor, implements the methods described in the various embodiments of this application.

[0105] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0106] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0107] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between units or modules may be electrical or other forms.

[0108] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0109] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0110] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0111] The above are merely preferred embodiments of this application. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A vehicle HUD display method, characterized in that, include: Collect the gaze behavior of the target object during the target time period and the vehicle driving status during the target time period; The driving scenario within the target time period is determined based on the vehicle driving status within the target time period, and the user profile of the target object in the driving scenario is determined based on the gaze behavior within the target time period, wherein the target object profile is used to represent the target object's preference for the head-up display (HUD) layout; The layout information of the display interface corresponding to the HUD is determined based on the driving scenario within the target time period and the target object profile.

2. The vehicle HUD display method according to claim 1, characterized in that, Determining the user profile of the target object in the driving scenario based on the gaze behavior during the target time period includes: The gaze behavior is subjected to feature extraction to obtain a gaze behavior feature vector; Obtain the Gaussian mixture model in the driving scenario, wherein the Gaussian mixture model contains multiple components, the components are used to represent the attention pattern of the target object, and the attention pattern is used to represent the target object's attention tendency towards the HUD display content; Obtain the posterior probability of the gaze behavior feature vector in the multiple components, and determine the component to which the gaze behavior feature vector belongs based on the posterior probability; The user profile of the target object in the driving scenario is determined based on the component to which the gaze behavior feature vector belongs.

3. The vehicle HUD display method according to claim 2, characterized in that, Determining the component to which the gaze behavior feature vector belongs based on the posterior probability includes: If the maximum value of the posterior probability among the plurality of components is greater than a preset probability threshold, the component with the maximum posterior probability is determined to be the component to which the gaze behavior feature vector belongs. If the maximum value of the posterior probability among the multiple components is less than a preset probability threshold, a new component is created in the Gaussian mixture model under the driving scenario, and the new component is determined as the component to which the gaze behavior feature vector belongs.

4. The vehicle HUD display method according to claim 3, characterized in that, The method further includes: After creating new components in the Gaussian mixture model under the driving scenario, obtain the total number of components in the Gaussian mixture model under the driving scenario; If the total number of components in the Gaussian mixture model under the driving scenario exceeds a preset threshold, the two most similar components in the Gaussian mixture model under the driving scenario will be merged.

5. The vehicle HUD display method according to claim 2, characterized in that, Determining the user profile of the target object in the driving scenario based on the components to which the gaze behavior feature vector belongs includes: The component to which the gaze behavior feature vector belongs is determined as the target component, and the gaze behavior feature corresponding to the target component is obtained; Extract the number of gazes, average single gaze duration, and total gaze duration of different regions in the HUD layout from the gaze behavior features corresponding to the target components; The semantic tag corresponding to the target component is determined based on the number of times the video is viewed, the average single gaze duration, and the total gaze duration. The user profile of the target object in the driving scenario is determined based on the semantic tags corresponding to the target components.

6. The vehicle HUD display method according to claim 1, characterized in that, The layout information of the display interface corresponding to the HUD is determined based on the driving scenario within the target time period and the target object profile, including: Obtain configurable attributes of multiple regions of interest within the layout of the HUD, wherein the configurable attributes include at least: display flag, center coordinates, scaling factor, transparency, and highlight flag; Based on the driving scenario within the target time period and the target object profile, the configurable attributes of each region of interest are determined, and the layout information of the display interface corresponding to the HUD is obtained.

7. The vehicle HUD display method according to claim 6, characterized in that, The method further includes: After obtaining the layout information of the display interface corresponding to the HUD, the behavior of the target object is collected within a preset time period; If the behavior of the target object within the preset time period indicates that the matching degree between the target object's attention frequency distribution for each region of interest and the preset frequency distribution is higher than the preset matching degree, then the layout information of the display interface corresponding to the HUD is determined to be positive feedback. If, during the preset time period, the behavior of the target object indicates that the frequency of the target object's gaze switching between multiple regions of interest is higher than a preset frequency threshold or that the layout information of the display interface corresponding to the HUD is modified, then the layout information of the display interface corresponding to the HUD is determined to be negative feedback. If the layout information of the display interface corresponding to the HUD is negative feedback, the layout information of the display interface corresponding to the HUD is rolled back to the previous layout configuration, and a negative sample is generated.

8. The vehicle HUD display method according to claim 1, characterized in that, Determining the driving scenario within the target time period based on the vehicle driving status within the target time period includes: Extract the vehicle's speed, route, and travel time from the vehicle's driving status within the target time period; The driving scenario is determined based on the vehicle's speed, route, and travel time.

9. A vehicle HUD display device, characterized in that, include: The acquisition module is used to acquire the gaze behavior of the target object during the target time period and the vehicle driving status during the target time period; The determination module is used to determine the driving scenario within the target time period based on the vehicle driving status within the target time period, and to determine the user profile of the target object in the driving scenario based on the gaze behavior within the target time period, wherein the target object profile is used to represent the target object's preference for the head-up display (HUD) layout; The layout module is used to determine the layout information of the display interface corresponding to the HUD based on the driving scenario within the target time period and the target object profile.

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

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

Images