An intelligent display method and system of an automobile air conditioner panel with environment self-learning function

Abstract

This invention discloses a method and system for intelligent display of an automotive air conditioning panel with environmental self-learning function. The method includes: collecting environmental parameters and user operation data in the cockpit through an onboard multimodal sensor, and generating a correlated perception data stream based on timestamps; dynamically constructing a personalized user profile including preferred temperature and frequently used function combinations based on the correlated perception data stream and user identity recognition results; predicting the user's air conditioning operation intention in a future preset time period through a time-series prediction model based on the personalized user profile and real-time environmental parameters, and generating an adaptive display strategy based on the prediction results; dynamically mapping the virtual button layout and function icon display hierarchy of the air conditioning panel according to the adaptive display strategy, and generating a personalized interactive interface matching the current scene and predicted intention. Using this invention, the human-computer interaction efficiency and personalization of the air conditioning panel can be improved, and the intelligent adaptation capability of the cabin environment can be enhanced.

Description

Technical Field

[0001] This invention belongs to the field of automotive technology, specifically a method and system for intelligent display of automotive air conditioning panels with environmental self-learning function. Background Technology

[0002] With the rapid development of automotive intelligence and cockpit human-machine interaction technology, the in-vehicle air conditioning panel, as the core interactive entry point for cockpit environmental control, directly impacts driving safety and user experience through its display and operation. Traditional automotive air conditioning panels mostly employ fixed physical buttons or simple touch layouts, lacking adaptability to individual user differences and dynamic environments in terms of function display priorities and interface arrangement. In recent years, some models have begun to introduce automatic air conditioning adjustment strategies based on single environmental parameters (such as outside temperature or inside humidity), but the panel display remains a static structure, unable to proactively adjust according to the user's historical operating habits, identity differences, and future operating intentions. Existing methods generally suffer from the following shortcomings: First, they lack deep modeling of user preferences and operation sequences, making it difficult to achieve personalized displays; second, interface adjustments lag behind environmental changes and user expectations, resulting in low interaction efficiency and long operation paths; and third, they fail to integrate timing prediction mechanisms, failing to proactively optimize the function entry layout before user operation. Summary of the Invention

[0003] The purpose of this invention is to provide a smart display method and system for automotive air conditioning panels with environmental self-learning function, so as to overcome the shortcomings of the prior art, improve the human-computer interaction efficiency and personalization of the air conditioning panel, and enhance the intelligent adaptation capability of the cabin environment.

[0004] One embodiment of this application provides a method for intelligent display of an automotive air conditioning panel with environmental self-learning function, the method comprising: The vehicle uses onboard multimodal sensors to collect environmental parameters and user operation data in the cockpit, and generates associated perception data streams based on timestamps. Based on the associated perception data stream and combined with the user identity recognition results, a personalized user profile containing preferred temperature and commonly used function combinations is dynamically constructed. Based on the user's personalized profile and real-time environmental parameters, the user's air conditioning operation intention in the future preset time period is predicted by the time-series prediction model, and an adaptive display strategy is generated by combining the prediction results. The strategy includes a function priority ranking and interface layout adjustment scheme. Based on the adaptive display strategy, the virtual button layout and function icon display hierarchy of the air conditioner panel are dynamically mapped to generate a personalized interactive interface that matches the current scene and predicted intent.

[0005] Optionally, the step of collecting environmental parameters and user operation data within the cockpit using onboard multimodal sensors and generating a correlated perception data stream based on timestamps includes: Real-time data on temperature, humidity, light intensity, and PM2.5 concentration in the cockpit are collected using temperature and humidity sensors, light sensors, and air quality sensors to generate a raw environmental parameter dataset. The system collects user operation data on the air conditioning system through the touch screen interaction module and physical buttons, including temperature set value, fan speed level, air outlet mode and circulation mode selection, and generates the original user operation dataset. A unified clock source is used to add timestamps to each data point in the original environmental parameter dataset and the original user operation dataset. The two types of data are then aligned along the time axis using a data fusion algorithm to generate a time-synchronized multimodal dataset. A sliding window segmentation process is applied to the time-synchronized multimodal dataset to extract the correlation between the environmental parameter sequence and the operation event sequence within a fixed time window, ultimately generating a correlated perception data stream.

[0006] Optionally, the step of dynamically constructing a personalized user profile containing preferred temperature and frequently used function combinations based on the associated perception data stream and user identity recognition results includes: The driver's facial image is captured by an in-vehicle camera, and facial recognition algorithm is used to extract facial features and compare them with a pre-registered user database to generate user identity recognition results. Based on the user identification results, historical operation records belonging to the current user are filtered from the associated perception data stream, and environmental parameters and operation actions corresponding to the operation time are extracted to generate a user historical behavior dataset. Statistical analysis was performed on the user's historical behavior dataset, and clustering algorithms were used to identify the user's preferred temperature range and commonly used function combination patterns under different environmental conditions, generating statistical results of preference features; By integrating user identification results with preference feature statistics, a mapping relationship is established between user identifiers and preference temperature ranges and commonly used function combinations, ultimately generating a personalized user profile that includes preference temperature and commonly used function combinations.

[0007] Optionally, based on the user's personalized profile and real-time environmental parameters, a time-series prediction model is used to predict the user's air conditioning operation intentions within a preset future time period. An adaptive display strategy is then generated based on the prediction results. This strategy includes a function priority ranking and interface layout adjustment scheme, including: The system extracts preferred temperature ranges and frequently used function combinations from the user's personalized profile, while simultaneously collecting real-time environmental parameter data to generate the input feature vector for the prediction model. The prediction model is input into the feature vector and then into the pre-trained temporal prediction model. This model uses a long short-term memory network architecture to learn the temporal correlation between the user's historical operation patterns and environmental changes, and generates the probability distribution of operation intentions within a preset future time period. Based on the probability distribution of operational intentions, a sorting algorithm is used to prioritize the various functions of the air conditioner from high to low according to the predicted probability of use, and a function priority sorting result is generated. Based on the functional priority ranking, combined with the screen size and interaction specifications of the air conditioner panel, an interface layout adjustment scheme is designed to place high-priority functions in easily accessible and prominent positions, ultimately generating an adaptive display strategy that includes both functional priority ranking and interface layout adjustment scheme.

[0008] Optionally, the step of dynamically mapping the virtual button layout and function icon display hierarchy of the air conditioner panel according to the adaptive display strategy, and generating a personalized interactive interface that matches the current scene and predicted intent, includes: The function priority ranking result in the adaptive display strategy is analyzed, and corresponding display level coefficients are assigned to each function. High-priority functions are assigned the highest level coefficients, and a function display level mapping table is generated. Based on the function display hierarchy mapping table and interface layout adjustment scheme, generate virtual button layout instructions, determine the coordinate position and display size of each function button on the air conditioner panel screen, and generate a virtual button layout scheme. Based on the function display hierarchy mapping table, the display style of function icons is dynamically adjusted, including high-priority functions using highlighted colors and dynamic effects, and low-priority functions using simplified icons or being included in the secondary menu, generating a function icon display configuration. By integrating the virtual button layout scheme and function icon display configuration, a personalized interactive interface is generated through the vehicle display screen driver module, and finally the air conditioning panel display interface that matches the current scene and predicted intent is output.

[0009] Another embodiment of this application provides an intelligent display system for an automotive air conditioning panel with an environmental self-learning function, the system comprising: The data acquisition module is used to collect environmental parameters and user operation data in the cockpit through on-board multimodal sensors, and generate associated perception data streams based on timestamps; The module is used to dynamically construct a personalized user profile that includes preferred temperature and commonly used function combinations based on the associated perception data stream and the user identity recognition results. The prediction module is used to predict the user's air conditioning operation intention in a future preset period of time based on the user's personalized profile and real-time environmental parameters through a time-series prediction model, and generate an adaptive display strategy based on the prediction results. The strategy includes a function priority ranking and interface layout adjustment scheme. The generation module is used to dynamically map the virtual button layout and function icon display hierarchy of the air conditioner panel according to the adaptive display strategy, and generate a personalized interactive interface that matches the current scene and predicted intent.

[0010] Another embodiment of this application provides a storage medium storing a computer program, wherein the computer program is configured to execute the method described in any of the preceding claims when running.

[0011] Another embodiment of this application provides an electronic device including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the method described in any of the preceding claims.

