Vehicle window control method and vehicle

Abstract

This application provides a window control method and vehicle, relating to the field of vehicle control technology. The method includes: acquiring driving environment data and occupant posture data of a target vehicle; determining the current driving scenario characteristics of the target vehicle based on the driving environment data and occupant posture data; inputting the current driving scenario characteristics into a pre-trained window opening inference model to obtain the initial window opening of the target vehicle output by the window opening inference model; correcting the initial window opening based on the safety constraints corresponding to the target vehicle to obtain the final window opening of the target vehicle; and controlling the window of the target vehicle based on the final window opening. This application can solve the problem that existing window control schemes have low intelligence levels and are difficult to meet the needs of drivers and passengers.

Description

Technical Field

[0001] This application relates to the field of vehicle control technology, and in particular to a window control method and a vehicle. Background Technology

[0002] As people's consumption levels rise and more families own cars, consumers are also demanding higher levels of comfort and intelligent features from their vehicles. Among these features, the control methods for car windows, as a core functional component of the vehicle, are evolving from manual operation to intelligent control.

[0003] In related technologies, intelligent window control solutions typically rely solely on single environmental thresholds such as rainfall and sunlight to trigger opening or closing. These solutions generally have a low level of intelligence and cannot meet the development trend of intelligent cockpits that enable seamless human-vehicle-environment collaboration and proactive intelligent services. To a certain extent, this restricts the further improvement of the overall vehicle intelligence level and driving experience. Summary of the Invention

[0004] This application provides a method and vehicle for controlling vehicle windows, in order to solve the problem that the existing window control schemes have low intelligence and are difficult to meet the needs of driving and riding.

[0005] In a first aspect, embodiments of this application provide a method for controlling vehicle windows, including: Acquire driving environment data and occupant posture data of the target vehicle; Based on the driving environment data and the occupant posture data, the current driving scene characteristics of the target vehicle are determined; The current driving scene features are input into a pre-trained window opening inference model to obtain the initial window opening of the target vehicle output by the window opening inference model; wherein, the window opening inference model is trained based on multiple historical driving scene features and their corresponding manual window openings. Based on the safety constraints corresponding to the target vehicle, the initial window opening is corrected to obtain the final window opening of the target vehicle; Based on the final window opening, the target vehicle's windows are controlled.

[0006] Based on the aforementioned technical content, this embodiment of the application simultaneously collects driving environment data and occupant posture data of the target vehicle, comprehensively integrates multi-dimensional information to accurately identify the characteristics of the current driving scenario, and relies on a window opening reasoning model built through training on massive historical driving scenario characteristics and manually generated window opening samples to intelligently reason and match an initial window opening suitable for the current scenario, conforming to the user's actual usage habits and scenario adaptation needs; simultaneously, it combines vehicle safety constraints to legally correct the initial window opening, avoiding driving safety hazards and environmental interference caused by excessive window opening and closing, and finally achieves automated and precise window control based on the corrected final window opening. This embodiment can achieve autonomous perception, intelligent decision-making, and safety constraints of window opening, improving cabin comfort, safety, and intelligence levels.

[0007] In one possible implementation, determining the current driving scenario characteristics of the target vehicle based on the driving environment data and the occupant posture data includes: Obtain the clustering results of pre-stored clustering of multiple historical driving environment data and historical occupant posture data of the target vehicle; wherein each clustering result corresponds to a historical driving scene feature; Based on the driving environment data, the occupant posture data, and the driving scene features corresponding to each clustering result, the current driving scene features are determined.

[0008] This embodiment pre-conducts cluster analysis on massive historical driving environment data and historical occupant posture data, and establishes a one-to-one mapping relationship between clustering results and driving scene features. This enables rapid matching and judgment of real-time collected driving environment data and occupant posture data, efficiently and accurately classifying and identifying the current driving scene features of the target vehicle. Compared with single-dimensional feature judgment or rule-based classification methods, clustering mining of the inherent correlation and scene distribution patterns of multi-source data can filter out abnormal data and improve the comprehensiveness and refinement of scene classification, reducing scene feature recognition bias and misjudgment problems, and providing an accurate and reliable scene input foundation for subsequent window opening model inference.

[0009] In one possible implementation, determining the current driving scene features based on the driving environment data, the occupant posture data, and the driving scene features corresponding to each clustering result includes: Based on the driving environment data, the occupant posture data, and each clustering result, the matching similarity is determined; Based on the matching similarity, the target clustering result is determined from each clustering result; The driving scene features corresponding to the target clustering results are used as the current driving scene features of the target vehicle.

[0010] This embodiment achieves precise scene matching by using similarity metric matching based on pre-stored clustering results. It improves the objectivity and accuracy of scene recognition by relying on the statistical regularity of clustering data, ensuring that the current driving scene feature recognition results are consistent with the actual driving environment and occupant status, and providing reliable and accurate pre-data support for subsequent intelligent reasoning and precise control of window opening.

[0011] In one possible implementation, before correcting the initial window opening based on the safety constraints corresponding to the target vehicle to obtain the final window opening of the target vehicle, the method further includes: Based on the driving environment data, the degree of danger of the current driving condition is determined, and based on the occupant posture data, the degree of danger of the occupant posture is determined. Based on the degree of danger of the current driving conditions and the degree of danger of the occupant posture, the safety constraints corresponding to the target vehicle are determined.

[0012] This embodiment quantifies the degree of danger of driving conditions by using real-time driving environment data, and accurately assesses the degree of danger of occupant posture based on occupant posture data. It integrates the two-dimensional danger levels to generate dynamic safety constraints that are adapted to the current driving state, replacing fixed safety threshold settings. It can dynamically adjust the constraint boundaries to fit real-time road conditions, external environment and actual occupant posture, making the safety constraints more real-time, targeted and comprehensive, effectively making up for the poor adaptability of static safety rules.

[0013] In one possible implementation, determining the safety constraints corresponding to the target vehicle by combining the degree of danger of the current driving condition and the degree of danger of the occupant posture includes: Based on the degree of danger of the current driving condition, a first window opening limit is determined, and based on the degree of danger of the occupant's posture, a second window opening limit is determined; wherein, the first window opening limit is negatively correlated with the degree of danger of the current driving condition, and the second window opening limit is negatively correlated with the degree of danger of the occupant's posture; The safety constraints corresponding to the target vehicle are determined based on the minimum value between the first window opening limit and the second window opening limit.

