Adjusting method and device of vehicle-mounted air conditioner, vehicle and storage medium

Abstract

The invention relates to an adjusting method and device for a vehicle-mounted air conditioner, a vehicle and a storage medium, and the method comprises the steps that image data and physiological data of a user under the current driving working condition, and the environment temperature inside the vehicle and environment data outside the vehicle under the current driving working condition are input into a pre-constructed comfort degree prediction model; the method comprises the following steps: obtaining the sendible temperature characteristics of a user, determining the sendible temperature state of the user and a temperature adjusting parameter corresponding to the sendible temperature state, and controlling a vehicle-mounted air conditioner of a vehicle to execute a corresponding temperature adjusting action based on the temperature adjusting parameter until the sendible temperature state meets a preset comfort condition. Therefore, the technical problems that in the related technology, the adjusting mode of the vehicle-mounted air conditioner lacks predictability, the situation that the temperature in the vehicle is not matched with the sensible temperature of the user after adjustment easily occurs, and the automatic adjusting parameters of the vehicle-mounted air conditioner gradually deviate from the comfort interval of the user due to the fact that the adjusting parameters without characterization are easily brought into the learning range in a unified mode are solved.

Description

Technical Field

[0001] This application relates to the field of vehicle technology, and in particular to a method, device, vehicle, and storage medium for regulating vehicle air conditioning. Background Technology

[0002] In related technologies, the adjustment method of vehicle air conditioning lacks predictability. It only determines the adjustment parameters by mapping the environment to a pre-set target temperature based on data collected by physical sensors such as in-vehicle and out-of-vehicle temperature sensors. This can easily lead to a mismatch between the adjusted in-vehicle temperature and the user's perceived temperature, forcing the user to manually adjust the temperature according to their own body temperature. Furthermore, in some cases, such as when the user is sick, the perceived temperature will be lower than usual. In this case, the change in the user's perceived temperature is not representative. The vehicle's optimization of adjustment parameters by memorizing and learning the user's adjustment parameters can easily cause the final automatic adjustment parameters of the vehicle air conditioning to gradually deviate from the user's comfort range, ultimately affecting the user's actual user experience. This urgently needs to be optimized. Summary of the Invention

[0003] This application provides a method, device, vehicle, and storage medium for adjusting a vehicle air conditioner, in order to solve the technical problems in the related art where the adjustment method of a vehicle air conditioner lacks predictability, easily leads to a mismatch between the temperature inside the vehicle and the user's perceived temperature after adjustment, and when learning the user's manual adjustment parameters, it is easy to include non-characteristic adjustment parameters in the learning scope, causing the automatic adjustment parameters of the vehicle air conditioner to gradually deviate from the user's comfort range, thus affecting the user's actual user experience.

[0004] The first aspect of this application provides a method for adjusting an in-vehicle air conditioner, comprising the following steps: acquiring image data and physiological data of a user under current driving conditions; obtaining the user's perceived temperature characteristics by combining the image data, the physiological data, the ambient temperature inside the vehicle under the current driving conditions, and the ambient temperature outside the vehicle; determining the user's perceived temperature state and a temperature adjustment parameter corresponding to the perceived temperature state based on the perceived temperature characteristics; and controlling the in-vehicle air conditioner of the vehicle to perform corresponding temperature adjustment actions based on the temperature adjustment parameters until the perceived temperature state meets preset comfort conditions.

[0005] Optionally, in one embodiment of this application, obtaining the user's perceived temperature characteristics from the image data, the physiological data, the ambient temperature inside the vehicle under the current driving condition, and the ambient temperature outside the vehicle includes: identifying human body key points in each frame of the image data; determining the user's behavioral feature vector based on the human body key points and the temporal information of each frame of the image; acquiring the user's physiological parameter change trend and fluctuation data under the current driving condition based on the physiological information, and determining the user's physiological feature vector based on the physiological parameter change trend and the fluctuation data; acquiring the ambient temperature and solar radiation intensity outside the vehicle based on the ambient temperature outside the vehicle; determining the vehicle's environmental feature vector by combining the ambient temperature inside the vehicle, the ambient temperature outside the vehicle, the solar radiation intensity, and the vehicle speed; and obtaining the perceived temperature characteristics by combining the behavioral feature vector, the physiological feature vector, and the environmental feature vector.

[0006] Optionally, in one embodiment of this application, the step of combining the behavioral feature vector, the physiological feature vector, and the environmental feature vector to obtain the perceived temperature feature includes: concatenating the behavioral feature vector, the physiological feature vector, and the environmental feature vector to obtain a comprehensive feature vector; performing dimensionality reduction processing on the comprehensive feature vector to obtain a feature vector to be processed; splitting the feature vector to be processed using a preset time step, and assigning corresponding weights to the feature vector to be processed at each time step using a temporal attention mechanism; and calculating a perceived temperature feature vector by combining the weights and the feature vector to be processed at each time step, so as to obtain the perceived temperature feature based on the perceived temperature feature vector.

[0007] Optionally, in one embodiment of this application, determining the user's perceived temperature state and the temperature adjustment parameter corresponding to the perceived temperature state based on the perceived temperature characteristics includes: converting the perceived temperature characteristics into a perceived probability distribution to obtain the perceived temperature state based on the perceived probability distribution; mapping the perceived temperature characteristics to a scalar of adjustment intensity that satisfies a preset continuity condition using a preset mapping rule; and combining the scalar of adjustment intensity and the perceived temperature state to obtain the temperature adjustment parameter.

[0008] Optionally, in one embodiment of this application, the method further includes: obtaining the actual temperature adjustment parameters of the vehicle air conditioner; and optimizing the preset mapping rule based on the actual temperature adjustment parameters.

[0009] Optionally, in one embodiment of this application, before controlling the vehicle's air conditioning system to perform a corresponding temperature adjustment action based on the temperature adjustment parameters, the method further includes: responding to the temperature adjustment parameters being within a preset safe adjustment limit range, controlling the vehicle's air conditioning system to perform a corresponding temperature adjustment action based on the temperature adjustment parameters; and responding to the temperature adjustment parameters not being within the preset safe adjustment limit range, determining a target temperature adjustment parameter for the vehicle's air conditioning system based on the preset safe adjustment limit range, so as to control the vehicle's air conditioning system to perform a corresponding temperature adjustment action using the target temperature adjustment parameter.