[0012] Compared with existing technologies, the present invention provides an intelligent display method for automotive air conditioning panels with environmental self-learning function, which can improve the human-computer interaction efficiency and personalization of air conditioning panels and enhance the intelligent adaptation capability of the cabin environment. Attached Figure Description

[0013] Figure 1 Hardware structure block diagram of a computer terminal for an intelligent display method for an automotive air conditioning panel with environmental self-learning function provided in an embodiment of the present invention; Figure 2 A flowchart illustrating an intelligent display method for an automotive air conditioning panel with environmental self-learning function, provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of an intelligent display system for an automotive air conditioning panel with environmental self-learning function, provided as an embodiment of the present invention. Detailed Implementation

[0014] The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0015] This invention first provides a method for intelligent display of an automotive air conditioning panel with environmental self-learning function. This method can be applied to electronic devices, such as computer terminals, specifically ordinary computers.

[0016] The following detailed explanation uses a computer terminal as an example. Figure 1 This is a hardware structure block diagram of a computer terminal for a smart display method for an automotive air conditioning panel with environmental self-learning function, provided as an embodiment of the present invention. (See diagram for reference.) Figure 1 As shown, the computer device includes a processor, memory, and network interface connected via a system bus, wherein the memory may include non-volatile storage media and internal memory.

[0017] See Figure 2The present invention provides a method for intelligent display of an automotive air conditioning panel with environmental self-learning function, which may include the following steps: S201 collects environmental parameters and user operation data in the cockpit through onboard multimodal sensors and generates associated perception data streams based on timestamps; Specifically, temperature, humidity, light intensity, and PM2.5 concentration data in the cockpit can be collected in real time using temperature and humidity sensors, light sensors, and air quality sensors to generate a raw environmental parameter dataset; The core of this step is to rely on various onboard environmental perception sensors to continuously acquire environmental status information inside the cockpit, converting physical environmental changes into quantifiable digital signals to form a complete dataset of raw environmental parameters. This provides foundational environmental dimension data for subsequent data fusion and behavioral analysis. The specific implementation method is as follows: The temperature and humidity sensor is the core sensing device for collecting temperature and humidity data in the cockpit. Utilizing a non-contact sensing principle, it can capture real-time environmental temperature and humidity values ​​in different areas of the cockpit. The sampling frequency is set to 1Hz, meaning data is collected once per second, ensuring timely recording of environmental changes. The temperature parameter sampling range is set to 0°C to 60°C, with an accuracy controlled to 0.1°C. This range covers the temperature variation range within the cockpit under different climatic conditions, and the accuracy meets the fine-grained needs of subsequent user preference analysis. The collected temperature value is denoted as T_n, where n represents the nth temperature data collection, and the subscript n is used to distinguish the temperature value at different sampling times. The humidity parameter sampling range is 0%RH to 100%RH, with an accuracy of 1%RH. RH represents relative humidity, a key indicator characterizing the water vapor content in the air. The collected humidity value is denoted as H_n, and the subscript n corresponds to the temperature sampling time, ensuring the correlation of temperature and humidity data at the same moment.

[0018] The light sensor is used to sense changes in light intensity inside the cockpit. It mainly collects light values ​​under scenarios such as direct sunlight and when the interior lights are on. The sampling frequency is also 1Hz, and the sampling range is set from 0 lux to 100,000 lux. Lux is the standard unit of light intensity. This range covers all interior lighting scenarios from no light at night to strong direct sunlight. The accuracy is controlled within 10 lux. The collected light intensity value is recorded as L_n, and the subscript n still corresponds to the sampling time, which is used to mark the light state at different times.

[0019] The air quality sensor focuses on detecting air cleanliness in the cockpit, primarily collecting PM2.5 concentration data. PM2.5 refers to particulate matter with an aerodynamic equivalent diameter of 2.5 micrometers or less, and is an important indicator for measuring in-vehicle air quality. The sensor collects data at a frequency of 0.5 Hz, meaning it collects data every two seconds, with a range of 0 μg / m³ to 500 μg / m³. μg / m³ is the unit of measurement for particulate matter concentration, with an accuracy of 1 μg / m³. The collected PM2.5 concentration value is denoted as P_n, where the subscript n is used to distinguish air quality data from different collection periods.

[0020] During the data collection process, various sensors continuously convert the sensed analog signals into digital electrical signals, which are then transmitted to the data processing unit via the vehicle bus. The data processing unit performs preliminary filtering on the sensor signals, eliminating abnormal values ​​caused by vehicle vibrations and electromagnetic interference to ensure the validity of the collected data. The four types of data—temperature, humidity, light intensity, and PM2.5 concentration—after preliminary processing, are stored sequentially according to the collection order. Each set of data includes the corresponding sensor type and the collected value. After integrating all the collected valid environmental data, a raw environmental parameter dataset is formed. This dataset completely records the continuous state of environmental parameters in the cockpit over time, without missing data or outlier interference.

[0021] The system collects user operation data on the air conditioning system through the touch screen interaction module and physical buttons, including temperature set value, fan speed level, air outlet mode and circulation mode selection, and generates the original user operation dataset. The core of this step is to capture the user's active operating behavior through the interactive hardware of the vehicle's air conditioning system, converting the user's various adjustment commands for the air conditioning into standardized operating data, forming a raw user operation dataset. This dataset complements the environmental parameter data, providing a basis for subsequent correlation analysis based on user behavior dimensions. The specific implementation method is as follows: The touchscreen interaction module is the core interface of the vehicle's air conditioning system. Utilizing capacitive touch sensing technology, it accurately recognizes user input commands such as clicks, swipes, and long presses. Users can directly adjust various air conditioning parameters through this module. The module's response latency is controlled within 30ms, ensuring rapid acquisition of input commands. Physical buttons, serving as auxiliary interaction hardware, are conveniently located on the air conditioning panel. These include physical buttons for fan speed adjustment and mode switching. Employing a mechanical press-triggered principle, their sensitivity has been calibrated to prevent accidental touches and invalid input. These two types of interaction hardware work together to comprehensively cover all user air conditioning operation scenarios.

[0022] The temperature setpoint is the core operational data when users adjust the air conditioning. It represents the target temperature in the cockpit that the user expects. The adjustment range of this value is 16 degrees Celsius to 32 degrees Celsius, with an adjustment step of 0.5 degrees Celsius. Users can complete the setting by sliding on the touch screen or adding or subtracting using physical buttons. The collected temperature setpoint is recorded as Ts_m, where m represents the mth user operation. The subscript m is used to distinguish the set temperature at different operation times. This value directly reflects the user's temperature preference.

[0023] The fan speed setting is the user's choice to adjust the airflow of the air conditioner. There are 6 settings: off, 1 to 5. The higher the value of the setting, the greater the airflow. Users can switch settings through the setting selection area on the touch screen or the physical fan speed button. The collected fan speed data is recorded as F_m, where the subscript m corresponds to the time of operation and represents the airflow requirement under different operations.

[0024] The air outlet mode is the user's setting and selection of the air outlet direction of the air conditioner. It includes four basic modes: face blowing mode, foot blowing mode, defogging mode, and face blowing and foot blowing mixed mode. Users can switch between different air outlet directions through the interactive module. The collected air outlet mode data is recorded as M_m. The four modes are respectively represented by numbers 1 to 4. The subscript m marks the operation sequence, clearly recording the user's air outlet direction selection behavior.

[0025] The circulation mode is the user's choice of how the air inside the vehicle is circulated. There are two modes: internal circulation mode and external circulation mode. Internal circulation mode means that the air inside the vehicle circulates independently, while external circulation mode means that the air inside and outside the vehicle circulates and exchanges. The collected circulation mode data is recorded as C_m. The two modes are represented by the numbers 1 and 2 respectively. The subscript m is consistent with the operation time, reflecting the user's changing needs for air circulation.

[0026] The data acquisition unit monitors the trigger signals from the touchscreen interaction module and physical buttons in real time. When a valid operation command is detected, it immediately extracts the corresponding operation parameters, discards consecutive and repeated invalid operation data, and only records the operation result finally confirmed by the user. The temperature setpoint, fan speed, airflow mode, and circulation mode data corresponding to each valid operation are integrated and stored. Each operation data entry is labeled with the operation type and specific parameters. After all user air conditioning operation data is summarized, a raw user operation dataset is generated. This dataset comprehensively records the user's adjustment behavior of the air conditioning system in different scenarios, providing detailed behavioral evidence for subsequent analysis of user operating habits.