[0014] By considering the degree of danger of driving conditions and the degree of danger of occupant posture, a first window opening limit and a second window opening limit that are negatively correlated are set. The higher the danger level, the stricter the opening limit. The minimum value of the two opening limits is then selected as the final safety constraint. This approach can take into account both road condition risks and occupant posture hazards, avoid the one-sidedness of single-dimensional constraints, and effectively suppress the safety hazards of excessive window opening in high-risk scenarios.

[0015] In one possible implementation, the step of correcting the initial window opening based on the safety constraints corresponding to the target vehicle to obtain the final window opening of the target vehicle includes: If the initial window opening is greater than the window opening limit corresponding to the safety constraint, then the window opening limit is taken as the final window opening. If the initial window opening is less than or equal to the window opening limit, then the initial window opening is taken as the final window opening.

[0016] In this embodiment, the window opening is corrected based on dynamic safety constraints, which can accurately avoid the safety risks caused by window opening and closing under different dangerous conditions. It takes into account both scenario adaptation needs and driving safety protection, further improving the rationality, flexibility and safety redundancy of the window control strategy, and comprehensively ensuring the driving safety of passengers during the driving process.

[0017] In one possible implementation, there are multiple window opening inference models, each window opening inference model corresponds to a corresponding user identity identifier, and different window opening inference models correspond to different user identity identifiers. The step of inputting the current driving scene features into a pre-trained window opening inference model to obtain the initial window opening of the target vehicle output by the window opening inference model includes: Obtain the identity identifier of the occupant; Based on the user identity identifier corresponding to each window opening inference model, determine the window opening inference model corresponding to the occupant's identity identifier from multiple window opening inference models; The current driving scene features are input into the window opening inference model corresponding to the occupant's identity to obtain the initial window opening of the target vehicle.

[0018] This embodiment identifies occupant identities and uses differentiated window opening inference models to independently allocate model resources for different drivers and passengers. Based on customized models corresponding to each identity, it can deeply learn the long-term window adjustment habits, preferences, and personalized usage patterns of different users. Combining this with the characteristics of the current driving scenario, it infers and outputs an initial window opening that aligns with individual usage habits, effectively avoiding the shortcomings of general models that lack adaptability and cannot accommodate individual preferences. This embodiment achieves personalized customization of window opening inference, significantly improving the humanization and adaptability of intelligent control strategies. It makes the automatic window adjustment effect more suitable for the driving and riding experience of different occupants, further optimizing the cabin riding experience on the basis of intelligent control and enhancing the refined service capabilities of the vehicle's intelligent cockpit.

[0019] In one possible implementation, after controlling the windows of the target vehicle according to the final window opening, the method further includes: Obtain the manual window opening degree corresponding to the current driving scene characteristics; Store the current driving scene features and their corresponding manual window opening degrees as an incremental training set; After the preset incremental training start conditions are met, the window opening inference model is incrementally trained using the incremental training set to update the window opening inference model.

[0020] This embodiment utilizes newly added real-world driving behavior data to incrementally iteratively train the window opening inference model, dynamically updating and optimizing model parameters. This approach eliminates the need for large-scale retraining with full datasets, efficiently saving computing resources and training costs. Simultaneously, it continuously captures users' dynamically changing personalized adjustment habits, usage preferences, and scenario adaptation needs in different scenarios, constantly correcting model inference biases and optimizing opening prediction accuracy. This allows the model to adaptively adapt to long-term user habits and diverse driving scenarios, continuously improving the accuracy, fit, and personalized adaptability of window opening inference. Ultimately, this enables the intelligent window control strategy to possess self-iterative optimization capabilities, significantly improving the intelligent driving experience in the long term.

[0021] In one possible implementation, before inputting the current driving scene features into a pre-trained window opening inference model to obtain the initial window opening of the target vehicle output by the window opening inference model, the method further includes: Acquire multiple historical driving scene features and their corresponding manual window opening degrees, and construct a training sample set based on the multiple historical driving scene features and their corresponding manual window opening degrees; Using the driving scene features in the training sample set as the input of the preset initial model, and the corresponding manual window opening degree as the model output label, the model parameters of the preset initial model are iteratively trained and updated until the preset convergence condition is met, thus obtaining the window opening degree inference model.

[0022] Secondly, embodiments of this application provide a vehicle window control device, including: The acquisition module is used to acquire driving environment data and occupant posture data of the target vehicle; The determination module is used to determine the current driving scene characteristics of the target vehicle based on the driving environment data and the occupant posture data; The inference module is used to input the current driving scene features into a pre-trained window opening inference model to obtain the initial window opening of the target vehicle output by the window opening inference model; wherein, the window opening inference model is trained based on multiple historical driving scene features and their corresponding manual window openings. The correction module is used to correct the initial window opening based on the safety constraints corresponding to the target vehicle, so as to obtain the final window opening of the target vehicle. The control module is used to control the windows of the target vehicle according to the final window opening degree.

[0023] In one possible implementation, the determining module is used to: Obtain the clustering results of pre-stored clustering of multiple historical driving environment data and historical occupant posture data of the target vehicle; wherein each clustering result corresponds to a historical driving scene feature; Based on the driving environment data, the occupant posture data, and the driving scene features corresponding to each clustering result, the current driving scene features are determined.

[0024] In one possible implementation, the determining module is used to: Based on the driving environment data, the occupant posture data, and each clustering result, the matching similarity is determined; Based on the matching similarity, the target clustering result is determined from each clustering result; The driving scene features corresponding to the target clustering results are used as the current driving scene features of the target vehicle.

[0025] In one possible implementation, before correcting the initial window opening based on the safety constraints corresponding to the target vehicle to obtain the final window opening of the target vehicle, the correction module is further configured to: Based on the driving environment data, the degree of danger of the current driving condition is determined, and based on the occupant posture data, the degree of danger of the occupant posture is determined. Based on the degree of danger of the current driving conditions and the degree of danger of the occupant posture, the safety constraints corresponding to the target vehicle are determined.