[0010] A second aspect of this application provides an adjustment device for an in-vehicle air conditioner, comprising: a first acquisition module for acquiring image data and physiological data of a user under current driving conditions; a calculation module for calculating the user's perceived temperature characteristics by combining the image data, the physiological data, the ambient temperature inside the vehicle under the current driving conditions, and the ambient temperature outside the vehicle; and an adjustment module for determining the user's perceived temperature state and a temperature adjustment parameter corresponding to the perceived temperature state based on the perceived temperature characteristics, and controlling the in-vehicle air conditioner of the vehicle to perform corresponding temperature adjustment actions based on the temperature adjustment parameters until the perceived temperature state meets preset comfort conditions.

[0011] Optionally, in one embodiment of this application, the calculation module includes: a recognition unit, configured to recognize human key points in each frame of the image data; a first determination unit, configured to determine the user's behavioral feature vector based on the human key points and the temporal information of each frame of the image; a second determination unit, configured to acquire the user's physiological parameter change trend and fluctuation data under the current driving condition based on the physiological information, and determine the user's physiological feature vector based on the physiological parameter change trend and the fluctuation data; a first acquisition unit, configured to acquire the ambient temperature and solar radiation intensity outside the vehicle based on the vehicle external environment data; a third determination unit, configured to determine the vehicle's environmental feature vector by combining the ambient temperature inside the vehicle, the ambient temperature outside the vehicle, the solar radiation intensity, and the vehicle speed; and a second acquisition unit, configured to obtain the perceived temperature feature by combining the behavioral feature vector, the physiological feature vector, and the environmental feature vector.

[0012] Optionally, in one embodiment of this application, the second acquisition unit includes: a splicing subunit, used to splice the behavioral feature vector, the physiological feature vector, and the environmental feature vector to obtain a comprehensive feature vector; a processing subunit, used to perform dimensionality reduction processing on the comprehensive feature vector to obtain a feature vector to be processed; a splitting subunit, used to split the feature vector to be processed using a preset time step, and to assign corresponding weights to the feature vector to be processed at each time step using a temporal attention mechanism; and a first calculation subunit, used to combine the weights and the feature vector to be processed at each time step to calculate a body temperature feature vector, so as to obtain the body temperature feature based on the body temperature feature vector.

[0013] Optionally, in one embodiment of this application, the adjustment module includes: a conversion unit, configured to convert the perceived temperature feature into a perceived probability distribution, so as to obtain the perceived temperature state based on the perceived probability distribution; a mapping unit, configured to map the perceived temperature feature into an adjustment intensity scalar that satisfies a preset continuity condition using a preset mapping rule; and a third acquisition unit, configured to combine the adjustment intensity scalar and the perceived temperature state to obtain the temperature adjustment parameter.

[0014] Optionally, in one embodiment of this application, it further includes: a second acquisition module, used to acquire the actual temperature adjustment parameters of the vehicle air conditioner; and an optimization module, used to optimize the preset mapping rule based on the actual temperature adjustment parameters.

[0015] Optionally, in one embodiment of this application, it further includes: a first control module, configured to control the vehicle's air conditioning system to perform a corresponding temperature adjustment action based on the temperature adjustment parameter when the temperature adjustment parameter is within a preset safe adjustment limit range; and a second control module, configured to determine a target temperature adjustment parameter for the vehicle air conditioning system based on the preset safe adjustment limit range when the temperature adjustment parameter is not within the preset safe adjustment limit range, so as to control the vehicle's air conditioning system to perform a corresponding temperature adjustment action using the target temperature adjustment parameter.

[0016] A third aspect of this application provides a vehicle, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the vehicle air conditioning adjustment method as described in the above embodiments.

[0017] A fourth aspect of this application provides a computer-readable storage medium storing computer instructions for causing the computer to perform the vehicle air conditioning adjustment method as described in the above embodiments.

[0018] A fifth aspect of this application provides a computer program product, including a computer program, which, when executed, is used to implement the above-described method for adjusting a vehicle air conditioner.

[0019] This application embodiment can input the user's image data, physiological data, and the ambient temperature inside and outside the vehicle under the current driving conditions into a pre-built comfort prediction model. By fusing visual behavior recognition, non-contact physiological parameters, and environmental data inside and outside the vehicle, it predicts the user's perceived temperature characteristics. This proactively triggers the adjustment of the vehicle's air conditioning before the user feels uncomfortable, reducing manual intervention by the user during driving and improving the driving experience. This solves the technical problems in related technologies where the adjustment method of vehicle air conditioning lacks predictability, easily leading to a mismatch between the adjusted interior temperature and the user's perceived temperature. Furthermore, when learning the user's manual adjustment parameters, it is easy to include non-representative adjustment parameters in the learning scope, causing the automatic adjustment parameters of the vehicle air conditioning to gradually deviate from the user's comfort range, affecting the user's actual user experience.

[0020] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0021] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0022] Figure 1 This is a flowchart illustrating a method for adjusting a vehicle air conditioner according to an embodiment of this application; Figure 2 This is a flowchart of a method for adjusting a vehicle air conditioner according to an embodiment of this application; Figure 3 This is a schematic diagram of the structure of an adjustment device for a vehicle air conditioner according to an embodiment of this application; Figure 4 This is a structural schematic diagram of a vehicle provided according to an embodiment of this application. Detailed Implementation

[0023] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0024] The following description, with reference to the accompanying drawings, illustrates an embodiment of a vehicle air conditioning adjustment method, apparatus, vehicle, and storage medium. Addressing the issues raised in the background section regarding the lack of predictability in vehicle air conditioning adjustment methods, which can lead to mismatches between the adjusted interior temperature and the user's perceived temperature, and the tendency to include non-representative adjustment parameters in the learning process when learning manual adjustment parameters, causing the automatic adjustment parameters to gradually deviate from the user's comfort range and negatively impacting the user's actual experience, this application provides a vehicle air conditioning adjustment method. This method inputs the user's image data, physiological data, and the vehicle's interior and exterior environmental data under the current driving conditions into a pre-built comfort prediction model. By fusing visual behavior recognition, non-contact physiological parameters, and in-vehicle and exterior environmental data, the model predicts the user's perceived temperature characteristics. This proactively triggers vehicle air conditioning adjustment before the user experiences discomfort, reducing manual intervention during driving and improving the user's driving experience. This solves the technical problems in related technologies, such as the lack of predictability in the adjustment method of vehicle air conditioning, which easily leads to a mismatch between the temperature inside the vehicle and the temperature felt by the user after adjustment. Furthermore, when learning the user's manual adjustment parameters, it is easy to include non-representative adjustment parameters in the learning scope, causing the automatic adjustment parameters of the vehicle air conditioning to gradually deviate from the user's comfort range and affect the user's actual user experience.

