Intelligent test evaluation method, device and equipment for automobile parts

Abstract

The application discloses an intelligent test evaluation method, device and equipment for automobile parts, and relates to the technical field of intelligent management, and the method comprises the following steps: collecting volatile organic compound component concentration data in a vehicle in real time through an electronic nose sensor array, and synchronously acquiring biological neural signals of a driver; inputting the data into a dynamic game evaluation model to obtain an adjustment instruction; starting a purification device when the VOC concentration exceeds a threshold value and the brain electrical inhibition feature reaches a preset intensity; and starting the tactile feedback and smell release device to work cooperatively when the VOC concentration and the pupil change feature have a low correlation degree.The application solves the technical problem that the existing volatile organic compound detection in a vehicle only relies on chemical sensing data, lacks active adjustment of physiological state linkage, and cannot realize individualized comfort management, and achieves the technical effect that the volatile organic compound concentration and the physiological signals of the driver are cooperatively evaluated through multi-modal sensing and dynamic game, and the comfort management precision is improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent management technology, and more specifically to intelligent testing and evaluation methods, devices, and equipment for automotive parts. Background Technology

[0002] With the continuous development of automotive technology, in-vehicle air quality management has become a crucial factor in enhancing the driving experience and ensuring health. Volatile organic compounds (VOCs) are common pollutants in vehicles, potentially originating from various sources such as interior materials and air conditioning systems, posing a threat to the health of drivers and passengers. Existing in-vehicle air quality monitoring largely relies on chemical sensors to detect VOC concentrations. This method only reflects the objective level of environmental pollution and fails to consider individual differences in occupant sensitivity to odors and subjective discomfort, leading to significant discrepancies between the detected data and actual human perception. Furthermore, the placement of a single sensor is susceptible to the influence of cabin airflow organization and temperature stratification, making it difficult to accurately capture the true exposure level in the occupant's breathing zone. Moreover, the lack of a linkage mechanism with human physiological responses prevents the precise and proactive adjustment of air conditioning and fragrance systems, failing to meet the technical demands for personalized comfort in an intelligent cockpit environment. Summary of the Invention

[0003] This application provides intelligent testing and evaluation methods, devices, and equipment for automotive parts, which solve the technical problem that existing in-vehicle volatile organic compound detection relies solely on chemical sensing data and lacks active adjustment linked to physiological states, thus failing to achieve personalized comfort management.

[0004] The first aspect of this application provides an intelligent testing and evaluation method for automotive components. The method includes: real-time acquisition of volatile organic compound (VOC) concentration data within the cabin using an electronic nose sensor array deployed in the headrest, dashboard, and headliner; simultaneous acquisition of driver bio-neural signals, including fingertip blood oxygen fluctuation characteristics monitored by a steering wheel photoplethysmography sensor, pupil dynamic change characteristics captured by an infrared camera, and posture adjustment characteristics detected by a seat pressure distribution matrix; inputting the VOC concentration data and the bio-neural signals into a dynamic game evaluation model and outputting adjustment commands; according to the adjustment commands, when the concentration of a specific component in the VOC concentration data exceeds a dynamic threshold and the brainwave inhibition characteristics in the bio-neural signals reach a preset intensity, triggering the deployment of a hidden purification device; and when the correlation between the VOC concentration data and the pupil dynamic change characteristics is below a critical value, activating a tactile feedback device and an odor release device to work collaboratively.

[0005] A second aspect of this application provides an intelligent testing and evaluation device for automotive parts, the device comprising: an organic component concentration monitoring module for real-time acquisition of volatile organic compound (VOC) concentration data in the cabin via an electronic nose sensor array arranged in the seat headrest, dashboard, and headliner; a driver biosignal acquisition module for synchronously acquiring driver bioneural signals, including fingertip blood oxygen fluctuation characteristics monitored by a steering wheel photoplethysmography sensor, pupil dynamic change characteristics captured by an infrared camera, and posture adjustment characteristics detected by a seat pressure distribution matrix; a dynamic game evaluation module for inputting the VOC concentration data and the bioneural signals into a dynamic game evaluation model and outputting adjustment commands; a hidden purification module for triggering the deployment of a hidden purification device when the concentration of a specific component in the VOC concentration data exceeds a dynamic threshold and the EEG inhibition characteristics in the bioneural signals reach a preset intensity, according to the adjustment commands; and a collaborative working module for activating a tactile feedback device and an odor release device to work collaboratively when the correlation between the VOC concentration data and the pupil dynamic change characteristics is lower than a critical value.