[0026] In one possible implementation, the correction module is used to: Based on the degree of danger of the current driving condition, a first window opening limit is determined, and based on the degree of danger of the occupant's posture, a second window opening limit is determined; wherein, the first window opening limit is negatively correlated with the degree of danger of the current driving condition, and the second window opening limit is negatively correlated with the degree of danger of the occupant's posture; The safety constraints corresponding to the target vehicle are determined based on the minimum value between the first window opening limit and the second window opening limit.

[0027] In one possible implementation, the correction module is used to: If the initial window opening is greater than the window opening limit corresponding to the safety constraint, then the window opening limit is taken as the final window opening. If the initial window opening is less than or equal to the window opening limit, then the initial window opening is taken as the final window opening.

[0028] In one possible implementation, there are multiple window opening inference models, each window opening inference model corresponds to a corresponding user identity identifier, and different window opening inference models correspond to different user identity identifiers. The reasoning module is used for: Obtain the identity identifier of the occupant; Based on the user identity identifier corresponding to each window opening inference model, determine the window opening inference model corresponding to the occupant's identity identifier from multiple window opening inference models; The current driving scene features are input into the window opening inference model corresponding to the occupant's identity to obtain the initial window opening of the target vehicle.

[0029] In one possible implementation, after controlling the windows of the target vehicle based on the final window opening, the acquisition module is further configured to: Obtain the manual window opening degree corresponding to the current driving scene characteristics; Store the current driving scene features and their corresponding manual window opening degrees as an incremental training set; After the preset incremental training start conditions are met, the window opening inference model is incrementally trained using the incremental training set to update the window opening inference model.

[0030] In one possible implementation, before inputting the current driving scene features into a pre-trained window opening inference model to obtain the initial window opening of the target vehicle output by the window opening inference model, the inference module is further configured to: Acquire multiple historical driving scene features and their corresponding manual window opening degrees, and construct a training sample set based on the multiple historical driving scene features and their corresponding manual window opening degrees; Using the driving scene features in the training sample set as the input of the preset initial model, and the corresponding manual window opening degree as the model output label, the model parameters of the preset initial model are iteratively trained and updated until the preset convergence condition is met, thus obtaining the window opening degree inference model.

[0031] Thirdly, embodiments of this application provide a vehicle including a memory and a processor. The memory stores a computer program that can run on the processor. When the processor executes the computer program, it implements the window control method as described in any of the first aspects.

[0032] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the window control method as described in any of the first aspects.

[0033] It is understood that the beneficial effects of the second to fourth aspects mentioned above can be found in the relevant descriptions in the first aspect mentioned above, and will not be repeated here.

[0034] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this specification. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0036] Figure 1 This is a schematic diagram of an application scenario provided by an embodiment of this application; Figure 2 This is a schematic flowchart of a window control method provided in an embodiment of this application; Figure 3 This is a schematic flowchart of a window control method provided in another embodiment of this application; Figure 4 This is a schematic flowchart of a window control method provided in another embodiment of this application; Figure 5 This is a schematic diagram of the structure of a window control device provided in one embodiment of this application; Figure 6 This is a schematic diagram of the structure of a vehicle provided in one embodiment of this application. Detailed Implementation

[0037] The present application will be described more clearly below with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the function of the present application, but do not limit the present application in any way. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the present application. These all fall within the protection scope of the present application.

[0038] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0039] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0040] In the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0041] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0042] Furthermore, the term "multiple" mentioned in the embodiments of this application should be interpreted as two or more.

[0043] The applicant has found that traditional car window controls rely on passive triggering methods such as manual buttons and voice commands, lacking autonomous perception and adjustment capabilities, and requiring continuous user intervention. Existing smart window solutions often use single environmental parameters as the driving force, such as triggering window closing / fine-tuning based solely on single environmental signals like rainfall, temperature, and vehicle speed, ignoring the coupling relationship between environmental factors and occupant status, resulting in insufficient decision-making accuracy; furthermore, they cannot adapt to user habits and complex scenarios, leading to a poor user experience. Therefore, it is necessary to consider a new method for automated intelligent decision-making regarding window opening.

[0044] To overcome the shortcomings of existing window control methods, such as passivity, low intelligence, and poor adaptability, the embodiments of this application use driving environment parameters and occupant posture data as collaborative perception input dimensions, breaking through the technical limitations of single environment perception or single occupant state recognition. At the same time, a pre-trained window opening inference model is introduced, which constructs feature vectors with environmental features and occupant posture features and completes the model input. Relying on the model's intelligent fuzzy inference capability, it adaptively solves the optimal opening of each window to adapt to the current working conditions, significantly improving the autonomy, intelligence level, and dynamic scene adaptability of window control.

[0045] First refer to Figure 1 , Figure 1 The schematic diagram illustrates an application scenario provided according to an embodiment of this application. The device involved in the application scenario includes a perception input layer 101, a vehicle controller 102, and an execution output layer 103.

[0046] The perception input layer 101 is used to complete the real-time acquisition of multi-source data. In this embodiment, this layer realizes the collaborative acquisition of multi-source driving environmental information and cabin occupant posture information, breaking the limitations of single data acquisition. Among them, the environmental dimension relies on vehicle-mounted sensing components to collect multi-dimensional environmental parameters such as temperature and humidity inside and outside the vehicle, light intensity, rainfall, driving speed, air quality inside and outside the vehicle, and sunlight incidence angle in real time; the occupant posture dimension integrates multiple sensing devices such as cabin cameras, millimeter-wave radar, seat pressure sensors, and infrared sensors to accurately identify the number of occupants, their seating positions (driver's seat, front passenger seat, rear seat area), head posture (forward tilt, backward tilt, side turn), body extension state, whether there is any leaning out of the window, and the occupant identity attributes such as adults and children, comprehensively and meticulously acquiring occupant posture and status data, providing complete and reliable underlying data support for subsequent scene recognition and intelligent window control.

[0047] The vehicle controller 102 undertakes core computing and decision-making functions. It can perform in-depth analysis of environmental data and occupant posture data collected by the perception input layer to identify driving scene characteristics. At the same time, it calls the window opening inference model and combines dynamically generated vehicle safety constraints to intelligently infer the optimal window opening for the current working condition, and outputs a precise and reliable window adaptive control strategy accordingly.