[0025] Understandably, with the rapid improvement of intelligent technology, users are increasingly demanding higher levels of cabin comfort. As a core component for regulating the cabin environment, the level of intelligence of the vehicle's air conditioning system directly affects the driving and riding experience.

[0026] Currently, in related technologies, intelligent adjustment of in-vehicle air conditioning mainly relies on data collected by physical sensors such as in-vehicle and out-of-vehicle temperature sensors, humidity sensors, and sunlight sensors. This data is then used for feedback adjustment through preset control logic (such as setting a target temperature). Some high-end models have introduced user recognition functions that can remember and recall different users' preset temperature preferences.

[0027] Furthermore, related technologies have emerged that utilize multi-source data to achieve adaptive air conditioning adjustment. These solutions acquire users' historical settings, environmental data, clothing information, and even health indicators such as heart rate and blood pressure, perform statistical analysis and fitting to generate personalized target air conditioning parameters. They can also predict the environment based on navigation information and dynamically compensate based on the user's health status.

[0028] However, the relevant technologies have significant limitations. First, the basic control mode is passive and lacks predictability, only responding after environmental parameters change, resulting in adjustment delays and an inability to intervene before the user actually feels discomfort. Second, it cannot perceive the user's real-time physiological state and ignores the dynamic differences in comfort needs influenced by factors such as clothing and activity intensity.

[0029] While more advanced solutions incorporate learning from health indicators and historical preferences, their regulatory decisions heavily rely on statistical analysis of historical data and responses to macroscopic health conditions. These solutions cannot capture and interpret the user's natural, subconscious micro-behaviors during driving (such as rubbing hands or wiping sweat) in real time. Therefore, they struggle to provide timely, proactive, and precise intervention when subtle changes in the user's subjective temperature sensation occur before health indicators become abnormal or manual adjustments are triggered. Furthermore, learning manual adjustment parameters under unconventional conditions can easily affect the compatibility between automatic adjustment parameters under normal conditions and user preferences. Finally, all of the above solutions may still require manual intervention when automatic adjustments fail to meet expectations, which can distract the user and create safety hazards during driving.

[0030] To address the aforementioned technical issues, the vehicle air conditioning adjustment method of this application embodiment can proactively predict the user's perceived physical state, thereby achieving personalized adaptation and providing a better in-vehicle environment.

[0031] Specifically, Figure 1 This is a schematic flowchart illustrating a method for adjusting a vehicle air conditioner according to an embodiment of this application.

[0032] like Figure 1 As shown, the method for adjusting the vehicle air conditioner includes the following steps: In step S101, image data and physiological data of the user under the current driving conditions are acquired.

[0033] It is understandable that users will have different driving mindsets under different driving conditions, which will affect their perceived body temperature. For example, users may experience anxiety when driving in traffic jams, and their body temperature may rise at this time.

[0034] Therefore, the embodiments of this application can combine the current driving conditions, the user's image data and physiological data to comprehensively determine the user's perceived temperature, thereby making more suitable adjustments to the parameters of the vehicle air conditioner.

[0035] In this embodiment of the application, the user's image data under the current driving conditions can be captured by a camera in the vehicle cabin to obtain the user's behavioral characteristics. The user's physiological data, such as facial temperature and heart rate, can be obtained by using infrared sensors, steering wheel sensors, or seat sensors.

[0036] In step S102, the user's perceived temperature characteristics are obtained by combining image data, physiological data, ambient temperature inside the vehicle under the current driving conditions, and ambient temperature outside the vehicle.

[0037] Furthermore, embodiments of this application can construct a comfort prediction model using multimodal feature fusion and attention mechanisms. By fusing data such as image data, physiological data, ambient temperature inside the vehicle under current driving conditions, and ambient temperature outside the vehicle, the model learns the complex nonlinear relationships between different modal features and predicts the user's perceived temperature characteristics, so as to convert the perceived temperature characteristics into adjustment parameters for the vehicle's air conditioning.

[0038] Optionally, in one embodiment of this application, the user's perceived temperature characteristics are obtained by combining image data, physiological data, ambient temperature inside the vehicle under the current driving conditions, and ambient temperature outside the vehicle. This includes: identifying key human body points in each frame of the image data; determining the user's behavioral feature vector based on the key human body points and the temporal information of each frame of the image; acquiring the user's physiological parameter change trends and fluctuation data under the current driving conditions based on physiological information, and determining the user's physiological feature vector based on the physiological parameter change trends and fluctuation data; acquiring the ambient temperature and solar radiation intensity outside the vehicle based on the ambient temperature inside the vehicle, the ambient temperature outside the vehicle, solar radiation intensity, and the vehicle speed to determine the vehicle's environmental feature vector; and combining the behavioral feature vector, physiological feature vector, and environmental feature vector to obtain the perceived temperature characteristics.

[0039] In this embodiment, the user's upper body video stream (continuous frames) captured by an in-cabin camera is utilized. A Lightweight OpenPose or HRNet model pre-trained on a large human pose and motion dataset is employed. This network does not directly process the raw image but receives a heatmap of human key points pre-processed by a preceding algorithm, focusing on key areas such as hands, face, and shoulders. The backbone network outputs a sequence of skeletal key point coordinates for each frame (e.g., hand position, head pose). Subsequently, a one-dimensional convolutional neural network combined with a gated recurrent unit (GRU) is used to perform temporal modeling of this coordinate sequence, capturing dynamic changes in behavior (e.g., the reciprocating motion of rubbing hands), ultimately outputting a fixed-dimensional behavior feature vector V_behavior.