[0027] A unified clock source is used to add timestamps to each data point in the original environmental parameter dataset and the original user operation dataset. The two types of data are then aligned along the time axis using a data fusion algorithm to generate a time-synchronized multimodal dataset. The core of this step is to resolve the time asynchrony between environmental parameters and user operation data. By standardizing time stamps through a unified clock source and relying on data fusion algorithms, the time axes of the two types of data are aligned, eliminating correlation analysis errors caused by time deviations and generating a multimodal dataset with a unified time dimension. The specific implementation method is as follows: A unified clock source is the core timing benchmark of the vehicle system. It uses the crystal oscillator clock of the vehicle's main control unit as the timing core, achieving millisecond-level accuracy and controlling timing errors within 1 millisecond. This provides a unified time stamp for all vehicle data, avoiding time discrepancies caused by differences in the clocks of different sensors and interaction modules. The timestamp uses a decimal number format (hour-minute-second-millisecond), accurately marking the generation time of each data point. A timestamp, St_n, is added to each environmental data point in the original environmental parameter dataset, where n corresponds to the data collection sequence number. Similarly, a timestamp, St_m, is added to each operation data point in the original user operation dataset, where m corresponds to the user operation sequence number, ensuring that each data point has a unique time identifier.

[0028] Data fusion algorithms are the core technology for aligning the timelines of two types of data. One such algorithm is a time nearest neighbor matching fusion algorithm. The core logic of this algorithm is to traverse each operation data entry in the original user operation dataset, using the operation data's timestamp St_m as a benchmark. It then searches the original environment parameter dataset for the environment data whose timestamp differs from St_m the smallest. The matching environment data is then associated and bound with the operation data, completing the timeline alignment. The algorithm sets a time matching threshold of 500 milliseconds, meaning it only matches environment data and operation data with a time difference within 500 milliseconds. Data exceeding this threshold is considered unrelated, ensuring the validity of the matching results.

[0029] During the fusion process, for continuously collected environmental data in the environmental parameter dataset, if no corresponding user operation data is found within the matching threshold, the environmental data is retained and marked as having no operation association. For user operation data, if no matching environmental data is found within the threshold, the most recent valid environmental data is retrieved for supplementary matching, ensuring that each operation data can be associated with the corresponding environmental state. After fusion, the associated environmental parameters and user operation data are integrated into a single data entry. Each entry contains nine types of information: timestamp, temperature, humidity, light intensity, PM2.5 concentration, temperature setpoint, fan speed, air outlet mode, and circulation mode. All integrated entries are arranged in ascending order of timestamp, and duplicate and abnormally associated entries are removed, ultimately forming a time-synchronized multimodal dataset. This dataset achieves the unification of environmental states and user operations in the time dimension, eliminates the time deviation of multi-source data collection, and lays a time benchmark for subsequent sliding window segmentation and association extraction.

[0030] A sliding window segmentation process is applied to the time-synchronized multimodal dataset to extract the correlation between the environmental parameter sequence and the operation event sequence within a fixed time window, ultimately generating a correlated perception data stream.

[0031] The core of this step is to segment the synchronized multimodal data using sliding window segmentation technology, mine the intrinsic correlation between environmental changes and user operations within a fixed time window, and transform the static dataset into a dynamic time-series data stream, forming a correlation-aware data stream that can be used to build user profiles. The specific implementation method is as follows: Sliding window segmentation is a common technique in time-series data processing. This study employs a sliding window pattern with a fixed window size and a fixed sliding step. The fixed time window duration is set to 300 seconds (5 minutes), which fully covers the complete cycle from an environmental change to a user's air conditioning operation, avoiding the loss of correlations due to an excessively short window or data redundancy due to an excessively long window. The sliding step is set to 60 seconds, meaning the window slides backward by 60 seconds each time. There is some data overlap between windows; this overlap ensures the continuity of the environmental parameter sequence and the operation event sequence, preventing the correlation from being interrupted by window segmentation.

[0032] During the segmentation process, based on the timestamps of the time-synchronized multimodal dataset, starting from the dataset's start time, all data within a 300-second timeframe is extracted as the first window. Subsequently, the data is slid out in 60-second increments, sequentially extracting subsequent windows until the entire multimodal dataset has been traversed. Each sliding window contains a continuous sequence of environmental parameters and a sequence of operational events. The environmental parameter sequence is a continuous set of values ​​for temperature, humidity, light intensity, and PM2.5 concentration arranged chronologically within the window. The operational event sequence is a set of parameters related to user air conditioning operations within the window. Time periods without operation are marked as steady-state operation.

[0033] For each segmented time window, a time-series correlation analysis algorithm is used to extract the correlation between environmental parameters and user operations. The algorithm establishes a correspondence between the magnitude of environmental changes and the frequency of operation events by calculating the rate of change of environmental parameters within the window and the frequency of operation events. For example, the correlation between the magnitude of temperature increase and the reduction of temperature setpoint, and the correlation between changes in light intensity and the switching of air outlet mode. The correlation is represented in the form of numerical correlation degree, with the correlation degree ranging from 0 to 1. The closer the value is to 1, the stronger the correlation.

[0034] The environmental parameter sequence, operation event sequence, and extracted correlations within each sliding window are output sequentially according to the window order, forming a continuous streaming data structure. The transmission rate of the streaming data matches the vehicle data acquisition frequency to ensure real-time data transmission. After integrating the streaming data processed by all windows, a final correlated perception data stream is generated. This data stream not only contains multimodal environmental and operational data but also includes temporal correlation features between data, which can be directly used for the dynamic construction of personalized user profiles, providing complete temporal data support for predicting air conditioning operation intentions.

[0035] S202, Based on the associated perception data stream and combined with the user identity recognition results, dynamically construct a personalized user profile that includes preferred temperature and commonly used function combinations; Specifically, the driver's facial image can be captured by the in-vehicle camera, facial recognition algorithm is used to extract facial features and compare them with a pre-registered user database to generate user identity recognition results; The core of this step is to acquire the driver's facial feature information using in-vehicle vision acquisition equipment, extract features and match users through facial recognition algorithms, and finally output accurate user identification results. This provides a basis for subsequent screening of exclusive user behavior data. The specific implementation method is as follows: The in-vehicle camera is mounted above the center console, near the windshield, providing a complete view of the driver's face. This avoids incomplete facial information capture due to perspective deviation. The camera's image resolution is set to 1280×720 pixels, with a frame rate of 15 frames per second and a 200-millisecond interval between each facial image capture. This ensures continuous facial information capture without excessive computational resources on the vehicle's system due to high-frequency capture. The captured raw facial images undergo preprocessing. Gaussian filtering removes noise, followed by grayscale conversion to reduce computational burden from color dimensions. Then, a face detection algorithm locates the facial region, cropping out irrelevant areas such as the background and interior trim. The final result is a standardized 256×256 pixel face image, ensuring accurate feature extraction in subsequent steps.

[0036] The facial recognition algorithm employs a lightweight deep convolutional neural network architecture, specifically optimized for in-vehicle embedded environments. It boasts high computational efficiency and recognition accuracy sufficient for in-vehicle use scenarios. The algorithm extracts a 128-dimensional facial feature vector F_1 from standardized facial images. Each dimension of the feature vector corresponds to the quantified result of key facial features, including facial contour curvature, eye distance ratio, nasal bridge height, and jawline curvature. Each dimension's value ranges from 0 to 1, with precision retained to four decimal places, uniquely representing the driver's facial biometrics. The pre-registered user database is a pre-stored database of user facial features. Each registered user is assigned a unique user identifier ID_n, where n is a positive integer, and the database also stores the corresponding 128-dimensional standard facial feature vector. The database uses a hash index structure for data storage, significantly improving the retrieval speed of feature comparison.

[0037] The feature comparison and matching process employs a cosine similarity algorithm to calculate the cosine similarity value S between the real-time extracted facial feature vector and each pre-registered feature vector in the database. The similarity value ranges from 0 to 1, with values ​​closer to 1 indicating a higher feature match. The similarity threshold set for this matching is 0.85, meaning a match is considered successful when the similarity value is greater than or equal to 0.85. During the comparison process, the algorithm iterates through all user features in the database, with each match taking less than 50 milliseconds. If a unique user matching the threshold is found, that user's identifier is directly output as the user identification result. If no user is matched, an identification tag for an unknown user is generated. If multiple users meet the similarity threshold, the user with the highest similarity value is selected as the final identification result, ensuring the uniqueness and accuracy of the identification.

[0038] Based on the user identification results, historical operation records belonging to the current user are filtered from the associated perception data stream, and environmental parameters and operation actions corresponding to the operation time are extracted to generate a user historical behavior dataset. The core of this step is to rely on the generated user identity recognition results to accurately filter the multimodal fusion-based associated perception data stream, extract the current user's unique environment and operation-related data, and after cleaning and organizing, form a standardized user historical behavior dataset, providing a data foundation for subsequent preference analysis. The specific implementation method is as follows: The associated sensing data stream is a time-series data sequence aligned with timestamps and segmented by a sliding window. Each data unit in the data stream contains a timestamp T_n generated by a unified clock source, an environmental parameter set E_n, and a user operation action set O_n. The timestamp T_n is in the format of hour:minute:second.millisecond, which can accurately mark the time when the data was generated. The environmental parameter set E_n contains four types of parameters: in-vehicle temperature T_e, humidity H_e, light intensity L_e, and PM2.5 concentration P_e. The operation action set O_n contains four types of operation data: air conditioning temperature setting T_o, fan speed G_o, air outlet mode M_o, and circulation mode C_o. All parameters are retained to one decimal place to ensure the precision of the data.