[0006] A third aspect of this application provides an electronic device comprising: a processor coupled to a memory for storing a program, wherein when the program is executed by the processor, the device is configured to perform the method described in any of the first aspects.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: The intelligent testing and evaluation method, device, and equipment for automotive parts provided in this application relate to the field of intelligent management technology. By constructing a multimodal fusion system of electronic nose array and biological neural signals, and realizing the adaptive weight allocation of objective chemical detection and subjective comfort feeling through a dynamic game evaluation model, and triggering the coordinated adjustment of purification, tactile, and odor devices based on a dual threshold judgment mechanism, a closed-loop intelligent control of in-vehicle air quality is formed. This solves the technical problem that existing in-vehicle volatile organic compound detection relies solely on chemical sensing data and lacks active adjustment linked to physiological state, thus failing to achieve personalized comfort management. It achieves the technical effect of improving the accuracy of comfort management by realizing the coordinated evaluation of volatile organic compound concentration and driver physiological signals through multimodal sensing and dynamic game. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 A schematic diagram of the intelligent testing and evaluation method for automotive parts provided in this application embodiment; Figure 2 A schematic diagram of the structure of an intelligent testing and evaluation device for automotive parts provided in an embodiment of this application; Figure 3 This application provides a schematic diagram of the structure of an electronic device.

[0010] Figure reference numerals: Organic component concentration monitoring module 11, driver biosignal acquisition module 12, dynamic game evaluation module 13, hidden purification module 14, collaborative work module 15, electronic device 300, memory 301, processor 302, communication interface 303, bus architecture 304. Detailed Implementation

[0011] This application provides intelligent testing and evaluation methods, devices, and equipment for automotive parts, which solve the technical problem that existing in-vehicle volatile organic compound detection relies solely on chemical sensing data and lacks active adjustment linked to physiological states, thus failing to achieve personalized comfort management.

[0012] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0013] It should be noted that the terms "first," "second," etc., in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, apparatus, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or devices.

[0014] Example 1, as Figure 1 As shown, this application provides an intelligent testing and evaluation method for automotive parts, the method comprising: P10: Real-time data collection of volatile organic compound concentrations in the cabin is achieved through an array of electronic nose sensors located in the seat headrests, instrument panel, and ceiling.

[0015] Specifically, an electronic nose sensor array collects real-time data on the concentration of volatile organic compounds (VOCs) in the vehicle cabin. This array consists of various types of chemical sensors, including metal oxide semiconductor sensors, electrochemical sensors, and photoionization detectors. These sensors work together to cover the detection needs of different VOC components, such as formaldehyde, benzene, toluene, and ethanol. By combining multiple sensor types, false positives caused by the cross-sensitivity of a single sensor can be effectively avoided, improving the accuracy and specificity of the detection.

[0016] The electronic nose sensor array consists of multiple sensor modules, each employing a different sensing technology, such as semiconductor sensors, metal oxide sensors, or optical sensors. Each sensor has different selectivity, enabling it to detect specific types of gas molecules. By reacting to these gas molecules, the sensors generate corresponding signals, which are used to calculate the specific VOC concentration. To enhance the system's accuracy, the sensor array is calibrated in real time based on changes in the temperature and humidity of the vehicle's interior environment, avoiding the influence of environmental factors on the measurement data.

[0017] The electronic nose sensor array employs a distributed arrangement, installed in three key locations: the headrest, the dashboard, and the headliner. The sensors in the headrest primarily target the driver's breathing zone, monitoring the concentration of low-density gases such as formaldehyde, an area prone to volatile substance accumulation due to respiration and thermal convection. The sensors in the dashboard, located near interior trim components such as plastics and leather, capture the concentration of benzene compounds released from these materials; these substances release at higher rates under high temperatures. The sensors in the headliner utilize thermal convection to monitor the settling trends of high-density volatile organic compounds, such as some aldehydes, ensuring coverage of areas with significant concentration gradients within the cabin. This arrangement was optimized based on computational fluid dynamics simulations to match the gas diffusion patterns within the cabin.