[0040] This application embodiment can acquire facial temperature sequences from infrared sensors and heart rate variability sequences from steering wheel / seat sensors. Two parallel one-dimensional convolutional neural networks are used to extract local features from the temperature and heart rate signals, respectively, and then fed into a shared Bi-directional LSTM (Bidirectional Long Short-Term Memory) network to capture short-term fluctuations and long-term trends of physiological parameters, thereby outputting a physiological feature vector V_physio.

[0041] This application embodiment can acquire scalar data from the vehicle bus, including in-vehicle temperature, outside vehicle temperature, solar radiation intensity, and vehicle speed. The acquired data is then normalized, and a fully connected encoder is used to map the environmental parameters into a high-dimensional environmental feature vector V_env.

[0042] The embodiments of this application can combine the behavioral feature vector V_behavior, the physiological feature vector V_physio, and the environmental feature vector V_env to obtain the perceived temperature features.

[0043] Optionally, in one embodiment of this application, the perceived temperature feature is obtained by combining behavioral feature vectors, physiological feature vectors, and environmental feature vectors, including: concatenating behavioral feature vectors, physiological feature vectors, and environmental feature vectors to obtain a comprehensive feature vector; performing dimensionality reduction processing on the comprehensive feature vector to obtain a feature vector to be processed; splitting the feature vector to be processed using a preset time step, and assigning corresponding weights to the feature vector to be processed at each time step using a temporal attention mechanism; and calculating the perceived temperature feature vector by combining the weights and the feature vector to be processed at each time step, so as to obtain the perceived temperature feature based on the perceived temperature feature vector.

[0044] As one possible implementation, embodiments of this application can concatenate V_behavior, V_physio, and V_env to form a comprehensive multimodal feature vector. Considering that the user's state changes continuously, embodiments of this application can introduce an attention mechanism. This mechanism can automatically learn and weight the importance of features at different time steps, thereby focusing more on the recent state most relevant to the current comfort judgment, rather than a simple averaging.

[0045] Here, the time step feature refers to the multimodal feature corresponding to each discrete time node in a continuous time series. In this embodiment, the user's continuous state data (video stream, physiological signal sequence, environmental parameter time series data) is processed. This continuous data is split into fixed time intervals, and each time step corresponds to a set of features (the multimodal feature concatenated from V_behavior, V_physio, and V_env at that moment). The features of all time steps constitute a time series multimodal feature sequence, and the feature of each time step is the time step feature.

[0046] The embodiments of this application can organize the complete multimodal feature vector after splicing and dimensionality reduction into a temporal feature sequence according to the time dimension, that is, each time step corresponds to a feature vector, which constitutes the temporal data of the model input.

[0047] This application embodiment can map the features of each time step into a low-dimensional attention vector, capture the correlation between the features of that time step and the comfort judgment, and assign corresponding weights to each time step to obtain a multimodal feature vector. The weighted multimodal feature vector is then input into a fully connected layer for deep nonlinear fusion, allowing the model to autonomously learn the correlation rules between the three types of features (such as the hand-rubbing behavior feature contributing more to the perceived cold under low-temperature environmental features). Finally, the output is a perceived temperature feature that integrates multi-dimensional time-series information (this feature is a high-dimensional vector that accurately represents the user's comprehensive perceived temperature state in the current driving environment).

[0048] In step S103, the user's perceived temperature state and the corresponding temperature adjustment parameters are determined based on the perceived temperature characteristics. The vehicle's air conditioning is then controlled to perform corresponding temperature adjustment actions based on the temperature adjustment parameters until the perceived temperature state meets the preset comfort conditions.

[0049] Furthermore, embodiments of this application can utilize the perceived temperature characteristics to determine the user's perceived temperature state, that is, to convert the data into a corresponding state description, such as cold, comfortable, or hot, and quantify the perceived temperature state into the temperature adjustment parameters of the vehicle air conditioner for automatic control of the vehicle air conditioner.

[0050] Optionally, in one embodiment of this application, determining the user's perceived temperature state and the temperature adjustment parameters corresponding to the perceived temperature state based on perceived temperature characteristics includes: converting the perceived temperature characteristics into a perceived probability distribution to obtain the perceived temperature state based on the perceived probability distribution; mapping the perceived temperature characteristics to a scalar of adjustment intensity that satisfies a preset continuity condition using a preset mapping rule; and combining the scalar of adjustment intensity and the perceived temperature state to obtain the temperature adjustment parameters.

[0051] In this embodiment, the Softmax activation function can be used to output a three-class probability distribution based on the perceived temperature characteristics: [P(cold), P(comfortable), P(warm)]. A specific adjustment suggestion intensity scalar value is also output to quantify the adjustment range.

[0052] This application embodiment uses Focal Loss to address the class imbalance problem (because data on comfortable states may be far more abundant than data on cold / hot states). Smooth L1 Loss is used to constrain the predicted adjustment intensity, matching it to the magnitude of subsequent manual adjustment by the user. The total loss is a weighted sum of the classification loss and the regression loss.

[0053] For example, in this embodiment of the application, the perceived temperature features can be mapped to a three-class probability distribution of cold, comfortable, and hot by using the classification output head of the fully connected layer and the Softmax activation function, so as to determine the user's most likely current perceived state (e.g., if P(cold) > 0.8, it is determined that the current perceived temperature is cold). Meanwhile, the embodiments of this application can map the perceived temperature characteristics into a scalar value of quantitative adjustment suggestion intensity through fully connected layers and linear regression in parallel (this scalar value is a continuous value, which is pre-calibrated through training to match the numerical range of air conditioning adjustment, such as ±1℃, ±2℃, air volume level 1-5, etc.), representing the temperature / air volume range that needs to be adjusted.

[0054] Furthermore, in this embodiment, the temperature control parameters can be determined by mapping and matching the scalar value of the quantitative adjustment suggestion based on the body sensation and the adjustment parameter library of the vehicle air conditioner. For example, if it is determined to be too cold and the quantitative adjustment intensity is +2℃, the matched temperature control parameter is to increase the set temperature by 2℃.

[0055] Optionally, in one embodiment of this application, the method further includes: obtaining the actual temperature adjustment parameters of the vehicle air conditioner; and optimizing the preset mapping rules based on the actual temperature adjustment parameters.