[0048] The execution output layer 103 mainly integrates hardware modules such as the window drive motor, transmission actuator, and drive control unit, serving as the terminal execution carrier for window adjustment. This layer receives window adaptive control strategies and opening commands from the vehicle controller 102 in real time, accurately responds to intelligent control commands, and drives each window mechanism to complete lifting and lowering adjustment and opening positioning actions. At the same time, relying on high-precision position closed-loop control logic, it performs real-time calibration and deviation correction of the window travel, stably constraining the control error between the actual window opening and the target opening within ±1mm, effectively ensuring the positioning accuracy, operational stability, and action consistency of window adjustment, and achieving refined and highly reliable window adaptive execution control.

[0049] In some embodiments, the execution entity of the control strategy of this application can be flexibly expanded, not limited to deployment on the local vehicle controller, but can also be migrated to the cloud server to complete the computing power support. For example, data can be uploaded to the cloud by relying on the vehicle network, and the cloud server can centrally perform scene feature extraction, model inference, safety constraint verification and opening calculation, and then send the generated control commands to the vehicle execution unit. This can reduce the computing power load of the local vehicle controller, and can also use the large computing power of the cloud to achieve model iterative optimization and high-precision calculation of complex scenarios, thereby improving the flexibility, scalability and computing performance of the solution.

[0050] The following is combined Figure 1 Application scenarios, refer to Figures 2-4 This application describes a window control method provided according to exemplary embodiments. It should be noted that the above application scenarios are shown only to facilitate understanding of the spirit and principles of this application, and the embodiments of this application are not limited in any way. Rather, the embodiments of this application can be applied to any applicable scenario.

[0051] It should be noted that the window control method provided according to the exemplary embodiments of this application can be executed on the same device or on different devices.

[0052] refer to Figure 2 , Figure 2 This is a schematic flowchart illustrating a window control method provided in one embodiment of this application. Figure 2 As shown, the method in the embodiments of this application may include: Step S201: Obtain the driving environment data and occupant posture data of the target vehicle.

[0053] Here, driving environment data is used to characterize the state of the surrounding environment inside and outside the vehicle during the driving process of the target vehicle. Its acquisition depends on various on-board sensing devices, including but not limited to: temperature and humidity parameters inside and outside the vehicle collected by on-board temperature and humidity sensors, light intensity parameters inside and outside the vehicle collected by light sensors, rainfall parameters collected by rain sensors, vehicle speed parameters collected by vehicle speed sensors, air quality parameters inside and outside the vehicle collected by air quality sensors, and various environmental data such as sunlight incident angle parameters collected by angle sensors.

[0054] Occupant posture data is used to characterize the seating status and body posture of occupants in the cabin of a target vehicle. Specifically, it may include, but is not limited to, various posture and status-related data such as the number of occupants in the cabin, occupant seating positions (such as driver's seat, front passenger seat, rear area, etc.), occupant head postures (such as forward tilt, backward tilt, side turn, etc.), occupant body extension degree, whether occupants are engaging in dangerous behaviors such as leaning out of the window, and occupant identity (such as adult, child, or identification mark, etc.) obtained through various sensing devices such as cabin cameras, millimeter-wave radar, seat pressure sensors, and infrared sensors.

[0055] To ensure the accuracy and reliability of subsequent data processing, feature extraction, and model inference, the collected driving environment data and occupant posture data need to be preprocessed. This preprocessing may include, but is not limited to: denoising the collected data to eliminate invalid data caused by environmental interference, equipment errors, and other factors; normalizing the denoised data to map data of different dimensions and units to the same numerical range, avoiding the impact of data magnitude differences on subsequent calculations; and performing time synchronization processing on the simultaneously collected multi-source data to ensure a one-to-one correspondence between the driving environment data and occupant posture data in the time dimension, guaranteeing data synergy and effectiveness.

[0056] Step S202: Determine the current driving scene characteristics of the target vehicle based on driving environment data and occupant posture data.

[0057] Based on the acquired and preprocessed driving environment data and occupant posture data, feature analysis and classification are performed on the multi-source fusion data. By combining the two dimensions of environmental conditions and occupant status, the current driving scene features corresponding to the target vehicle in real time are extracted and determined.

[0058] The driving scenario features include environmental scenario features and occupant status features. Environmental scenario features can be identified and categorized based on driving environment data, specifically including but not limited to typical environments and driving conditions such as sunny days, rainy days, high-temperature environments, high-speed driving, and urban congestion driving. Occupant status features can be obtained by analyzing occupant posture data, specifically including but not limited to the actual occupant status types in the cabin such as single-person driving, multiple-person riding, presence of children, and occupant resting postures. This forms a comprehensive and refined representation of driving scenario features, providing accurate feature input basis for subsequent window opening model inference.

[0059] Step S203: Input the current driving scene features into the pre-trained window opening inference model to obtain the initial window opening of the target vehicle output by the window opening inference model; wherein, the window opening inference model is trained based on multiple historical driving scene features and their corresponding manual window openings.

[0060] The window opening inference model is an intelligent inference model that has been pre-trained offline and deployed. This model uses massive amounts of historical driving scenario features as input samples and the manual window openings actually performed by the user in corresponding historical scenarios as label samples. It completes iterative training and parameter optimization through supervised learning, fully learning the inherent mapping relationship between environmental features, posture features, and user window adjustment habits. After the model training converges, it can perform autonomous calculations and intelligent inference based on real-time input features of the current driving scenario, adaptively outputting an initial window opening that matches the current driving scenario and user preferences. This achieves data-driven opening prediction independent of manually set rules, providing a basic opening basis for subsequent safety corrections and precise window control.

[0061] Step S204: Based on the safety constraints corresponding to the target vehicle, the initial window opening is corrected to obtain the final window opening of the target vehicle.