[0018] During data acquisition, the system uses a timing synchronization module to align the output signals of each sensor, eliminating data delays caused by differences in installation location. Simultaneously, recognizing the susceptibility of chemical sensors to environmental temperature and humidity, a temperature-humidity compensation algorithm is employed to preprocess the raw signals. This algorithm, based on a polynomial fitting correction model, dynamically adjusts the sensor's response curve by real-time monitoring of cabin temperature and humidity parameters, reducing the impact of environmental fluctuations on detection accuracy. After preprocessing, the electronic nose array outputs a multi-dimensional concentration vector, including data on multiple dimensions such as formaldehyde, benzene, and toluene concentrations, rather than a single total value. This allows subsequent steps to differentiate the varying effects of different components on the driver's physiological state.

[0019] Through the arrangement and operation of this sensor array, the system can monitor the air quality inside the vehicle in all aspects and provide real-time environmental data support, providing a basis for subsequent automatic adjustment and intelligent decision-making.

[0020] P20: Simultaneously acquire driver bio-neural signals, including fingertip blood oxygen fluctuation characteristics monitored by steering wheel photoplethysmography sensor, pupil dynamic change characteristics captured by infrared camera, and posture adjustment characteristics detected by seat pressure distribution matrix.

[0021] It should be understood that multimodal sensing devices are used to simultaneously acquire the driver's bio-neural signals to capture the body's physiological responses to changes in the cabin environment. Specifically, three different sensor technologies can be used to monitor the driver's physiological responses: a steering wheel photoplethysmography sensor to monitor fingertip blood oxygen fluctuations, an infrared camera to capture dynamic changes in pupil size, and a seat pressure distribution matrix to detect posture adjustments. These signals will provide more accurate and personalized data support for subsequent in-vehicle environment adjustments.

[0022] First, a photoplethysmography (PPG) sensor on the steering wheel is used to monitor blood oxygen fluctuations in the driver's fingertips. This sensor estimates the blood oxygen concentration in the fingertips by detecting the light absorption characteristics in the blood. Specifically, photoplethysmography (PPG) uses a combination of a light source and a photodiode sensor to measure changes in blood volume in blood vessels by reflecting light through the skin, thereby calculating blood oxygen fluctuations. Fluctuations in blood oxygen concentration in the fingertips can reflect the driver's physiological state, such as fatigue or stress. Therefore, real-time monitoring of this characteristic helps to assess the driver's health and emotional state. By comparing changes in blood oxygen fluctuations, it is possible to identify whether the driver is under stress or tension, allowing for timely environmental adjustments. For example, a photoplethysmography sensor is integrated into the steering wheel grip area. This sensor uses a dual-wavelength light emission module, emitting red light at 660 nm and near-infrared light at 880 nm, and receiving reflected light signals from the fingertip tissue. By calculating the periodicity of light absorption rate changes, the fluctuation characteristics of blood oxygen saturation are extracted, with the fluctuation frequency range set from 0.01 Hz to 0.15 Hz to cover low-frequency autonomic nervous system regulatory signals. When the driver is stimulated by volatile organic compounds, sympathetic nerve excitation may lead to peripheral vasoconstriction, manifested as a decrease in the amplitude and prolongation of blood oxygen fluctuations. The sensor records these characteristics in real time at a sampling rate of 100 Hz.

[0023] Next, the dynamic changes in pupil size are captured by an infrared camera. Changes in the pupil, especially its response under different lighting conditions, can reflect the driver's emotions and physiological state. The infrared camera captures these dynamic changes non-contactly, including pupil dilation and constriction, reflecting the driver's attention, fatigue level, and emotional fluctuations. For example, pupil constriction may be associated with high stress or intense focus, while pupil dilation may indicate relaxation or fatigue. By analyzing this data on pupil dynamic changes, the driver's emotional state can be assessed and used as a reference for comfort adjustment. For example, an infrared camera mounted in the center of the dashboard captures images of the driver's eyes at a rate of 30 frames per second. The camera is equipped with an 850nm active near-infrared light source to avoid visible light interference. The image processing module first locates the pupil region using a Haar feature classifier, then calculates the pupil diameter using an ellipse fitting algorithm with an accuracy of 0.1 mm. The pupil diameter is tracked over time, and feature parameters including baseline diameter, maximum constriction rate, and recovery half-life are extracted. For example, when formaldehyde is present, the pupil may contract rapidly and then recover slowly; this characteristic is correlated with the concentration of the chemical substance.