[0056] In actual implementation, this application embodiment can continuously monitor the user's behavior after predictive adjustment. If the user does not manually override the adjustment within 2 minutes, the adjustment is recorded as a success case to strengthen the current mapping relationship between the user's perceived temperature and the adjustment parameter. Conversely, if the user manually adjusts the temperature back, it is considered a negative feedback, thereby adjusting the mapping relationship so that different adjustment strategies are adopted in similar situations in the future.

[0057] Optionally, in one embodiment of this application, before controlling the vehicle's air conditioning system to perform the corresponding temperature adjustment action based on the temperature adjustment parameters, the method further includes: responding to the temperature adjustment parameters being within a preset safe adjustment limit range, controlling the vehicle's air conditioning system to perform the corresponding temperature adjustment action based on the temperature adjustment parameters; and responding to the temperature adjustment parameters not being within the preset safe adjustment limit range, determining a target temperature adjustment parameter for the vehicle's air conditioning system based on the preset safe adjustment limit range, so as to control the vehicle's air conditioning system to perform the corresponding temperature adjustment action using the target temperature adjustment parameter.

[0058] In some embodiments, all adjustments to the vehicle's air conditioning system can be set with safety boundaries, i.e., safe adjustment limits. For example, it will not perform large-scale changes in airflow direction or mode when the vehicle is being driven aggressively (such as cornering at high speed). The adjustment speed is also carefully calibrated to avoid sudden strong winds or intense hot or cold stimuli affecting driving safety.

[0059] Combination Figure 2 As shown, an embodiment of the vehicle air conditioning adjustment method of this application will be described.

[0060] This application embodiment can construct a prediction model that integrates multimodal feature fusion, multi-head attention mechanism and dual output head, so as to obtain the user's image data, physiological data and combine the environmental data inside and outside the vehicle, and output the adjustment parameters of the vehicle air conditioner to achieve automatic adjustment of the vehicle air conditioner.

[0061] To achieve the above technical solutions, embodiments of this application may include: installing a high-resolution in-cabin vision camera in the cockpit to accurately identify the user's micro-behaviors (such as rubbing hands or wiping sweat) and facial orientation; installing an infrared sensor to measure facial skin temperature, or using integrated sensors in the steering wheel / seat to monitor heart rate variability, etc., to obtain the user's physiological data; using a vehicle bus (such as a CAN bus) for efficient communication to obtain contextual information such as vehicle speed, ambient light, and outside temperature, and sending control commands to actuators such as the air conditioner and seats; and using a predictive model to align and fuse visual, physiological, and environmental data over time to form a unified and reliable judgment of the user's state.

[0062] The predictive model can use sample data from a database that correlates user behavior, physiological and environmental parameters with subjective comfort feedback as the basis for model training. It can also combine user manual intervention records to dynamically adjust and optimize the predictive model, achieving the effect of becoming more and more accurate with use. Furthermore, it can transform the predictive decision results (such as feeling too cold) into specific and gentle actuator instructions (such as slightly raising the air conditioning temperature or turning on the seat heating) to avoid abruptness.

[0063] Building upon this, the embodiments of this application can also obtain sample data through testing. First, a compliance assessment is conducted to ensure that each step complies with relevant constraints and standards, confirming that the automatic adjustment function will not adversely affect driving safety. Then, a standardized data collection and anonymization process is created, allowing for model training and optimization while protecting user privacy. A clear user notification and authorization mechanism is established to ensure the legality and compliance of data processing, and a transparent user privacy policy is created.

[0064] The testing process can be as follows: Comprehensive system integration testing was conducted in laboratory and real-vehicle environments to verify the coordination and accuracy of the entire chain, from data perception and algorithm decision-making to actuator control.

[0065] Develop automated test cases covering different climates, users, and driving scenarios to verify the robustness and reliability of the system under various extreme and corner cases.

[0066] Conduct large-scale user road tests and comprehensively evaluate usability and user acceptance through subjective questionnaires and objective behavioral data (such as the frequency of manual intervention).

[0067] Establish an effective user feedback channel, incorporate users' explicit settings and implicit behavioral data into the model optimization loop, and continuously improve system performance and user experience.

[0068] With these technological reserves and detailed planning, corresponding vehicle air conditioning adjustment strategies can be developed to significantly improve cabin comfort and personalization, and enhance driving safety by reducing user distraction.

[0069] like Figure 2 As shown, this embodiment of the application can capture the user's micro-behavioral characteristics and physiological signals in real time by fusing an in-cabin visual camera and biosensors. Combined with a preset predictive model, it proactively determines the user's thermal comfort needs and generates optimized air conditioning control commands. This achieves a leap from passive environmental compensation to proactive personalized comfort, enabling pre-adjustment of airflow, temperature, and seat functions before the user subjectively feels discomfort. This not only significantly improves the comfort and intelligent experience of the cabin environment but also effectively reduces manual intervention during driving, enhancing driving safety. Simultaneously, the continuous learning mechanism continuously optimizes the personalized model, making its adjustments increasingly precise. Specifically, it may include the following steps: Step S201: Acquire user behavior, physiological data, and vehicle environmental data. This application embodiment can acquire user image data, physiological data, and in-vehicle and out-of-vehicle environmental data through cameras, infrared sensors, temperature sensors, etc. For example, after the vehicle starts, the in-cabin vision camera continuously captures the user's upper body image at a frame rate of no less than 15fps; infrared sensors located on the steering wheel or A-pillar periodically (e.g., every 10 seconds) scan the user's face to acquire temperature distribution data; the vehicle's CAN bus synchronously provides environmental parameters such as in-vehicle and out-of-vehicle temperature, sunlight intensity, and vehicle speed.

[0070] Step S202: Input user behavior, physiological, and environmental data into the prediction model. Further, in this embodiment, the acquired data can be input into a pre-built prediction model. The prediction model can first convert the acquired data into corresponding feature vectors, such as behavioral feature vectors, environmental feature vectors, and physiological feature vectors, and then fuse these feature vectors to form a comprehensive multimodal feature vector.