[0062] The safety constraints here are window opening limits determined based on the target vehicle's real-time driving conditions and occupant status. Their core purpose is to balance the comfort of window adjustment with driving safety. Specific application scenarios may include, but are not limited to, the following two categories: Firstly, dangerous posture detection adapts to various scenarios: real-time analysis of occupant posture data. If dangerous postures such as occupants leaning out of the window or extending body parts out of the window are detected, or if potentially dangerous operations such as child occupants accidentally touching window controls or attempting to open the window are detected, the forced window closing or window opening restriction strategy in the safety constraints can be triggered to urgently correct the initial window opening, limit the window opening range, or directly close the window to avoid safety risks such as occupants falling or their bodies being disturbed by external factors.

[0063] Secondly, high-risk operating conditions: For special operating conditions such as high-speed driving, heavy rain, direct sunlight, and poor air quality, the preset safe opening standards in the safety constraints are prioritized, and the initial window opening is strictly corrected. For example, in high-speed driving conditions, the window opening is corrected to within the preset safe threshold to avoid problems such as increased wind resistance, noise interference, and foreign object intrusion caused by excessive window opening; in heavy rain conditions, the window opening is appropriately reduced or the windows are closed to prevent rainwater from seeping into the cabin; in direct sunlight conditions, a reasonable opening can be maintained to balance ventilation and sunshade, achieving a dual balance of comfort and safety.

[0064] Through the above safety constraint correction process, the limitations of the initial window opening that only focuses on scenario adaptation and user preferences can be effectively compensated, ensuring that the final output window opening not only conforms to user habits but also meets driving safety requirements, thus protecting the personal safety of passengers and the safety of vehicle operation.

[0065] Step S205: Control the windows of the target vehicle according to the final window opening.

[0066] Based on the final window opening degree, precise control commands are sent to the window actuator of the target vehicle to complete the adaptive lifting and lowering adjustment of the window, achieving intelligent and safe control of the window opening degree. Simultaneously, an anti-pinch algorithm can be linked to ensure that the entire adjustment process is smooth, safe, and reliable, balancing driving comfort and safety.

[0067] Based on the aforementioned technical content, this embodiment of the application simultaneously collects driving environment data and occupant posture data of the target vehicle, comprehensively integrates multi-dimensional information to accurately identify the characteristics of the current driving scenario, and relies on a window opening reasoning model built through training on massive historical driving scenario characteristics and manually generated window opening samples to intelligently reason and match an initial window opening suitable for the current scenario, conforming to the user's actual usage habits and scenario adaptation needs; simultaneously, it combines vehicle safety constraints to legally correct the initial window opening, avoiding driving safety hazards and environmental interference caused by excessive window opening and closing, and finally achieves automated and precise window control based on the corrected final window opening. This embodiment can achieve autonomous perception, intelligent decision-making, and safety constraints of window opening, improving cabin comfort, safety, and intelligence levels.

[0068] In addition, to improve the objectivity and accuracy of scene feature recognition, this application embodiment considers using clustering to identify scene features. Simultaneously, it combines an incremental learning mechanism to enable the model to adaptively adapt to users' long-term driving habits and various complex and diverse driving scenarios, continuously iterating and optimizing model parameters to improve the accuracy and adaptability of window opening inference results.

[0069] Figure 3 A schematic flowchart of a window control method provided in another embodiment of this application is shown below. Figure 3 As shown, the method includes: Step S301: Obtain the driving environment data and occupant posture data of the target vehicle.

[0070] For details on how to implement this step, please refer to [link / reference]. Figure 2 The descriptions in the embodiments will not be repeated in this embodiment.

[0071] Step S302: Obtain the pre-stored clustering results of multiple historical driving environment data and historical occupant posture data of the target vehicle; wherein each clustering result corresponds to a historical driving scene feature; determine the current driving scene feature based on the driving environment data and occupant posture data, as well as the driving scene feature corresponding to each clustering result.

[0072] Here, clustering operations are performed on massive historical driving environment data and historical occupant posture data in advance. Driving scenarios with highly similar data distribution patterns, environmental conditions and occupant states are classified into the same category, and standardized driving scenario features are uniformly assigned to each clustering result. For example, the driving scenario features of a certain clustering result are "heavy rain, high altitude, smooth traffic, low temperature, single person, driving". In this way, a binding relationship between clustering results and fixed driving scenario features is established.

[0073] Furthermore, the real-time collected driving environment data and occupant posture data are used to perform distance calculations with the cluster centers of each clustering result to quantify the matching similarity of the current real-time data relative to each cluster center; the clustering result with the highest matching similarity value (e.g., 90%) is selected as the target clustering result, and then the scene feature label bound to the target clustering result is retrieved as the current driving scene feature of the target vehicle.

[0074] In one possible implementation, if the matching similarity of all clustering results does not reach a preset similarity threshold, the current driving condition is determined to be a completely new and unknown scenario. In this case, a preset number of clustering results can be selected in descending order of matching similarity, and the average value of the corresponding historical manual window opening can be calculated and used as a temporary window opening control value. At the same time, the data under this new scenario can be included in the incremental training set for subsequent model iteration and optimization, realizing autonomous learning and gradual adaptation of new scenarios, and continuously improving the coverage of the scenario clustering system and window control strategy.

[0075] Compared to single-dimensional feature judgment or rule-based classification, clustering can reduce the bias and misjudgment of scene feature recognition by mining the inherent correlation and scene distribution patterns of multi-source data. At the same time, clustering analysis can eliminate invalid samples, providing accurate and reliable input for subsequent window opening model inference.

[0076] Step S303: Input the current driving scene features into the pre-trained window opening inference model to obtain the initial window opening of the target vehicle output by the window opening inference model; wherein, the window opening inference model is trained based on multiple historical driving scene features and their corresponding manual window openings.

[0077] For details on how to implement this step, please refer to [link / reference]. Figure 2 The descriptions in the embodiments will not be repeated in this embodiment.

[0078] Step S304: Determine the degree of danger of the current driving condition based on the driving environment data, and determine the degree of danger of the occupant posture based on the occupant posture data; combine the degree of danger of the current driving condition and the degree of danger of the occupant posture to determine the safety constraints corresponding to the target vehicle.