[0024] Finally, the seat pressure distribution matrix detects the driver's posture adjustment characteristics through an array of pressure sensors built into the seat. This array monitors the pressure distribution across various parts of the driver's body in contact with the seat in real time, determining if their posture has changed. The seat pressure distribution data reflects the driver's physical condition, such as whether they have been in an uncomfortable posture for an extended period or have adjusted their posture to alleviate discomfort. By analyzing pressure changes, it's possible to identify whether the driver has changed their posture due to discomfort, thereby assessing their comfort level. If prolonged posture adjustments or uncomfortable postures are detected, the system can suggest or automatically adjust the seat angle and height to improve driver comfort. For example, a pressure distribution matrix sensor is embedded in the seat surface. This matrix consists of 16 rows × 16 columns of pressure-sensitive capacitive units, each with a resolution of 256 pressure levels. The system acquires pressure distribution maps at a frequency of 10 times per second, identifying posture adjustment characteristics by calculating the movement trajectory of the pressure center point and the variance of local pressure changes. When the driver feels discomfort, the frequency of pressure center point movement typically increases, and the variance of pressure distribution in the buttocks area increases; these characteristics are quantified as a posture adjustment intensity index.

[0025] By simultaneously acquiring these three bio-neural signals, a comprehensive understanding of the driver's physiological responses and comfort levels can be obtained, providing a more accurate and personalized basis for subsequent in-vehicle environment adjustments. Real-time data from each sensor will be integrated and analyzed to provide a multi-dimensional assessment of the driver's physical condition, mood, and health.

[0026] Furthermore, in processing the driver's biological neural signals, step P20 of this embodiment also includes: P21: Perform multi-scale decomposition on the fingertip blood oxygen fluctuation characteristics, extract specific frequency band components related to autonomic nervous system activity, and eliminate motion artifacts caused by steering wheel grip force; P22: Establish a dual-channel evaluation model for pupil dynamic changes to process the pupil dynamic change characteristics; P23: The dual-channel evaluation model includes a brightness adaptation channel and an emotion response channel. The brightness adaptation channel compensates for changes in ambient light intensity through iris texture features, and the emotion response channel identifies aversion response characteristics through the mapping relationship between pupil oscillation frequency and a known emotion database.

[0027] Optionally, the driver's bio-neural signals, including fingertip blood oxygen fluctuation characteristics, pupil dynamic change characteristics, and posture adjustment characteristics, can be further processed to more accurately assess the driver's comfort and emotional state.

[0028] First, the pulse oximetry characteristics of the fingertip are decomposed into multiple scales. Pulse oximetry signals are captured by photoplethysmography sensors and typically contain multiple frequency components. Through multi-scale decomposition, the signal can be broken down into different scales to extract specific frequency band components related to autonomic nervous system activity. Autonomic nervous system activity often manifests as low-frequency fluctuations; therefore, during multi-scale decomposition, the system focuses on extracting low-frequency components to better reflect the driver's physiological and emotional changes. These low-frequency components help reveal the driver's state of tension, stress, or relaxation. For example, an empirical mode decomposition algorithm is used to decompose the original signal into multiple intrinsic mode functions (IMFs), from which components with a frequency range of 0.01 Hz to 0.15 Hz are selected, corresponding to the low-frequency regulatory activity of the autonomic nervous system.

[0029] Furthermore, steering wheel grip force can cause motion artifacts, interfering with the accuracy of blood oxygen fluctuation signals. Changes in steering wheel grip force introduce high-frequency components, causing artifacts and affecting signal clarity. By using multi-scale decomposition and calculating the sample entropy values ​​of each component, motion artifact components caused by changes in steering wheel grip force can be identified and eliminated, such as signal step interference caused by sudden changes in grip force. This ensures that the final extracted blood oxygen fluctuation signal is more accurate and can truly reflect the driver's physiological response.