[0039] The filtering process uses the user identifier from the user identification results as the core basis, traversing all historical data in the associated perception data stream. It distinguishes the operational behaviors of different users through data binding tags, retaining only data entries generated during the driving period bound to the current user identifier. Non-target operation data from the front passenger and rear passengers are automatically removed. The filtering time range is set to the past 30 days of historical data, balancing data sufficiency and the timeliness of behavioral preferences. After filtering, the data undergoes cleaning processing, with outlier judgment rules set to remove obviously abnormal data entries such as temperature settings exceeding the conventional air conditioning adjustment range of 16℃ to 32℃ and negative PM2.5 concentrations. Simultaneously, a small amount of missing environmental parameter data is supplemented using the mean-filling method of adjacent timestamp data to ensure the integrity of the data sequence.

[0040] The cleaned and valid data are sorted and organized in chronological order by timestamp. The timestamps, environmental parameter sets, and operation action sets in each data entry are integrated to form a standardized user history behavior dataset. Each data unit in the dataset fully records the user's air conditioning operation behavior in a specific environment. In the example, a data unit has a timestamp of 14:35:22.120, environmental parameters of 24℃ temperature, 45% humidity, light intensity of 300 lux, and PM2.5 concentration of 35 μg / m³, and operation actions of 26℃ temperature setting, fan speed level 2, airflow mode to face, and internal circulation mode. The dataset is stored in a time-series structure, which can intuitively reflect the correlation between user operations and environmental changes, providing high-quality input data for subsequent cluster analysis and preference statistics.

[0041] Statistical analysis was performed on the user's historical behavior dataset, and clustering algorithms were used to identify the user's preferred temperature range and commonly used function combination patterns under different environmental conditions, generating statistical results of preference features; The core of this step is to mine potential patterns in users' historical behavior data through statistical analysis and clustering algorithms, divide users' preference temperature ranges under different environmental scenarios, summarize frequently used function combinations, and finally form quantitative statistical results of preference features. The specific implementation method is as follows: First, basic statistical analysis was conducted on the user historical behavior dataset. This included analyzing the distribution ranges of parameters such as in-vehicle ambient temperature, light intensity, and PM2.5 concentration, as well as the frequency of user air conditioning temperature settings, fan speed levels, airflow modes, and recirculation modes within each range. The percentage of each operation under different environments was calculated, rounded to two decimal places, to gain a preliminary understanding of the basic distribution characteristics of user operations. Subsequently, the K-means clustering algorithm was used for in-depth data mining. This algorithm can group similar environments and operational behaviors into the same category, accurately classifying user behavioral preference patterns. In this clustering, the number of clusters K was set to 4, corresponding to four typical in-vehicle environment scenarios: low temperature and low light intensity, high temperature and high light intensity, normal temperature and normal light intensity, and enclosed space with high PM2.5 concentration.

[0042] During cluster analysis, in-vehicle temperature, light intensity, and PM2.5 concentration were selected as cluster feature dimensions D_1, D_2, and D_3, respectively. The original values ​​of these three dimensions were first normalized, converting parameters of different dimensions into values ​​between 0 and 1 to eliminate the influence of dimensional differences on the clustering results. Clustering was performed based on user temperature settings, and preferred temperature ranges were defined according to the values ​​of the cluster centers. In the example, three core preferred temperature ranges were obtained after clustering: 22℃ to 24℃, 24℃ to 26℃, and 26℃ to 28℃. The probability of using each range in the corresponding environmental scenario was also calculated; for example, the probability of using the 26℃ to 28℃ range in a high-temperature environment was 0.75, and the probability of using the 22℃ to 24℃ range in a low-temperature environment was 0.80.

[0043] For identifying commonly used function combinations, fan speed, airflow mode, and circulation mode are combined and coded. For example, fan speed level 2 combined with face blowing mode and internal circulation is coded as combination code C_1, and fan speed level 3 combined with foot blowing mode and external circulation is coded as combination code C_2. Clustering algorithms are used to count the frequency of each combination code in historical data. Function combinations with a frequency exceeding 20% ​​are identified as commonly used function combinations. In the example, combination code C_1 and fan speed level 1 combined with automatic airflow mode and automatic circulation mode (C_3) are selected as commonly used function combinations. Finally, data such as preferred temperature range, probability of use within the range, commonly used function combinations, and probability of combination triggering are integrated to form a complete statistical result of preference characteristics. All quantitative values ​​in the result are rounded to two decimal places, clearly presenting the user's air conditioning usage preferences in different environments.

[0044] By integrating user identification results with preference feature statistics, a mapping relationship is established between user identifiers and preference temperature ranges and commonly used function combinations, ultimately generating a personalized user profile that includes preference temperature and commonly used function combinations.

[0045] The core of this step is to associate and bind user identity information with preference features, build a standardized mapping relationship, and dynamically generate personalized user profiles that can be updated in real time. This provides a core basis for predicting subsequent air conditioning operation intentions. The specific implementation method is as follows: Using the unique user identifier from the user identification results as the core index, data such as preferred temperature ranges, frequently used function combinations, and scene trigger probabilities from the preference feature statistics are bound one-to-one with the user identifier to establish a stable mapping relationship. The mapping relationship clearly marks the triggering environmental conditions corresponding to each preferred temperature range. For example, the preferred temperature range of 22℃ to 24℃ corresponds to scenarios where the in-vehicle ambient temperature is below 20℃ and the light intensity is below 300 lux, while the preferred temperature range of 26℃ to 28℃ corresponds to scenarios where the in-vehicle ambient temperature is above 28℃ and the PM2.5 concentration is above 50μg / m³. At the same time, each frequently used function combination is marked with the corresponding high-frequency triggering scenarios. For example, combination code C_1 has the highest trigger frequency in urban congested road sections and scenarios where the in-vehicle temperature is around 25℃, while combination code C_3 has the highest usage frequency in scenarios where the vehicle is traveling at high speed and the in-vehicle lighting is uniform.

[0046] After integration, the mapping relationships are standardized and encapsulated, redundant statistical data is removed, and four core information categories are retained: user identifier, preferred temperature ranges for multiple scenarios, commonly used function combinations, and rules for associating environment and preferences, forming a personalized user profile. The profile employs a dynamic update mechanism. Whenever the in-vehicle system collects new and valid air conditioning operation data from the current user, the new data is added to the historical behavior dataset in real time, statistical analysis and clustering operations are re-executed, and the preferred temperature ranges and function combination information in the profile are updated synchronously to ensure that the profile always reflects the user's latest usage habits.

[0047] The personalized user profile generated in the example includes the following details: User ID_1; preferred temperature range of 22℃ to 24℃ for low-temperature, low-light environments (0.80 probability); preferred temperature range of 26℃ to 28℃ for high-temperature, high-pollution environments (0.75 probability); and commonly used function combinations C_1 and C_3, with trigger probabilities of 0.35 and 0.28 respectively. The profile data format is compatible with the calling specifications of the vehicle system, allowing for rapid reading by subsequent time-series prediction models and achieving precise alignment between user habits and air conditioning display strategies.

[0048] S203, based on the user's personalized profile and real-time environmental parameters, predict the user's air conditioning operation intention in a future preset time period through a time-series prediction model, and generate an adaptive display strategy based on the prediction results. The strategy includes a function priority ranking and interface layout adjustment scheme. Specifically, preferred temperature ranges and frequently used function combinations can be extracted from user personalized profiles, while real-time environmental parameter data at the current moment can be collected to generate the input feature vector for the prediction model. The core of this step is to integrate personalized preference features formed from users' historical behavior with the current real-time state of the cockpit environment, transforming these two types of key information into standardized feature vectors. This provides accurate and well-organized input data for subsequent time-series prediction models, ensuring that the models can predict user operation intentions based on effective features. The specific implementation method is as follows: The user's personalized profile stores information on their unique air conditioning usage preferences, obtained through cluster analysis. When extracting features from this profile, the first step is to locate the unique data fields corresponding to the user's identifier. The preferred temperature range is a core feature representing the user's comfortable temperature under different environments. This range is represented by a continuous numerical range in degrees Celsius, with an accuracy controlled to 0.5 degrees Celsius. In the example, the extracted user preferred temperature range is 22.5 degrees Celsius to 24.5 degrees Celsius. This range is the core comfort range obtained by clustering frequently set temperature values ​​under different environmental conditions during multiple historical driving sessions, directly reflecting the user's core temperature needs. The frequently used function combination feature is the set of air conditioning functions that the user habitually uses while driving. This feature is stored in the form of function combination tags. In the example, the frequently used function combination extracted is the combination of automatic fan speed, windshield defroster, and recirculation mode. This combination is the most frequently selected function combination when the light intensity is high and the humidity inside the car is high. It belongs to the frequently used function mode. During the extraction process, temporary functions used only once are removed, and only function combinations with a frequency exceeding 60% are retained to ensure the representativeness of the features.