[0071] The fusion algorithm of the predictive model set in the cockpit domain controller can perform timestamp alignment and comprehensive analysis on the aforementioned multi-source data. For example, when the visual algorithm recognizes the hand-rubbing behavior, and infrared data shows a decreasing trend in facial temperature during the same time period, while the in-vehicle thermometer reading remains stable, it can be concluded that the user may feel cold due to personal comfort rather than environmental factors.

[0072] Considering that user states change continuously, the attention mechanism of the prediction model can automatically learn and weight the importance of features at different time steps, making the model pay more attention to the recent state most relevant to the current comfort judgment, rather than a simple average.

[0073] The fusion and decision layers in the predictive model can perform deep fusion of weighted multimodal features through a series of fully connected layers. Within these layers, the model learns the complex nonlinear relationships between different modal features.

[0074] It should be noted that the embodiments of this application may employ a confidence-based triggering mechanism. The fused comprehensive state can be scored, such as determining the interval of each feature, obtaining the corresponding interval score, and then obtaining the total score of all features to obtain the comprehensive state score. The prediction model will only be activated when the comprehensive state score exceeds a preset threshold (e.g., cold confidence > 80%) and the state lasts for a certain period of time (e.g., 5 seconds).

[0075] Step S203: Output the prediction and adjustment parameters for the user's comfort state.

[0076] The output layer of the prediction model, i.e., the final fully connected layer, can use the Softmax activation function to output a three-class probability distribution: [P(cold), P(comfortable), P(hot)]. Simultaneously, a parallel regression output head outputs a specific scalar value for the intensity of the adjustment proposal, used to quantify the adjustment magnitude.

[0077] It's important to note that a combination of loss functions can be used during model training, including: classification loss, since data on comfortable states may be far more abundant than those on the cold / hot side, Focal Loss can be used to address class imbalance; and regression loss, using Smooth L1 Loss to constrain the predicted adjustment intensity, ensuring it matches the user's subsequent manual adjustments. The total loss is a weighted sum of the classification and regression losses.

[0078] During training, the predictive model is first pre-trained in the cloud using a large-scale, anonymized cockpit dataset. After deployment in the vehicle, transfer learning techniques are used to fine-tune the last one or two layers of the model using anonymized data from specific users of the vehicle, achieving personalized adaptation.

[0079] The predictive model can output the optimal adjustment command based on the state. For example, in a cold environment, instead of directly setting the air conditioner to the highest temperature, it will gently raise the set temperature by 1.5°C and automatically turn on the seat heating (if equipped) to a low level. This is intended to alleviate discomfort imperceptibly and avoid attracting the user's attention or causing new discomfort due to overly aggressive adjustments.

[0080] Step S204: Execute the adjustment parameters. To increase user trust, this embodiment of the application can provide a concise prompt on the vehicle's infotainment screen when adjusting the parameters, such as "We have detected that you may feel slightly cold and have increased the temperature for you," and allow the user to cancel the operation with one click. Users can also turn off such prompts in the settings to achieve a completely seamless experience.

[0081] In this embodiment, all adjustment actions are set with safety boundaries. For example, large-scale wind direction or mode switching will not be performed when the vehicle is being driven aggressively (such as cornering at high speed). The adjustment speed has also been carefully calibrated to avoid sudden strong winds or strong cold or heat stimuli affecting driving safety.

[0082] Furthermore, embodiments of this application can also disable the automatic adjustment function of the vehicle's air conditioning. For example, when a key sensor (such as a vision camera) fails or is obstructed, a prompt is issued to the user, and the system automatically downgrades to a traditional automatic air conditioning mode based on environmental sensors and the user's preset preferences, ensuring the availability of basic air conditioning functions. Alternatively, if the user does not require the automatic adjustment function, they can completely disable it at any time in the vehicle settings, reverting to a traditional automatic air conditioning mode, or only disable specific functions (such as disabling only vision-based behavior recognition).

[0083] Building upon this foundation, embodiments of this application can also establish user profiles. For example, upon initial vehicle delivery or the addition of a new user, the user can be bound to a specific identity via facial recognition or account login. The user is guided through a brief calibration process: under a set comfortable ambient temperature, the baseline facial infrared temperature characteristics and typical sitting posture image of the user in this state are recorded. Simultaneously, the user can set their basic temperature preference (e.g., a preference for coolness or warmth) as initial parameters for the personalized model. The vehicle can also be pre-installed with a general user behavior-comfort prediction model trained on massive amounts of data. This model includes an initial correspondence between common micro-behaviors (e.g., rubbing hands, wiping sweat) and air conditioning adjustment needs. The general user behavior-comfort prediction model can periodically (e.g., monthly) upload anonymized training data to the cloud when the vehicle is off and connected to Wi-Fi. The cloud uses a large amount of vehicle data to retrain and optimize the prediction model, and then delivers the updated model to the vehicle via OTA (Over-The-Air) updates, enabling continuous evolution of the entire model.

[0084] This application embodiment can continuously monitor user behavior after predictive adjustments. If the user does not perform any manual overwrite operation within 2 minutes after the adjustment, the adjustment is recorded as a success case to reinforce the current behavior-adjustment mapping relationship. Conversely, if the user manually adjusts the temperature back, it is considered a negative feedback, fine-tuning the model so that different adjustment strategies are adopted in similar situations in the future, and the user's behavior is recorded in the user profile.

[0085] It is important to note that in this embodiment, all data involving user biometrics (such as facial images and infrared temperature) is processed and transformed within the cockpit domain controller. Only the extracted abstract feature values ​​(such as behavioral codes and temperature trends) are used for model calculations. The original image data is discarded immediately after processing and is not stored or uploaded. Furthermore, all data uploaded for model optimization undergoes strict anonymization and cannot be associated with any specific vehicle or individual. All vehicle-to-cloud communications employ high-strength encryption protocols to ensure data transmission security.

[0086] In summary, the embodiments of this application can construct a multimodal user state perception system by integrating visual behavior recognition and non-contact physiological parameter detection, and establish a predictive model based on this system. This enables proactive and timely triggering of air conditioning adjustments before the user experiences noticeable discomfort. The system can learn and memorize the preferred air conditioning settings of different users under specific behaviors, physiological signals, and environmental parameters. Through a continuous learning mechanism, it continuously optimizes decisions, thereby achieving truly adaptive environmental adjustments that perfectly match individual user differences and real-time states, enhancing the exclusivity and intelligence of the cabin experience. It minimizes manual intervention by the user during driving. Through highly accurate predictive automatic adjustment, the system can meet most of the user's dynamic comfort needs, eliminating the need for manual operation and effectively reducing distraction caused by adjusting the air conditioning, thus improving driving safety.