[0079] For example, a first window opening limit can be determined based on the degree of danger of the current driving conditions, and a second window opening limit can be determined based on the degree of danger of the occupant's posture. The first window opening limit is negatively correlated with the degree of danger of the current driving conditions, and the second window opening limit is negatively correlated with the degree of danger of the occupant's posture. The safety constraints corresponding to the target vehicle are determined based on the minimum value between the first and second window opening limits.

[0080] For example, corresponding hazard levels are pre-configured for various typical environmental characteristics (such as vehicle speed, rainfall, air quality, tunnels / construction sections, etc.). After real-time identification of the current driving environment characteristics, the maximum hazard level corresponding to each environmental characteristic is used as the overall hazard level of the driving condition, and the first window opening limit is determined accordingly. This limit is negatively correlated with the hazard level of the driving condition; that is, the higher the risk of the condition, the smaller the maximum allowable window opening. At the same time, corresponding hazard levels are preset for various occupant states in the cabin (such as leaning out of the window, child occupants, abnormal body postures, etc.). After identifying the occupant state, the maximum hazard level corresponding to each occupant state is used as the overall hazard level of the occupant posture, and the second window opening limit is determined accordingly. This limit is also negatively correlated with the hazard level of the occupant posture; the higher the occupant posture risk, the smaller the maximum allowable window opening. Finally, the minimum value between the first window opening limit and the second window opening limit is selected as the comprehensive safety constraint for the target vehicle.

[0081] This embodiment quantifies the degree of danger of driving conditions by using real-time driving environment data, and accurately assesses the degree of danger of occupant posture based on occupant posture data. It integrates the two-dimensional danger levels to generate dynamic safety constraints that are adapted to the current driving state, replacing fixed safety threshold settings. It can dynamically adjust the constraint boundaries to fit real-time road conditions, external environment and actual occupant posture, making the safety constraints more real-time, targeted and comprehensive, effectively making up for the poor adaptability of static safety rules.

[0082] Step S305: Based on the safety constraints corresponding to the target vehicle, the initial window opening is corrected to obtain the final window opening of the target vehicle.

[0083] This step may include: if the initial window opening is greater than the window opening limit corresponding to the safety constraint, then the window opening limit is used as the final window opening; if the initial window opening is less than or equal to the window opening limit, then the initial window opening is used as the final window opening.

[0084] Step S306: Control the windows of the target vehicle according to the final window opening.

[0085] For details on how to implement this step, please refer to [link / reference]. Figure 2 The descriptions in the embodiments will not be repeated in this embodiment.

[0086] Step S307: Obtain the manual window opening degree corresponding to the current driving scene features; store the current driving scene features and their corresponding manual window opening degrees as an incremental training set; after meeting the preset incremental training start conditions, incrementally train the window opening degree inference model through the incremental training set to update the window opening degree inference model.

[0087] In this embodiment, after the vehicle completes adaptive window control based on the final window opening, if the currently automatically adjusted window opening does not meet the user's actual driving needs, the user can manually adjust the opening. This solution can simultaneously collect the current driving scene features and the user's adjusted actual manual window opening, linking and storing the two as an incremental training set. When preset trigger conditions are met, such as reaching a fixed running time or the cumulative amount of newly added sample data reaching a specified threshold, the system automatically retrieves the incremental training set to perform incremental iterative training on the window opening inference model, dynamically updating the model parameters and feature mapping relationship. This allows for updating the user's personalized adjustment habits and scene adaptation preferences, enabling the model to continuously learn the optimal opening rules, optimize output results, achieve autonomous iterative upgrades of the model, and continuously improve the accuracy and fit of window opening inference in subsequent different driving scenarios.

[0088] In some embodiments, historical driving scene features and their corresponding manual window opening degrees can be classified by user, and a window opening degree inference model can be trained for each user. In practical applications: by obtaining the identity identifier of the occupant (e.g., facial recognition result), based on the user identity identifier corresponding to each window opening degree inference model, the window opening degree inference model corresponding to the occupant identity identifier is determined from multiple window opening degree inference models. The current driving scene features are input into the window opening degree inference model corresponding to the occupant identity identifier to obtain the initial window opening degree of the target vehicle.

[0089] This embodiment identifies occupant identities and differentiates them with dedicated window opening inference models, allocating model resources independently for different drivers and passengers. Based on customized models corresponding to each identity, it can deeply learn the long-term window adjustment habits, preferences, and personalized usage patterns of different users. Combining these with the characteristics of the current driving scenario, it infers and outputs an initial window opening that aligns with individual usage habits, effectively avoiding the shortcomings of general models that lack adaptability and cannot accommodate individual preferences. This embodiment achieves personalized customization of window opening inference, significantly improving the humanization and adaptability of intelligent control strategies. It makes the automatic window adjustment effect more suitable for the driving and riding experience of different occupants, further optimizing the cabin riding experience on the basis of intelligent control and enhancing the refined service capabilities of the vehicle's intelligent cockpit.

[0090] Figure 4 This is a schematic diagram of the overall logic flow of a window control method provided in another embodiment of this application, as shown below. Figure 4 As shown, the method includes: (1) Power-on initialization After the system is powered on, it first executes the initialization process, loads the pre-built historical scene sample library and the trained window opening inference model, and completes the initialization configuration of each vehicle sensor, controller and actuator to ensure that the system enters the ready state.

[0091] (2) Real-time data acquisition and preprocessing The system collects driving environment data and occupant posture data of the target vehicle at preset intervals. The environmental data can be further integrated with roadside information to expand the environmental perception dimension, providing more comprehensive external information support for opening decisions. After collection, the multi-source data undergoes preprocessing operations such as denoising, time synchronization, and normalization to eliminate noise interference and dimensional differences, providing high-quality input data for subsequent scene matching and model inference.

[0092] (3) Scene matching and opening degree decision The real-time data is compared with the cluster centers of each scene in the sample library. When the matching similarity is greater than a preset threshold (e.g., 90%), it is considered a successful match. The window opening inference model is called to directly output the initial window opening adapted to the current scene. If the matching similarity does not reach the threshold, it is considered a new scene. At this time, a weighted calculation is performed based on several scenes with the highest similarity ranking in the sample library to generate a temporary window opening, and the current working condition is marked as a new scene to be learned.