[0049] Real-time environmental parameter data is acquired using vehicle-mounted multimodal sensors. The acquisition time is the millisecond-level time node of the current system operation, with a timestamp accuracy of 1 millisecond, ensuring time synchronization between environmental data and user characteristics. The acquired real-time environmental parameters include four core indicators: temperature, humidity, light intensity, and PM2.5 concentration. Each parameter has a clearly defined numerical range and unit. The temperature parameter is in degrees Celsius, with a value range of 10 to 40 degrees Celsius; in this example, the current real-time temperature is 26 degrees Celsius. The humidity parameter is in percentage, with a value range of... The humidity ranges from 30% to 80%, with the current real-time humidity in the example being 65%; the light intensity is measured in lux, ranging from 0 lux to 100,000 lux, with the current real-time light intensity in the example being 35,000 lux; the PM2.5 concentration is measured in micrograms per cubic meter, ranging from 0 micrograms per cubic meter to 100 micrograms per cubic meter, with the current real-time PM2.5 concentration in the example being 35 micrograms per cubic meter. All parameters are calibrated after collection to remove outliers caused by sensor noise, ensuring that the parameters accurately reflect the current environmental conditions in the cockpit.

[0050] The construction of the prediction model input feature vector adopts a standardized feature concatenation method. The extracted upper limit value of the preferred temperature range T_h, the lower limit value of the preferred temperature range T_l, the commonly used function combination encoding value F_c, and the real-time temperature T_r, real-time humidity H_r, real-time light intensity L_r, and real-time PM2.5 concentration P_r are arranged in a fixed order to form a one-dimensional feature vector V_1. Among them, the function combination encoding value uses one-hot encoding logic to assign a unique value to different function combinations. In the example, the combination encoding value of automatic fan speed, windshield defrost, and internal circulation is 12. The final generated prediction model input feature vector is (22.5, 24.5, 12, 26, 65, 35000, 35). Each value in the vector corresponds to a clear feature meaning. The numerical dimensions are uniform and regular, with no missing or abnormal values, which can meet the requirements of the time series prediction model for input data format and lay the foundation for subsequent model calculations.

[0051] The prediction model is input into the feature vector and then into the pre-trained temporal prediction model. This model uses a long short-term memory network architecture to learn the temporal correlation between the user's historical operation patterns and environmental changes, and generates the probability distribution of operation intentions within a preset future time period. The core of this step is to utilize a pre-trained Long Short-Term Memory (LSTM) network model to mine the temporal dependencies between environmental changes and user operations in the input features. Through model inference, the probability of various air conditioning operations performed by the user in the future is calculated, forming a quantified probability distribution of operational intentions. This provides data support for subsequent function ranking. The specific implementation method is as follows: The pre-trained time-series prediction model is a deep learning model trained on a massive stream of historical user-related sensory data. During the model training phase, iterative optimization of the weight parameters was performed. The training data covers air conditioning operations and environmental changes from different seasons, time periods, and users. After convergence, the error rate is controlled within 5%, ensuring the accuracy of the prediction results. The Long Short-Term Memory (LSTM) network architecture is the core structure of this model. This architecture effectively handles the long-term dependency problem of time-series data, avoiding the gradient vanishing problem that occurs in traditional neural networks when processing time-series sequences. Internally, it contains three core gating units: an input gate, a forget gate, and an output gate. The forget gate filters out invalid environmental and operational association information from historical time-series data. The input gate receives the feature vector data at the current moment, and the output gate outputs the prediction result of the operation intent at the next moment. These three gating units work together to achieve deep learning of the temporal correlation between users' historical operation patterns and environmental changes.

[0052] After the generated prediction model is input into the feature vector, the model first normalizes the feature vector, mapping all feature values ​​to the range of 0 to 1. This eliminates computational bias caused by differences in the dimensions of different features. For example, light intensity of 35,000 lux is normalized to 0.35, PM2.5 concentration of 35 micrograms per cubic meter is normalized to 0.35, and values ​​such as temperature and humidity are also normalized proportionally. The normalized feature data is more suitable for the model's computational logic. The preset time period is the time range for model prediction, set at 30 seconds. This duration matches the response cycle of air conditioning operation during driving, ensuring that the prediction is neither too short (meaningless) nor too long (distorted).

[0053] The model uses its internal time-series feature extraction module to analyze the matching degree between current environmental parameters and user preference features. Combined with the user's operation patterns in similar environments in historical time-series data, it infers the air conditioning operations that the user may perform in the next 30 seconds, including more than ten operations such as temperature adjustment, air volume switching, air outlet mode change, circulation mode adjustment, defogging switch, and A / C cooling switch. It also generates a corresponding probability value for each operation, with the probability value ranging from 0 to 1. The closer the value is to 1, the higher the probability that the user will perform the operation. The probability distribution of the operation intentions generated by the model in the example is as follows: the probability of adjusting the temperature is 0.82, the probability of increasing the fan speed is 0.75, the probability of turning on the windshield defroster is 0.68, the probability of switching the internal circulation is 0.45, the probability of turning off the A / C cooling is 0.21, the probability of switching the air outlet mode is 0.13, and the probabilities of the remaining operations are all below 0.1. This probability distribution fully quantifies the user's operation tendencies in the future time period. All probability values ​​are obtained through the fully connected layer of the model and are normalized by the softmax function to ensure that the sum of all operation probability values ​​is 1, thus ensuring the rationality of the probability distribution.

[0054] Based on the probability distribution of operational intentions, a sorting algorithm is used to prioritize the various functions of the air conditioner from high to low according to the predicted probability of use, and a function priority sorting result is generated. The core of this step is to prioritize air conditioning functions based on the probability of user intent, using a sorting algorithm to categorize them by priority. Functions that users are more likely to use are listed as high priority, resulting in a clear hierarchical function sorting result. This provides a priority basis for subsequent interface layout design. The specific implementation method is as follows: The probability distribution of operational intent contains the predicted usage probability of all air conditioning functions. The sorting algorithm selected is the quicksort algorithm, which has the advantages of high computational efficiency and low time complexity. It can complete the sorting of function probabilities in milliseconds, which is suitable for the real-time operation requirements of the vehicle system. The core logic of the algorithm is to select the benchmark probability value in the probability distribution, divide the functions with probabilities higher than the benchmark value into the preorder sequence, and divide the functions with probabilities lower than the benchmark value into the postorder sequence. The overall sorting is completed through recursive iteration, and finally the descending order of probability from high to low is achieved.

[0055] During the sorting process, the effective functional items in the probability distribution of operation intentions are first extracted. Low-probability invalid functions with a probability value below 0.1 are removed to reduce the amount of sorting data and improve processing efficiency. In the example, the remaining effective functions after removal are six: temperature reduction, fan speed increase, windshield defogger, recirculation switch, A / C cooling off, and air outlet mode switch. Then, the probability values ​​of the six functions are sorted in descending order using a quicksort algorithm. The sorting process strictly follows the rule of comparing probability values, with functions having higher probability values ​​appearing earlier and having higher priority. After sorting, each function is assigned a corresponding priority number, represented by positive integers. The smaller the number, the higher the priority. In the example, the priority order of the sorted functions is as follows: priority 1 is temperature reduction, priority 2 is fan speed increase, priority 3 is windshield defogger, priority 4 is recirculation switch, priority 5 is A / C cooling off, and priority 6 is air outlet mode switch.

[0056] The function priority ranking results are stored in a standardized format. Each function is associated with a corresponding priority number and predicted probability value, ensuring that subsequent interface layout modules can directly call the ranking information. At the same time, the ranking results are updated in real time. If the real-time environmental parameters change, the ranking algorithm will immediately re-execute the ranking after the model regenerates the probability distribution, ensuring that the priority results always match the current scene. There are no function omissions or ranking errors during the ranking process. The priority division of all functions is strictly based on the predicted probability value, ensuring the objectivity and accuracy of the ranking results.

[0057] Based on the functional priority ranking, combined with the screen size and interaction specifications of the air conditioner panel, an interface layout adjustment scheme is designed to place high-priority functions in easily accessible and prominent positions, ultimately generating an adaptive display strategy that includes both functional priority ranking and interface layout adjustment scheme.