[0087] The vehicle air conditioning adjustment method proposed in this application can input the user's image data, physiological data, and the vehicle's interior and exterior environmental data under the current driving conditions into a pre-built comfort prediction model. By fusing visual behavior recognition, non-contact physiological parameters, and in-vehicle and exterior environmental data, the model predicts the user's perceived temperature characteristics. This proactively triggers the adjustment of the vehicle air conditioning before the user feels uncomfortable, reducing manual intervention by the user during driving and improving the driving experience. This solves the technical problems in related technologies where the vehicle air conditioning adjustment method lacks predictability, easily leading to a mismatch between the adjusted interior temperature and the user's perceived temperature. Furthermore, when learning the user's manual adjustment parameters, it is easy to include non-representative adjustment parameters in the learning scope, causing the automatic adjustment parameters of the vehicle air conditioning to gradually deviate from the user's comfort range, affecting the user's actual user experience.

[0088] Next, referring to the accompanying drawings, we describe the vehicle air conditioning regulating device according to an embodiment of this application.

[0089] Figure 3 This is a block diagram of the regulating device of the vehicle air conditioner according to an embodiment of this application.

[0090] like Figure 3 As shown, the vehicle air conditioning adjustment device 10 includes: a first acquisition module 100, a calculation module 200 and an adjustment module 300.

[0091] Specifically, the first acquisition module 100 is used to acquire the user's image data and physiological data under the current driving conditions.

[0092] The calculation module 200 is used to obtain the user's perceived temperature characteristics by combining image data, physiological data, ambient temperature inside the vehicle under the current driving conditions, and ambient temperature outside the vehicle.

[0093] The adjustment module 300 is used to determine the user's perceived temperature state and the corresponding temperature adjustment parameters based on the perceived temperature characteristics, and to control the vehicle's air conditioning to perform corresponding temperature adjustment actions based on the temperature adjustment parameters until the perceived temperature state meets the preset comfort conditions.

[0094] Optionally, in one embodiment of this application, the calculation module 200 includes: an identification unit, a first determination unit, a second determination unit, a first acquisition unit, a third determination unit, and a second acquisition unit.

[0095] The recognition unit is used to identify key human body points in each frame of the image data.

[0096] The first determining unit is used to determine the user's behavioral feature vector based on human body key points and temporal information of each frame image.

[0097] The second determining unit is used to acquire the trend and fluctuation data of the user's physiological parameters under the current driving conditions based on physiological information, and to determine the user's physiological feature vector based on the trend and fluctuation data of the physiological parameters.

[0098] The first acquisition unit is used to acquire the ambient temperature and solar radiation intensity outside the vehicle based on the external environment data.

[0099] The third determining unit is used to determine the vehicle's environmental feature vector by combining the ambient temperature inside the vehicle, the ambient temperature outside the vehicle, the solar radiation intensity, and the vehicle speed.

[0100] The second acquisition unit is used to combine behavioral feature vectors, physiological feature vectors, and environmental feature vectors to obtain the perceived temperature features.

[0101] Optionally, in one embodiment of this application, the second acquisition unit includes: a splicing subunit, a processing subunit, a splitting subunit, and a first calculation subunit.

[0102] The splicing sub-unit is used to splice behavioral feature vectors, physiological feature vectors, and environmental feature vectors to obtain a comprehensive feature vector.

[0103] The processing subunit is used to reduce the dimensionality of the comprehensive feature vector to obtain the feature vector to be processed.

[0104] The sub-unit is used to split the feature vector to be processed using a preset time step, and to assign corresponding weights to the feature vector to be processed at each time step using a temporal attention mechanism.

[0105] The first computational subunit is used to combine the weights and the feature vector to be processed at each time step to calculate the perceived temperature feature vector, so as to obtain the perceived temperature feature based on the perceived temperature feature vector.

[0106] Optionally, in one embodiment of this application, the adjustment module 300 includes: a conversion unit, a mapping unit, and a third acquisition unit.

[0107] The conversion unit is used to convert the perceived temperature characteristics into a perceived probability distribution, so as to obtain the perceived temperature state based on the perceived probability distribution.

[0108] The mapping unit is used to map the perceived temperature characteristics into a scalar of adjustment intensity that meets preset continuity conditions using preset mapping rules.

[0109] The third acquisition unit is used to combine the adjustment intensity scalar and the perceived temperature state to obtain the temperature adjustment parameters.

[0110] Optionally, in one embodiment of this application, the vehicle air conditioning regulating device 10 further includes a second acquisition module and an optimization module.

[0111] The second acquisition module is used to acquire the actual temperature adjustment parameters of the vehicle air conditioner.

[0112] The optimization module is used to optimize preset mapping rules based on actual temperature adjustment parameters.

[0113] Optionally, in one embodiment of this application, the vehicle air conditioning regulating device 10 further includes: a first control module and a second control module.

[0114] The first control module is used to control the vehicle's air conditioning to perform corresponding temperature adjustment actions based on the temperature adjustment parameters when the temperature adjustment parameters are within the preset safe adjustment limit range.

[0115] The second control module is used to determine the target temperature adjustment parameter of the vehicle air conditioner based on the preset safety adjustment limit range in response to the temperature adjustment parameter not being within the preset safety adjustment limit range, so as to control the vehicle air conditioner to perform the corresponding temperature adjustment action using the target temperature adjustment parameter.

[0116] It should be noted that the explanation of the above-mentioned embodiment of the vehicle air conditioning adjustment method also applies to the vehicle air conditioning adjustment device of this embodiment, and will not be repeated here.