[0030] Next, a dual-channel evaluation model is established to handle the dynamic changes in the driver's pupils. Pupil changes are typically influenced by both ambient light intensity variations and emotional responses; therefore, a multi-channel analysis of pupil dynamics is necessary. First, the brightness adaptation channel adjusts the analysis results of pupil changes based on ambient light variations. Since pupil contraction and dilation are normal physiological responses to changes in ambient light intensity inside and outside the vehicle, they may not fully reflect emotional states. Therefore, the brightness adaptation channel compensates for the impact of ambient light variations on the pupils by analyzing pupil iris texture features, such as the contrast and spatial frequency of iris folds. This allows the system to eliminate interference from light intensity changes and accurately capture dynamic pupil changes.

[0031] Simultaneously, pupillary oscillation frequency is further analyzed through the emotional response channel. Pupillary oscillation is a high-frequency response in pupillary changes, usually closely related to emotional reactions such as anxiety and anger. By mapping pupillary oscillation frequency to an emotional database, features associated with negative emotions such as aversion can be identified. For example, when faced with unpleasant stimuli or situations, the driver's pupillary oscillation frequency may change significantly. Based on these changes in oscillation frequency, the emotional response channel determines whether the driver is in a negative emotional state and then adjusts the system's response strategy accordingly.

[0032] Ultimately, the dual-channel evaluation model accurately identifies the driver's physiological and emotional state, taking into account both ambient light variations and emotional fluctuations. The combined use of the brightness adaptation channel and the emotional response channel allows the system to accurately capture the driver's emotional reactions, especially signals related to negative emotions such as aversion and anxiety. The two types of outputs are weighted and combined to form the final pupil dynamic evaluation index, which considers both environmental adaptability and emotional specificity, avoiding misjudgment based on a single feature. For example, when ambient light intensity is stable but the probability of an aversion reaction suddenly increases, it is determined to be a physiological response caused by chemical stimulation, rather than the influence of light changes. All processing steps can employ a real-time sliding window mechanism with a window length of 5 seconds and a step size of 1 second to ensure timely feature updates. Model parameters are calibrated through pre-experiments; for example, the similarity threshold for the emotional response channel is set to 0.75, and a value higher than this triggers an aversion reaction flag.

[0033] By comprehensively processing these steps, the driver's physiological condition and emotional response can be fully assessed, providing accurate data support for subsequent adjustments to the in-vehicle environment.

[0034] P30: Input the volatile organic compound concentration data and the biological neural signal into the dynamic game evaluation model, and output the adjustment command.

[0035] Furthermore, in the dynamic game evaluation model, step P30 of this application embodiment also includes: P31: The first player's payoff function generates an air quality deviation score by calculating the Euclidean distance between the concentrations of each volatile organic compound component and the preset standard concentration curve; P32: The second player's payoff function generates a physiological discomfort score by weightedly fusing the frequency domain energy distribution of the fingertip blood oxygen fluctuation characteristics, the contraction rate of the pupil dynamic change characteristics, and the frequency of the sitting posture adjustment characteristics; P33: The third player's payoff function generates an energy efficiency score by establishing a linear relationship matrix between the external particulate matter concentration gradient and the power consumption of the air conditioning system.

[0036] Specifically, the pre-processed volatile organic compound (VOC) concentration data and biological neural signals are input into a dynamic game-theoretic evaluation model. This model generates regulatory commands through multi-objective optimization. The dynamic game-theoretic evaluation model includes three players, representing demands in three dimensions: air quality, physiological comfort, and energy efficiency. The model comprehensively evaluates the balance between air quality, driver physiological state, and energy efficiency through the payoff functions of the three players, ultimately outputting regulatory commands to achieve intelligent environmental control.

[0037] In this system, the payoff function for the first player is calculated by euclidean distance between the concentrations of each volatile organic compound (VOC) component and a preset standard concentration curve, generating an air quality deviation score. The purpose of this air quality deviation score is to assess the difference between the VOC concentrations in the vehicle's interior and the predetermined health standard concentrations. The preset standard concentration curve represents the ideal concentration range of various harmful gases that should be maintained in the vehicle. By calculating the euclidean distance between the actual concentration data and the standard concentration, the degree of air quality deviation can be quantified. This deviation score serves as the payoff function for the first player, reflecting the potential impact of the current in-vehicle air quality on occupants. The greater the deviation, the worse the air quality, and the stronger the need for model adjustments.