[0093] (4) Safety constraint verification The initial window opening obtained from the model output or weighted calculation is validated for safety constraints. Based on the safety constraints corresponding to the target vehicle, the initial window opening is corrected to obtain the final window opening of the target vehicle, prioritizing the safety of passengers.

[0094] (5) Adjustment of window opening Based on the final window opening after safety verification, control commands are sent to the window drive motor and transmission mechanism to drive each window to independently complete the lifting and lowering adjustment action.

[0095] (6) The system continuously monitors whether the user performs manual opening correction. If the user performs manual intervention, the current driving scene features are associated with the user's corrected actual opening and stored in the incremental database. When the preset trigger conditions are met (such as the number of new samples reaching the target or the fixed running cycle being reached), the system automatically starts the incremental learning process, uses the new samples to iteratively train the model, and dynamically updates the model weight parameters so that the model can continuously adapt to the user's personalized usage habits and new scenarios.

[0096] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0097] Figure 5 This is a schematic diagram of the structure of a window control device provided in one embodiment of this application. Figure 5 As shown, the window control device 50 provided in this embodiment may include: The acquisition module 51 is used to acquire driving environment data and occupant posture data of the target vehicle; The determination module 52 is used to determine the current driving scene characteristics of the target vehicle based on driving environment data and occupant posture data; The inference module 53 is used to input the current driving scene features into the pre-trained window opening inference model to obtain the initial window opening of the target vehicle output by the window opening inference model; wherein, the window opening inference model is trained based on multiple historical driving scene features and their corresponding manual window openings. The correction module 54 is used to correct the initial window opening based on the safety constraints corresponding to the target vehicle, so as to obtain the final window opening of the target vehicle. The control module 55 is used to control the windows of the target vehicle based on the final window opening.

[0098] In one possible implementation, the determining module 52 is used for: Obtain the clustering results of pre-stored clustering of multiple historical driving environment data and historical occupant posture data of the target vehicle; where each clustering result corresponds to a historical driving scene feature. Based on driving environment data, occupant posture data, and driving scene characteristics corresponding to each clustering result, the current driving scene characteristics are determined.

[0099] In one possible implementation, the determining module 52 is used for: Based on driving environment data, occupant posture data, and the results of each cluster, the matching similarity is determined; Based on the matching similarity, the target clustering result is determined from the various clustering results; The driving scene features corresponding to the target clustering results are used as the current driving scene features of the target vehicle.

[0100] In one possible implementation, before correcting the initial window opening based on the safety constraints corresponding to the target vehicle to obtain the final window opening of the target vehicle, the correction module 54 is further configured to: Based on driving environment data, determine the degree of danger of the current driving condition, and based on occupant posture data, determine the degree of danger of the occupant posture. Based on the degree of danger of the current driving conditions and the degree of danger of the occupants' posture, determine the corresponding safety constraints for the target vehicle.

[0101] In one possible implementation, the correction module 54 is used for: Based on the degree of danger of the current driving conditions, a first window opening limit is determined, and based on the degree of danger of the occupant's posture, a second window opening limit is determined; wherein, the first window opening limit is negatively correlated with the degree of danger of the current driving conditions, and the second window opening limit is negatively correlated with the degree of danger of the occupant's posture. The safety constraints corresponding to the target vehicle are determined based on the minimum value between the first window opening limit and the second window opening limit.

[0102] In one possible implementation, the correction module 54 is used for: If the initial window opening is greater than the window opening limit corresponding to the safety constraint, then the window opening limit will be used as the final window opening. If the initial window opening is less than or equal to the window opening limit, the initial window opening will be used as the final window opening.

[0103] In one possible implementation, there are multiple window opening inference models, each of which corresponds to a user identity identifier. Different window opening inference models correspond to different user identity identifiers. Inference module 53 is used for: Obtain occupant identification; Based on the user identity identifier corresponding to each window opening inference model, determine the window opening inference model corresponding to the occupant's identity identifier from multiple window opening inference models; By inputting the current driving scene characteristics into the window opening inference model corresponding to the occupant's identity, the initial window opening of the target vehicle can be obtained.

[0104] In one possible implementation, after controlling the windows of the target vehicle based on the final window opening, the acquisition module 51 is further used for: Obtain the manual window opening degree corresponding to the current driving scene characteristics; Store the current driving scene features and their corresponding manual window opening degrees as an incremental training set; After the preset incremental training start conditions are met, the window opening inference model is incrementally trained using the incremental training set to update the window opening inference model.

[0105] In one possible implementation, before inputting the current driving scene features into a pre-trained window opening inference model to obtain the initial window opening of the target vehicle output by the window opening inference model, the inference module 53 is further used for: Acquire features of multiple historical driving scenarios and their corresponding manual window opening degrees, and construct a training sample set based on the features of multiple historical driving scenarios and their corresponding manual window opening degrees; The driving scene features in the training sample set are used as the input of the preset initial model, and the corresponding manual window opening degree is used as the model output label. The model parameters of the preset initial model are iteratively trained and updated until the preset convergence condition is met, and the window opening degree inference model is obtained.

[0106] Based on the aforementioned technical content, this embodiment of the application simultaneously collects driving environment data and occupant posture data of the target vehicle, comprehensively integrates multi-dimensional information to accurately identify the characteristics of the current driving scenario, and relies on a window opening reasoning model built through training on massive historical driving scenario characteristics and manually generated window opening samples to intelligently reason and match an initial window opening suitable for the current scenario, conforming to the user's actual usage habits and scenario adaptation needs; simultaneously, it combines vehicle safety constraints to legally correct the initial window opening, avoiding driving safety hazards and environmental interference caused by excessive window opening and closing, and finally achieves automated and precise window control based on the corrected final window opening. This embodiment can achieve autonomous perception, intelligent decision-making, and safety constraints of window opening, improving cabin comfort, safety, and intelligence levels.

[0107] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0108] Figure 6 This is a schematic diagram of the structure of a vehicle provided in one embodiment of this application. Figure 6As shown, the vehicle 600 in this embodiment includes a processor 610 and a memory 620, wherein the memory 620 stores a computer program 621 that can run on the processor 610. When the processor 610 executes the computer program 621, it implements the steps in any of the above method embodiments. Alternatively, when the processor 610 executes the computer program 621, it implements the functions of each module in the above device embodiments.