[0058] The core of this step is to combine functional priorities, hardware screen parameters, and interaction design standards to plan the interface layout of the air conditioner panel, placing high-priority functions in easily accessible areas, and integrating priority and layout information to form a complete adaptive display strategy, providing an execution basis for subsequent interface rendering. The specific implementation method is as follows: The screen size of the air conditioning panel is the hardware foundation of the interface layout, measured in inches. In this example, the screen size is 10.25 inches with a resolution of 1920 pixels by 720 pixels. The screen display area is divided into three parts: the main operation area, the secondary operation area, and the secondary storage area. The main operation area is located in the lower center of the screen, within direct reach of the driver's fingers. This area occupies 60% of the total screen width and 50% of the total screen height, making it the most convenient position for the driver. The secondary operation area is located on both sides of the screen, each occupying 20% ​​of the total screen width, offering slightly less convenient operation. The secondary storage area is located at the top of the screen, occupying 20% ​​of the total screen height, requiring swiping or clicking to access, and is used to house lower-priority functions. Interaction specifications are a common design standard for automotive air conditioning panels, stipulating that the minimum touch size for function buttons is no less than 40 pixels by 40 pixels, and the button spacing is no less than 10 pixels to avoid misoperation due to buttons being too small or too dense. Simultaneously, the button size in the main operation area can be appropriately enlarged to improve operational comfort.

[0059] The interface layout adjustment scheme strictly follows the priority adaptation principle, placing high-priority functions (priority 1 to 3) in the main operation area, functions with priority 4 in the secondary operation area, and functions with priority 5 to 6 in the secondary storage area. Specifically, the temperature adjustment function (priority 1) is placed in the center of the main operation area, with a button size of 80 pixels by 80 pixels (maximum size); the fan speed adjustment function (priority 2) is placed to the right of the temperature adjustment function, with a button size of 70 pixels by 70 pixels; the windshield defogger function (priority 3) is placed to the left of the temperature adjustment function, with a button size of 70 pixels by 70 pixels; the recirculation switching function (priority 4) is placed in the secondary operation area on the left side of the screen, with a button size of 50 pixels by 50 pixels; the A / C cooling off function (priority 5) and the airflow mode switching function (priority 6) are stored in the secondary menu at the top of the screen, requiring the click of the storage icon to expand, with a button size of 40 pixels by 40 pixels. The position coordinates of all buttons are precisely calculated based on the screen resolution to ensure the layout conforms to interaction specifications and has no overlap.

[0060] The generation of an adaptive display strategy is a process of integrating the function priority ranking results with the interface layout adjustment scheme. The strategy first clarifies the priority number and predicted probability value of each function, then details the screen coordinates, size, and area of ​​each function button, forming a complete strategy text. This strategy is stored in a standardized format recognizable by the vehicle system and can be directly transmitted to the panel display control module. The strategy content balances user operation intention prediction with actual interaction experience, ensuring the convenience of high-probability functions while complying with the hardware limitations and interaction specifications of the vehicle screen. After strategy generation, layout verification is performed to check whether buttons exceed screen boundaries, whether spacing is adequate, and whether high-priority functions are located in convenient areas. Once verification is successful, it becomes the final adaptive display strategy, which can be used for subsequent dynamic mapping of virtual buttons and icons.

[0061] S204, according to the adaptive display strategy, dynamically map the virtual button layout and function icon display hierarchy of the air conditioner panel to generate a personalized interactive interface that matches the current scene and predicted intent.

[0062] Specifically, it can parse the function priority sorting results in the adaptive display strategy, assign corresponding display level coefficients to each function, assign the highest level coefficient to high priority functions, and generate a function display level mapping table; The core of this step is to accurately analyze the function priority ranking results within the adaptive display strategy, and match corresponding display level coefficients to different air conditioning functions according to their priority. This ultimately forms a one-to-one mapping relationship between functions and level coefficients, providing a core basis for subsequent virtual button layout and icon style adjustments. The specific implementation method is as follows: The adaptive display strategy parsing unit is the core processing module for this step. Its main function is to extract the air conditioning function priority sequence after sorting algorithm processing in the strategy. This sequence covers all commonly used functions of the car air conditioning system, including temperature adjustment, fan speed switching, air outlet mode adjustment, internal and external circulation switching, windshield defrosting, rear windshield heating, in-vehicle PM2.5 purification, seat ventilation and heating, etc. The parsing unit identifies the function identifiers in the sequence through a text feature matching algorithm. The algorithm performs accurate matching through a preset function keyword library, with a matching accuracy of over 99%, avoiding function identification errors or sorting misalignments, and ensuring that the priority ranking results are extracted completely and accurately.

[0063] The display hierarchy coefficient is a core parameter for measuring the importance of a function's display. This coefficient is set to a continuous range of positive integers, from 1 to 5. The larger the value, the higher the display hierarchy of the function and the stronger the user's visual attention. The highest level coefficient is fixed at 5, which is specifically allocated to the core functions with the highest predicted usage probability. The coefficient value decreases sequentially as the function priority decreases, with the lowest level coefficient being 1, corresponding to the secondary functions with the lowest predicted usage probability. In the actual allocation process, the priority ranking number is matched with the predicted usage probability value. The predicted usage probability value is a numerical value that represents the likelihood of the user's operation intention. The value ranges from 0 to 1. The closer the value is to 1, the higher the probability that the user will perform the operation. In the example, the function priority ranking and corresponding probability values ​​are as follows: temperature adjustment 0.92, fan speed switching 0.85, internal and external circulation switching 0.71, windshield defroster 0.68, PM2.5 purification 0.53, air outlet mode adjustment 0.47, rear window heating 0.32, and seat ventilation and heating 0.25. According to the allocation rules, the temperature adjustment function with the highest priority is assigned a level coefficient of 5, the fan speed switching function with the second priority is assigned a level coefficient of 4, the internal and external circulation switching function and the windshield defroster function with the third and fourth priorities are both assigned a level coefficient of 3, the PM2.5 purification function and the air outlet mode adjustment function with the fifth and sixth priorities are assigned a level coefficient of 2, and the rear window heating function and the seat ventilation and heating function with the last two priorities are assigned a level coefficient of 1.

[0064] The generation of the function display hierarchy mapping table involves associating and integrating four types of information: function name, priority sorting number, predicted usage probability value, and display hierarchy coefficient. The mapping relationship is stored in a coherent textual association format, rather than presented in a table format. Instead, each function's corresponding parameters are clearly stated through statements. For example, the temperature adjustment function corresponds to sorting number 1, probability value 0.92, and hierarchy coefficient 5, while the fan speed setting function corresponds to sorting number 2, probability value 0.85, and hierarchy coefficient 4. The mapping association of all functions is completed in sequence to form a complete function display hierarchy mapping relationship. This mapping relationship is stored in a standardized data format to ensure that subsequent modules can directly call it without data ambiguity.

[0065] Based on the function display hierarchy mapping table and interface layout adjustment scheme, generate virtual button layout instructions, determine the coordinate position and display size of each function button on the air conditioner panel screen, and generate a virtual button layout scheme. The core of this step is to combine the functional display hierarchy mapping relationship with the preset interface layout specifications to generate layout instructions that control the position and size of virtual buttons, accurately determine the coordinates and size of each button on the screen, and finally form a virtual button layout scheme that conforms to the interaction logic. The specific implementation method is as follows: The interface layout adjustment scheme is based on layout guidelines formulated according to the parameters of the air conditioning panel screen and the vehicle interaction specifications. The air conditioning panel adopts a vehicle central control touch display screen with 1920 horizontal pixels and 1080 vertical pixels, and a physical display size of 12.3 inches. The vehicle interaction specifications clearly require that the minimum spacing between virtual buttons be no less than 15 pixels to prevent accidental operation by the driver. At the same time, the minimum size of the touch response area of ​​a single button is no less than 80×80 pixels to ensure the sensitivity of touch recognition. The virtual button layout instruction generation module will simultaneously call the function display hierarchy mapping table and the interface layout adjustment scheme, and use a spatial layout optimization algorithm to divide the screen area. The algorithm divides the screen into four areas according to the display hierarchy coefficient: core interaction area, secondary core interaction area, regular interaction area, and secondary storage area. The core interaction area corresponds to functions with a hierarchy coefficient of 5, located in the lower center of the screen where the driver can easily touch it, with pixel coordinates ranging from 800 to 1120 horizontally and 600 to 880 vertically; the secondary core interaction area corresponds to functions with a hierarchy coefficient of 4, and is divided into... The first layer is located on the left and right sides of the core interaction area, with pixel coordinates ranging from 480 to 780 and 1140 to 1440 horizontally and from 600 to 880 vertically. The second layer is located on the upper half of the screen, corresponding to the functions of layer coefficient 3, with pixel coordinates ranging from 320 to 620 and 1220 to 1520 horizontally and from 300 to 580 vertically. The third layer is located at the corners of the screen, corresponding to the functions of layer coefficients 2 and 1, with pixel coordinates ranging from 160 to 300 and 1540 to 1760 horizontally and from 100 to 280 vertically.