[0038] For example, firstly, a preset standard concentration curve is set based on relevant standards, including curves showing the changes in limits for components such as formaldehyde, benzene, and toluene over time. For each volatile organic compound component, the Euclidean distance between the current concentration sequence and the standard curve is calculated, i.e., the square root of the sum of the squares of the concentration differences at each point. Subsequently, the distance values ​​of each component are normalized, and weights are assigned according to the component's toxicity coefficient. For example, formaldehyde, which has higher toxicity, is assigned a weight of 0.4, benzene is assigned a weight of 0.3, and other components share the remaining weights. After weighted summation, an air quality deviation score is generated, with a higher score indicating a more severe pollution level.

[0039] Next, the payoff function of the second player generates a physiological discomfort score by weightedly fusing the frequency domain energy distribution of fingertip blood oxygen fluctuations, the constriction rate of pupil dynamics, and the frequency of posture adjustments. This scoring system focuses on the driver's physiological responses, comprehensively analyzing physiological characteristics such as blood oxygen fluctuations, pupil dynamics, and posture adjustments to assess the driver's comfort and health status from multiple dimensions. Specifically, the frequency domain energy distribution of fingertip blood oxygen fluctuations reflects the driver's physiological stress, while the pupil constriction rate reveals their emotional response. The frequency of posture adjustments reflects whether the driver adjusts their posture due to discomfort. By weightedly fusing these features, the model can comprehensively consider the driver's physiological discomfort and generate a comprehensive physiological discomfort score. This score reflects the driver's current comfort level and health status, providing data support for subsequent environmental adjustments.

[0040] For example, firstly, frequency domain analysis is performed on the fingertip blood oxygen fluctuation characteristics to extract the energy integral value of a specific low-frequency band; the pupillary contraction rate per unit time is calculated for the pupillary dynamic change characteristics; and the frequency of occurrence within a time window is statistically analyzed for the sitting posture adjustment characteristics. Then, the three feature values ​​are normalized to the 0-1 range and fused according to pre-calibrated weights: blood oxygen fluctuation energy weight 0.4, pupillary contraction rate weight 0.3, and sitting posture adjustment frequency weight 0.3. The weighted fusion yields a physiological discomfort score, with an increased score indicating increased physiological discomfort for the driver.

[0041] Finally, the payoff function of the third party generates an energy efficiency score by establishing a linear relationship matrix between the external particulate matter concentration gradient and the power consumption of the air conditioning system. This step focuses on evaluating the energy efficiency of the air conditioning system, especially under the influence of changes in external particulate matter concentration. The external particulate matter concentration gradient directly affects the air quality inside the vehicle; typically, when the particulate matter concentration is high, the air conditioning system requires higher power consumption for air purification and regulation. By establishing a linear relationship between the particulate matter concentration gradient and the power consumption of the air conditioning system, the system can calculate the energy efficiency score in real time. This score helps optimize the operation of the air conditioning system, ensuring a balance between energy efficiency and comfort. A higher energy efficiency score indicates that the air conditioning system can effectively control the in-vehicle environment under low energy consumption conditions.

[0042] For example, firstly, PM2.5 concentration gradient data is acquired using an external particulate matter sensor, and a pre-calibrated linear relationship matrix is ​​queried. This matrix stores the theoretical minimum power consumption of the air conditioning system corresponding to different concentration gradients. Then, the ratio of the theoretical minimum power consumption corresponding to the current external particulate matter concentration to the actual power consumption of the air conditioning system is calculated. A higher ratio indicates better energy efficiency. For instance, when the actual power consumption is close to the theoretical minimum, the score is close to 1.0; if the actual power consumption exceeds the theoretical value by 50%, the score drops to 0.67.