[0117] The vehicle air conditioning adjustment device proposed in this application can input the user's image data, physiological data, and the ambient temperature inside and outside the vehicle under the current driving conditions into a pre-built comfort prediction model. By fusing visual behavior recognition, non-contact physiological parameters, and environmental data inside and outside the vehicle, it predicts the user's perceived temperature characteristics. This proactively triggers the adjustment of the vehicle air conditioning before the user feels uncomfortable, reducing manual intervention by the user during driving and improving the driving experience. This solves the technical problems in related technologies where the adjustment method of vehicle air conditioning lacks predictability, easily leading to a mismatch between the adjusted interior temperature and the user's perceived temperature. Furthermore, when learning the user's manual adjustment parameters, it is easy to include non-representative adjustment parameters in the learning scope, causing the automatic adjustment parameters of the vehicle air conditioning to gradually deviate from the user's comfort range, affecting the user's actual user experience.

[0118] Figure 4 A schematic diagram of the structure of a vehicle provided in an embodiment of this application. The vehicle may include: The memory 401, the processor 402, and the computer program stored on the memory 401 and capable of running on the processor 402.

[0119] When the processor 402 executes the program, it implements the vehicle air conditioning adjustment method provided in the above embodiments.

[0120] Furthermore, the vehicle also includes: Communication interface 403 is used for communication between memory 401 and processor 402.

[0121] The memory 401 is used to store computer programs that can run on the processor 402.

[0122] Memory 401 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0123] If the memory 401, processor 402, and communication interface 403 are implemented independently, then the communication interface 403, memory 401, and processor 402 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 4 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0124] Optionally, in a specific implementation, if the memory 401, processor 402, and communication interface 403 are integrated on a single chip, then the memory 401, processor 402, and communication interface 403 can communicate with each other through an internal interface.

[0125] Processor 402 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0126] This embodiment also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the above-described method for adjusting the vehicle air conditioner.

[0127] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the vehicle air conditioning adjustment method provided in this embodiment of the invention.

[0128] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0129] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0130] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0131] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0132] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0133] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0134] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0135] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

Claims

1. A method for adjusting a vehicle air conditioner, characterized in that, Includes the following steps: Acquire image and physiological data of the user under the current driving conditions; The user's perceived temperature characteristics are obtained by combining the image data, the physiological data, the ambient temperature inside the vehicle under the current driving conditions, and the ambient temperature outside the vehicle. Based on the perceived temperature characteristics, the user's perceived temperature state and the corresponding temperature adjustment parameters are determined. Based on the temperature adjustment parameters, the vehicle's air conditioning is controlled to perform corresponding temperature adjustment actions until the perceived temperature state meets the preset comfort conditions.

2. The method according to claim 1, characterized in that, The step of obtaining the user's perceived temperature characteristics from the image data, the physiological data, the ambient temperature inside the vehicle under the current driving conditions, and the ambient temperature outside the vehicle includes: Identify key human body points in each frame of the image data; Based on the key points of the human body and the temporal information of each frame of the image, the user's behavioral feature vector is determined; Based on the physiological information, the physiological parameter change trend and fluctuation data of the user under the current driving condition are obtained, and the physiological feature vector of the user is determined based on the physiological parameter change trend and fluctuation data. Based on the external environment data, the ambient temperature and sunlight intensity outside the vehicle are obtained. The environmental feature vector of the vehicle is determined by combining the ambient temperature inside the vehicle, the ambient temperature outside the vehicle, the solar radiation intensity, and the vehicle speed. The perceived temperature feature is obtained by combining the behavioral feature vector, the physiological feature vector, and the environmental feature vector.

3. The method according to claim 2, characterized in that, The process of combining the behavioral feature vector, the physiological feature vector, and the environmental feature vector to obtain the perceived temperature feature includes: By concatenating the behavioral feature vector, the physiological feature vector, and the environmental feature vector, a comprehensive feature vector is obtained. The dimensionality reduction of the comprehensive feature vector is performed to obtain the feature vector to be processed; The feature vector to be processed is split using a preset time step, and a temporal attention mechanism is used to assign corresponding weights to the feature vector to be processed at each time step. By combining the weights and the feature vector to be processed at each time step, a perceptual temperature feature vector is calculated to obtain the perceptual temperature feature based on the perceptual temperature feature vector.

4. The method according to claim 1, characterized in that, The process of determining the user's perceived temperature state and the corresponding temperature adjustment parameters based on the perceived temperature characteristics includes: The perceived temperature characteristics are converted into a perceived probability distribution, and the perceived temperature state is obtained based on the perceived probability distribution; Using a preset mapping rule, the perceived temperature feature is mapped to an adjustment intensity scalar that satisfies a preset continuity condition; The temperature regulation parameters are obtained by combining the adjustment intensity scalar and the perceived temperature state.

5. The method according to claim 4, characterized in that, Also includes: Obtain the actual temperature adjustment parameters of the vehicle air conditioner; The preset mapping rule is optimized based on the actual temperature adjustment parameters.

6. The method according to claim 1, characterized in that, Before controlling the vehicle's onboard air conditioning to perform the corresponding temperature adjustment action based on the temperature adjustment parameters, the method further includes: In response to the temperature adjustment parameter being within a preset safe adjustment limit range, the vehicle's air conditioning is controlled to perform corresponding temperature adjustment actions based on the temperature adjustment parameter. In response to the temperature adjustment parameter not being within the preset safety adjustment limit range, a target temperature adjustment parameter for the vehicle air conditioner is determined based on the preset safety adjustment limit range, so as to control the vehicle air conditioner to perform corresponding temperature adjustment actions using the target temperature adjustment parameter.

7. A regulating device for a vehicle air conditioner, characterized in that, include: The acquisition module is used to acquire the user's image data and physiological data under the current driving conditions; The calculation module is used to obtain the user's perceived temperature characteristics from the image data, the physiological data, the ambient temperature inside the vehicle under the current driving conditions, and the ambient temperature outside the vehicle. The adjustment module is used to determine the user's perceived temperature state and the temperature adjustment parameters corresponding to the perceived temperature state based on the perceived temperature characteristics, and to control the vehicle's air conditioning to perform corresponding temperature adjustment actions based on the temperature adjustment parameters until the perceived temperature state meets the preset comfort conditions.

8. A vehicle, characterized in that, include: A memory, a processor, and a computer program stored in the memory and capable of running on the processor, the processor executing the program to implement the method for adjusting an in-vehicle air conditioner as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the vehicle air conditioning adjustment method as described in any one of claims 1-6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed, it is used to implement the vehicle air conditioning adjustment method as described in any one of claims 1-6.

Images