[0109] For example, computer program 621 may be divided into one or more modules / units, one or more of which are stored in memory 620 and executed by processor 610 to complete this application. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of computer program 621 in vehicle 600.

[0110] Those skilled in the art will understand that Figure 6 This is merely an example of a vehicle and does not constitute a limitation on the vehicle. It may include more or fewer components than shown, or combinations of certain components, or different components, such as input / output devices, network access devices, buses, etc.

[0111] The processor 610 can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0112] The memory 620 can be an internal storage unit of the vehicle, such as a hard drive or memory, or an external storage device, such as a plug-in hard drive, smart media card (SMC), secure digital (SD) card, flash card, etc. The memory 620 can also include both internal and external storage devices. The memory 620 is used to store computer programs and other programs and data required by the vehicle. The memory 620 can also be used to temporarily store data that has been output or will be output.

[0113] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments 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. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0114] An embodiment of this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described window control method.

[0115] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0116] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0117] In the embodiments provided in this application, it should be understood that the disclosed devices / vehicles and methods can be implemented in other ways. For example, the device / vehicle embodiments described above are merely illustrative. For instance, the division of modules or units is only 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 mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0118] 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 network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0119] 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.

[0120] If the integrated module / 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, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0121] The above-described 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, and should all be included within the protection scope of this application.

Claims

1. A method for controlling vehicle windows, characterized in that, include: Acquire driving environment data and occupant posture data of the target vehicle; Based on the driving environment data and the occupant posture data, the current driving scene characteristics of the target vehicle are determined; The current driving scene features are input into a pre-trained window opening inference model to obtain the initial window opening of the target vehicle output by the window opening inference model; wherein, the window opening inference model is trained based on multiple historical driving scene features and their corresponding manual window openings. Based on the safety constraints corresponding to the target vehicle, the initial window opening is corrected to obtain the final window opening of the target vehicle; Based on the final window opening, the target vehicle's windows are controlled.

2. The vehicle window control method according to claim 1, characterized in that, The step of determining the current driving scene characteristics of the target vehicle based on the driving environment data and the occupant posture data includes: Obtain the clustering results of pre-stored clustering of multiple historical driving environment data and historical occupant posture data of the target vehicle; wherein each clustering result corresponds to a historical driving scene feature; Based on the driving environment data, the occupant posture data, and the driving scene features corresponding to each clustering result, the current driving scene features are determined.

3. The window control method according to claim 2, characterized in that, The step of determining the current driving scene features based on the driving environment data, the occupant posture data, and the driving scene features corresponding to each clustering result includes: Based on the driving environment data, the occupant posture data, and each clustering result, the matching similarity is determined; Based on the matching similarity, the target clustering result is determined from each clustering result; The driving scene features corresponding to the target clustering results are used as the current driving scene features of the target vehicle.

4. The vehicle window control method according to any one of claims 1-3, characterized in that, Before correcting the initial window opening based on the safety constraints corresponding to the target vehicle to obtain the final window opening of the target vehicle, the method further includes: Based on the driving environment data, the degree of danger of the current driving condition is determined, and based on the occupant posture data, the degree of danger of the occupant posture is determined. Based on the degree of danger of the current driving conditions and the degree of danger of the occupant posture, the safety constraints corresponding to the target vehicle are determined.

5. The window control method according to claim 4, characterized in that, The determination of the safety constraints corresponding to the target vehicle by combining the degree of danger of the current driving condition and the degree of danger of the occupant posture includes: Based on the degree of danger of the current driving condition, a first window opening limit is determined, and based on the degree of danger of the occupant's posture, a second window opening limit is determined; wherein, the first window opening limit is negatively correlated with the degree of danger of the current driving condition, and the second window opening limit is negatively correlated with the degree of danger of the occupant's posture; The safety constraints corresponding to the target vehicle are determined based on the minimum value between the first window opening limit and the second window opening limit.

6. The vehicle window control method according to any one of claims 1-3, characterized in that, The step of correcting the initial window opening based on the safety constraints corresponding to the target vehicle to obtain the final window opening of the target vehicle includes: If the initial window opening is greater than the window opening limit corresponding to the safety constraint, then the window opening limit is taken as the final window opening. If the initial window opening is less than or equal to the window opening limit, then the initial window opening is taken as the final window opening.

7. The vehicle window control method according to any one of claims 1-3, characterized in that, There are multiple window opening inference models, and each window opening inference model corresponds to a corresponding user identity identifier. Different window opening inference models correspond to different user identity identifiers. The step of inputting the current driving scene features into a pre-trained window opening inference model to obtain the initial window opening of the target vehicle output by the window opening inference model includes: Obtain the identity identifier of the occupant; Based on the user identity identifier corresponding to each window opening inference model, determine the window opening inference model corresponding to the occupant's identity identifier from multiple window opening inference models; The current driving scene features are input into the window opening inference model corresponding to the occupant's identity to obtain the initial window opening of the target vehicle.

8. The vehicle window control method according to any one of claims 1-3, characterized in that, After controlling the windows of the target vehicle based on the final window opening, the method further includes: Obtain the manual window opening degree corresponding to the current driving scene characteristics; Store the current driving scene features and their corresponding manual window opening degrees as an incremental training set; After the preset incremental training start conditions are met, the window opening inference model is incrementally trained using the incremental training set to update the window opening inference model.

9. The vehicle window control method according to any one of claims 1-3, characterized in that, Before inputting the current driving scene features into a pre-trained window opening inference model to obtain the initial window opening of the target vehicle output by the window opening inference model, the method further includes: Acquire multiple historical driving scene features and their corresponding manual window opening degrees, and construct a training sample set based on the multiple historical driving scene features and their corresponding manual window opening degrees; Using the driving scene features in the training sample set as the input of the preset initial model, and the corresponding manual window opening degree as the model output label, the model parameters of the preset initial model are iteratively trained and updated until the preset convergence condition is met, thus obtaining the window opening degree inference model.

10. A vehicle comprising a memory and a processor, the memory storing a computer program executable on the processor, characterized in that, When the processor executes the computer program, it implements the window control method as described in any one of claims 1-9.

Images