[0043] By using the payoff functions of the three players in the game, the dynamic game evaluation model can comprehensively assess the relationship between in-vehicle air quality, driver's physiological state, and air conditioning system energy efficiency. The model generates adjustment commands based on these evaluation indicators to achieve intelligent adjustment of the in-vehicle environment. In practical applications, when the model assesses a high deviation in air quality, driver discomfort, or insufficient energy efficiency, the system will automatically optimize the in-vehicle environment. For example, it may adjust the power of the air conditioning system to improve air quality, or adjust temperature and humidity to enhance driver comfort, thereby achieving a more personalized and intelligent in-vehicle environment management system.

[0044] Furthermore, in the construction of the dynamic game evaluation model, step P30 of this application embodiment also includes: P34: In a closed testing environment, volatile organic compounds of different concentration gradients are generated using a controllable release device. Simultaneously, changes in the driver's facial micro-expressions, finger electromyography signals, and respiratory rhythm are recorded as test data. P35: Based on the recorded test data, a three-stage parameter optimization mechanism is established, including: P35-1: In the first stage, the initial weights of each player are adjusted using the gradient descent method so that the model output matches the expert evaluation results to the first threshold. P35-2: In the second stage, a time decay factor is introduced to exponentially decay the contribution of historical data. P35-3: In the third stage, an online learning mechanism is established so that when the actual adjustment effect deviates from the predicted value by more than the second threshold, the weights are recalibrated.

[0045] In one possible embodiment of this application, the construction of the dynamic game evaluation model further includes a model parameter optimization process, which improves the model's accuracy through controlled experiments and a phased optimization mechanism. By collecting volatile organic compound component data at different concentration gradients in a closed test environment, while simultaneously recording the driver's physiological response data, rich test data can be provided for the model. Through a three-stage parameter optimization mechanism, the model can continuously self-adjust in practical applications, ensuring that its output adjustment commands accurately reflect changes in the in-vehicle environment and the driver's needs.

[0046] First, a controlled release device is deployed in a closed testing environment. This device can precisely release volatile organic compound (VOC) components at different concentration gradients, such as formaldehyde at 0.02 ppm, 0.05 ppm, and 0.1 ppm, and benzene at 0.01 ppm, 0.03 ppm, and 0.06 ppm. By controlling the release device, multiple different air quality conditions can be created in the simulated in-vehicle environment. These different concentrations of VOCs affect the in-vehicle air quality and produce different physiological responses at different concentrations. Simultaneously, changes in the driver's facial micro-expressions, finger electromyography (EMG) signals, and respiratory rhythm are recorded under these conditions. These physiological signals are important indicators reflecting the driver's emotions and physiological state. For example, facial micro-expressions can reflect the driver's emotional fluctuations, such as anxiety or unease; finger EMG signals can show the physiological responses caused by tension or discomfort; and changes in respiratory rhythm directly reflect the driver's physiological comfort or stress level. By recording these physiological signals, the system obtains rich test data, providing a basis for subsequent model optimization.

[0047] Next, based on the recorded test data, a three-stage parameter optimization mechanism was established. The goal is to continuously optimize the model's parameters so that it can adapt to changes in the in-vehicle environment and accurately output adjustment commands.

[0048] In the first stage, the initial weights of each player in the model are adjusted using gradient descent. Gradient descent, as an optimization algorithm, continuously adjusts the model's parameters by calculating the error between the model's output and the actual expert evaluation results, making the model's predictions as close as possible to the expert evaluations. The goal of this stage is to allow the model to initially align with the expert evaluation results, ensuring that the model can produce reasonable outputs in the early training phase. For example, firstly, the dataset is divided into a training set and a validation set, with expert evaluation results used as labels. Experts comprehensively assess the discomfort level based on physiological data. The difference between the model output and the expert evaluation is calculated using mean squared error loss, and the weights are iteratively updated using gradient descent. For example, the initial weights for the air quality player are 0.4, the physiological discomfort player's weight is 0.4, and the energy consumption player's weight is 0.2. Iteration stops when the validation set's fit reaches a first threshold of 95%.