[0066] The coordinates and display size of the virtual buttons are set according to the level coefficient. The coordinates are in pixels on the screen, using the center coordinates of the horizontal X-axis and the vertical Y-axis. In the example, the temperature adjustment button (level coefficient 5) has center coordinates of X_960, Y_740, located in the center of the core interaction area; the fan speed switch button (level coefficient 4) has center coordinates of X_600, Y_740, located on the left side of the secondary core interaction area; the internal / external air circulation switch and the windshield defroster button (level coefficient 3) are also located in the same area. The center coordinates of the buttons for PM2.5 purification and airflow mode adjustment (level 2) are X_460, Y_440 and X_1380, Y_440, respectively, located on the left and right sides of the regular interaction area. The center coordinates of the buttons for PM2.5 purification and airflow mode adjustment (level 2) are X_220, Y_180 and X_1620, Y_180, respectively, located on the upper side of the secondary storage area. The center coordinates of the buttons for rear window heating and seat ventilation heating (level 1) are X_180, Y_120 and X_1700, Y_120, respectively, located at the corner of the secondary storage area. The display size follows the rule that the higher the level, the larger the size. The button size for level 5 is 120×120 pixels, level 4 is 100×100 pixels, level 3 is 90×90 pixels, level 2 is 70×70 pixels, and level 1 is 60×60 pixels. All sizes meet the minimum requirements for in-vehicle touch interaction.

[0067] The virtual button layout instructions include the area affiliation, center coordinate parameters, and display size parameters for each function button. The instructions are transmitted to subsequent modules in a standardized text format. The layout scheme integrates the position and size information of all buttons and checks the button spacing and overlap through a layout verification algorithm. The algorithm will automatically fine-tune coordinate values ​​that exceed the specifications to ensure that there is no button overlap and the spacing is compliant. After the verification is passed, the final virtual button layout scheme is formed.

[0068] Based on the function display hierarchy mapping table, the display style of function icons is dynamically adjusted, including high-priority functions using highlighted colors and dynamic effects, and low-priority functions using simplified icons or being included in the secondary menu, generating a function icon display configuration. The core of this step is to adjust the visual style of icons based on the functional display hierarchy coefficients, enhance the visual effects of high-priority functions, and simplify or consolidate low-priority functions, ultimately forming an icon display configuration that adapts to the hierarchy. The specific implementation method is as follows: The function icon style adjustment module will call up the function display hierarchy mapping table and match the preset visual display rules according to the hierarchy coefficient. The color display adopts the HSV color mode, where the hue H value ranges from 0 to 360, the saturation S value ranges from 0 to 100%, and the brightness V value ranges from 0 to 100%. The standard for setting the bright colors is high saturation and high brightness, while low priority functions use regular colors with low saturation and low brightness. In the example, the temperature adjustment icon (level 5) has a color tone of H_210, saturation of S_90%, and brightness of V_95%, presenting a bright blue highlight visual effect; the fan speed switch icon (level 4) has a color tone of H_120, saturation of S_85%, and brightness of V_90%, presenting a bright green highlight effect; the function icon (level 3) has a color tone of H_30, saturation of S_80%, and brightness of V_85%, presenting a warm yellow semi-highlight effect; and the function icons (levels 2 and 1) have a color tone of H_0, saturation of S_30%, and brightness of V_60%, presenting a light gray conventional visual effect to avoid interfering with the visual focus of high-priority functions.

[0069] The loading of dynamic effects uses a frame animation generation algorithm. The types of dynamic effects are divided into three types: breathing flash, pulse amplification, and rotation prompt. The dynamic effect cycle and frame rate are set to parameters adapted to in-vehicle displays. The frame rate is fixed at 30 frames per second to ensure smooth and lag-free dynamic effects. The core function of level 5 uses a breathing flashing animation effect with a cycle of 1.5 seconds, meaning it completes a cycle of alternating light and dark every 1.5 seconds to continuously attract the driver's attention. The secondary core function of level 4 uses a pulse amplification animation effect with a cycle of 2 seconds. The icon will be slightly enlarged and then shrunk by 1.1 times its original size to enhance recognizability. The functions of level 3 do not load any dynamic animation effects and remain in a static display state. The functions of level 2 use icon simplification processing, removing the decorative outline of the icon through a line simplification algorithm, retaining only the core identification graphic. For example, the PM2.5 purification icon is simplified to only retain the fan-shaped purification core logo, omitting the circular decorative lines. The functions of level 1 are directly included in the second-level menu. The first-level interface only displays a folded menu icon with a size of 40×40 pixels. Only after clicking the icon can the full icon be expanded to be viewed.

[0070] The function icon display configuration integrates information such as the color HSV parameters, animation type and cycle, icon simplification level, and whether it is included in the secondary menu for each function. The algorithm checks the rationality of the style through an icon visual verification algorithm. The algorithm controls the animation cycle within the range of 0.5 to 2 seconds and the color brightness does not exceed 95% to avoid problems such as glaring colors and excessively fast animations that cause visual fatigue. After verification, a complete function icon display configuration is generated, which can be directly matched and linked with the virtual button layout scheme.

[0071] By integrating the virtual button layout scheme and function icon display configuration, a personalized interactive interface is generated through the vehicle display screen driver module, and finally the air conditioning panel display interface that matches the current scene and predicted intent is output.

[0072] The core of this step is to fuse the virtual button layout information with the icon display configuration, and then render the interface through the vehicle display driver module to finally output an air conditioning panel display interface that fits the current environment and the user's operating intentions. The specific implementation method is as follows: The interface data integration module receives the virtual button layout scheme and function icon display configuration. It uses a time-series data matching algorithm to associate the two types of data one-to-one by function name, ensuring that the coordinates and size parameters of each function button are accurately bound to the corresponding icon color, animation effect, and simplification rules, without any mismatches. During the integration process, all parameters are converted into integer numerical formats recognizable by the vehicle display screen, retaining numerical precision to the nearest whole number to eliminate rendering errors caused by decimal places. The vehicle display screen driver module is the core of the vehicle's central control system's display control. It is responsible for receiving the integrated interface data and calling the 2D vector rendering engine to perform interface drawing operations. The rendering engine's working resolution is consistent with the screen's physical resolution, at 1920×1080 pixels, with a stable rendering frame rate of 60 frames per second, ensuring smooth interface transitions without ghosting or stuttering.

[0073] The rendering process follows a layer-priority drawing logic. First, the screen background layer is drawn, using a light gray background with a brightness of V_50% to reduce visual interference with the function icons. Then, each functional element is drawn sequentially according to its display layer coefficient from high to low. First, the temperature adjustment button and its highlighted breathing animation icon (layer 5) are drawn. Next, the function buttons and static icons for layers 4 and 3 are drawn in sequence. Finally, the simplified icon for layer 2 and the secondary menu icon for layer 1 are drawn. This drawing order ensures that high-priority functional elements are always displayed on the top layer of the interface and are not obscured by other elements. The personalized interactive interface generated after rendering perfectly matches the current cockpit environment and the predicted user's air conditioning operation intentions. For example, if the current cockpit temperature is low and the predicted user intention is to increase the air conditioning temperature, the temperature adjustment button is placed in a prominent position with a highlighted animation, followed closely by the fan speed adjustment button, perfectly matching the user's operating habits and needs.

[0074] Finally, the integrity verification is completed through the interface output verification module. The verification content includes the accuracy of button coordinates, the matching degree of icon style, the running status of animation, and the absence of missing or disordered interface elements. After the verification is passed, the air conditioning panel display interface is output to the vehicle touch screen in real time. The interface response delay is controlled within 30 milliseconds to ensure timely interface feedback when the driver operates. Finally, an intelligent and personalized air conditioning panel display interface that is highly matched with the current scene and the predicted user operation intention is presented.

[0075] Another embodiment of the present invention provides an intelligent display system for an automotive air conditioning panel with an environmental self-learning function, see [link to relevant documentation]. Figure 3 The system may include: The acquisition module 301 is used to acquire environmental parameters and user operation data in the cockpit through on-board multimodal sensors, and generate associated perception data streams based on timestamps; The construction module 302 is used to dynamically construct a personalized user profile that includes a combination of preferred temperature and commonly used functions based on the associated perception data stream and the user identity recognition result. The prediction module 303 is used to predict the user's air conditioning operation intention in a future preset period of time based on the user's personalized profile and real-time environmental parameters through a time-series prediction model, and generate an adaptive display strategy based on the prediction results. The strategy includes a function priority ranking and interface layout adjustment scheme. The generation module 304 is used to dynamically map the virtual button layout and function icon display hierarchy of the air conditioner panel according to the adaptive display strategy, and generate a personalized interactive interface that matches the current scene and predicted intent.

[0076] This invention also provides a storage medium storing a computer program, wherein the computer program is configured to execute the steps in any of the above method embodiments when running.

[0077] This invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0078] Specifically, the aforementioned electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the aforementioned processor, and the input / output device is connected to the aforementioned processor.

[0079] The above description, based on the embodiments shown in the figures, details the structure, features, and effects of the present invention. The above description is only a preferred embodiment of the present invention, but the present invention is not limited to the scope of implementation shown in the figures. Any changes made in accordance with the concept of the present invention, or equivalent embodiments modified to have equivalent changes, that do not exceed the spirit covered by the specification and figures, should be within the protection scope of the present invention.

Claims

1. A method for intelligent display of an automotive air conditioning panel with environmental self-learning function, characterized in that, The method includes: The vehicle uses onboard multimodal sensors to collect environmental parameters and user operation data in the cockpit, and generates associated perception data streams based on timestamps. Based on the associated perception data stream and combined with the user identity recognition results, a personalized user profile containing preferred temperature and commonly used function combinations is dynamically constructed. Based on the user's personalized profile and real-time environmental parameters, the user's air conditioning operation intention in the future preset time period is predicted by the time-series prediction model, and an adaptive display strategy is generated by combining the prediction results. The strategy includes a function priority ranking and interface layout adjustment scheme. Based on the adaptive display strategy, the virtual button layout and function icon display hierarchy of the air conditioner panel are dynamically mapped to generate a personalized interactive interface that matches the current scene and predicted intent.

2. The method according to claim 1, characterized in that, The process of collecting environmental parameters and user operation data within the cockpit using onboard multimodal sensors and generating a correlated perception data stream based on timestamps includes: Real-time data on temperature, humidity, light intensity, and PM2.5 concentration in the cockpit are collected using temperature and humidity sensors, light sensors, and air quality sensors to generate a raw environmental parameter dataset. The system collects user operation data on the air conditioning system through the touch screen interaction module and physical buttons, including temperature set value, fan speed level, air outlet mode and circulation mode selection, and generates the original user operation dataset. A unified clock source is used to add timestamps to each data point in the original environmental parameter dataset and the original user operation dataset. The two types of data are then aligned along the time axis using a data fusion algorithm to generate a time-synchronized multimodal dataset. A sliding window segmentation process is applied to the time-synchronized multimodal dataset to extract the correlation between the environmental parameter sequence and the operation event sequence within a fixed time window, ultimately generating a correlated perception data stream.

3. The method according to claim 2, characterized in that, The step of dynamically constructing a personalized user profile, including preferred temperature and frequently used function combinations, based on the associated sensing data stream and user identity recognition results includes: The driver's facial image is captured by an in-vehicle camera, and facial recognition algorithm is used to extract facial features and compare them with a pre-registered user database to generate user identity recognition results. Based on the user identification results, historical operation records belonging to the current user are filtered from the associated perception data stream, and environmental parameters and operation actions corresponding to the operation time are extracted to generate a user historical behavior dataset. Statistical analysis was performed on the user's historical behavior dataset, and clustering algorithms were used to identify the user's preferred temperature range and commonly used function combination patterns under different environmental conditions, generating statistical results of preference features; By integrating user identification results with preference feature statistics, a mapping relationship is established between user identifiers and preference temperature ranges and commonly used function combinations, ultimately generating a personalized user profile that includes preference temperature and commonly used function combinations.

4. The method according to claim 3, characterized in that, Based on the user's personalized profile and real-time environmental parameters, a time-series prediction model is used to predict the user's air conditioning operation intentions within a preset future time period. An adaptive display strategy is then generated based on the prediction results. This strategy includes a function priority ranking and interface layout adjustment scheme, including: The system extracts preferred temperature ranges and frequently used function combinations from the user's personalized profile, while simultaneously collecting real-time environmental parameter data to generate the input feature vector for the prediction model. The prediction model is input into the feature vector and then into the pre-trained temporal prediction model. This model uses a long short-term memory network architecture to learn the temporal correlation between the user's historical operation patterns and environmental changes, and generates the probability distribution of operation intentions within a preset future time period. Based on the probability distribution of operational intentions, a sorting algorithm is used to prioritize the various functions of the air conditioner from high to low according to the predicted probability of use, and a function priority sorting result is generated. Based on the functional priority ranking, combined with the screen size and interaction specifications of the air conditioner panel, an interface layout adjustment scheme is designed to place high-priority functions in easily accessible and prominent positions, ultimately generating an adaptive display strategy that includes both functional priority ranking and interface layout adjustment scheme.

5. The method according to claim 4, characterized in that, The step of dynamically mapping the virtual button layout and function icon display hierarchy of the air conditioner panel according to the adaptive display strategy to generate a personalized interactive interface that matches the current scene and predicted intent includes: The function priority ranking result in the adaptive display strategy is analyzed, and corresponding display level coefficients are assigned to each function. High-priority functions are assigned the highest level coefficients, and a function display level mapping table is generated. Based on the function display hierarchy mapping table and interface layout adjustment scheme, generate virtual button layout instructions, determine the coordinate position and display size of each function button on the air conditioner panel screen, and generate a virtual button layout scheme. Based on the function display hierarchy mapping table, the display style of function icons is dynamically adjusted, including high-priority functions using highlighted colors and dynamic effects, and low-priority functions using simplified icons or being included in the secondary menu, generating a function icon display configuration. By integrating the virtual button layout scheme and function icon display configuration, a personalized interactive interface is generated through the vehicle display screen driver module, and finally the air conditioning panel display interface that matches the current scene and predicted intent is output.

6. A smart display system for an automotive air conditioning panel with environmental self-learning function, characterized in that, The system includes: The data acquisition module is used to collect environmental parameters and user operation data in the cockpit through on-board multimodal sensors, and generate associated perception data streams based on timestamps; The module is used to dynamically construct a personalized user profile that includes preferred temperature and commonly used function combinations based on the associated perception data stream and the user identity recognition results. The prediction module is used to predict the user's air conditioning operation intention in a future preset time period based on the user's personalized profile and real-time environmental parameters through a time-series prediction model, and generate an adaptive display strategy based on the prediction results. The strategy includes a function priority ranking and interface layout adjustment scheme. The generation module is used to dynamically map the virtual button layout and function icon display hierarchy of the air conditioner panel according to the adaptive display strategy, and generate a personalized interactive interface that matches the current scene and predicted intent.

7. The system according to claim 6, characterized in that, The acquisition module is specifically used for: Real-time data on temperature, humidity, light intensity, and PM2.5 concentration in the cockpit are collected using temperature and humidity sensors, light sensors, and air quality sensors to generate a raw environmental parameter dataset. The system collects user operation data on the air conditioning system through the touch screen interaction module and physical buttons, including temperature set value, fan speed level, air outlet mode and circulation mode selection, and generates the original user operation dataset. A unified clock source is used to add timestamps to each data point in the original environmental parameter dataset and the original user operation dataset. The two types of data are then aligned along the time axis using a data fusion algorithm to generate a time-synchronized multimodal dataset. A sliding window segmentation process is applied to the time-synchronized multimodal dataset to extract the correlation between the environmental parameter sequence and the operation event sequence within a fixed time window, ultimately generating a correlated perception data stream.

8. The system according to claim 7, characterized in that, The building module is specifically used for: The driver's facial image is captured by an in-vehicle camera, and facial recognition algorithm is used to extract facial features and compare them with a pre-registered user database to generate user identity recognition results. Based on the user identification results, historical operation records belonging to the current user are filtered from the associated perception data stream, and environmental parameters and operation actions corresponding to the operation time are extracted to generate a user historical behavior dataset. Statistical analysis was performed on the user's historical behavior dataset, and clustering algorithms were used to identify the user's preferred temperature range and commonly used function combination patterns under different environmental conditions, generating statistical results of preference features; By integrating user identification results with preference feature statistics, a mapping relationship is established between user identifiers and preference temperature ranges and commonly used function combinations, ultimately generating a personalized user profile that includes preference temperature and commonly used function combinations.

9. A storage medium, characterized in that, The storage medium stores a computer program, wherein the computer program is configured to execute the method of any one of claims 1-5 when it is run.

10. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to run the computer program to perform the method of any one of claims 1-5.

Images