[0049] The second stage introduces a time decay factor to exponentially reduce the contribution of historical data. Over time, changes in the in-vehicle environment, the driver's physiological state, and the external environment can gradually reduce the reference value of historical data. Through the time decay factor, the system can dynamically adjust the influence of historical data in model calculations, allowing newer data to have a greater impact on model predictions. This mechanism ensures that the model responds more quickly to changes in the current environment and the driver, without being affected by outdated data, thereby improving the model's real-time adaptability and adjustment accuracy. For example, a decay coefficient of 0.8 is set, meaning that the weight of historical data decreases by 20% every 24 hours. This mechanism ensures that the model prioritizes recent data, adapting to changes in driver physiological characteristics, such as sensitivity differences caused by seasonal variations.

[0050] The third stage establishes an online learning mechanism. When the actual adjustment effect deviates from the predicted value by more than a set second threshold, the model automatically triggers a weight recalibration process. This process is based on real-time data feedback, ensuring that the model can learn and improve in real time. Specifically, when the system detects a significant gap between the output adjustment command and the actual effect, it will recalibrate the weights through the online learning mechanism, adjusting the model's parameters so that it can more accurately output adjustment commands that meet actual needs in the next prediction. For example, the actual adjustment effect is evaluated through subsequent sensor data. For instance, if the rate of decrease in volatile organic compound concentration is less than 30% of the predicted value after the purification device is activated, or if the physiological discomfort score does not decrease as expected, the recalibration process is triggered. The calibration process uses an incremental learning algorithm, fine-tuning the weights based on the latest 100 sets of data to avoid the computational burden of global retraining. All optimization stages use 5-fold cross-validation, with 4-fold data used for training and 1-fold data for validation in each iteration, and the average of the 5 validation results is taken as the model performance index. The optimized model weights are stored in the storage area of ​​the vehicle control unit and a version number is set, supporting comparison of version differences during OTA updates. This process is a continuous optimization cycle, where the model constantly corrects itself based on real-time feedback, thereby improving its accuracy in practical use.

[0051] Through this three-stage parameter optimization mechanism, the system can adapt to constantly changing in-vehicle environment, driver status, and external conditions during continuous optimization and adjustment. Ultimately, the dynamic game evaluation model can output reasonable adjustment commands based on the complex relationship between volatile organic compound concentration, physiological signals, and energy efficiency requirements, thereby achieving intelligent control of the in-vehicle environment.

[0052] Furthermore, to optimize the weight allocation relationship of the return function, step P35 of this application embodiment also includes: P35-4: Establish an external air quality classification table, dividing particulate matter concentration into multiple continuous intervals, with each interval corresponding to a different energy consumption weight coefficient adjustment step size; P35-5: When a change in external particulate matter concentration across intervals is detected, dynamically adjust the weight coefficient of the third party according to the step size corresponding to the current interval, while limiting the adjustment range to not exceed the preset upper limit value.

[0053] Optionally, to further optimize the weight allocation relationship of the revenue function, a dynamic adjustment mechanism for the classification of external air quality and energy consumption weights can be established. This process ensures that the model continuously optimizes the energy efficiency of in-vehicle air conditioning under the influence of changes in the external environment, balancing the relationship between comfort and energy consumption.

[0054] First, an outdoor air quality classification table is established, dividing particulate matter concentration into multiple continuous intervals. Specifically, outdoor particulate matter concentration is typically affected by factors such as weather and traffic emissions, exhibiting a certain degree of fluctuation. Therefore, in order to dynamically adjust the in-vehicle environment regulation strategy based on changes in outdoor air quality, the system divides particulate matter concentration into multiple intervals, such as: a low concentration interval (0-50). g / m 3 ), medium concentration range (51-150) g / m 3 ) and high concentration range (151) g / m 3 (Above). Each concentration range represents a different air quality level, and the energy efficiency requirements for the vehicle's air conditioning system also differ. The system defines a corresponding energy consumption weighting coefficient adjustment step size for each range, which determines the extent to which the air conditioning system adjusts its energy consumption. For example, in a low concentration range, the outside air quality is good, and the air conditioning system can appropriately reduce power consumption; in a high concentration range, the particulate matter concentration is high, and the system needs to increase the air conditioning power to filter the air, resulting in a corresponding increase in the energy consumption weighting coefficient. Through this hierarchical management, the system can flexibly adjust the air conditioning energy efficiency settings according to different particulate matter concentration levels.

[0055] When a change in the concentration of particulate matter outside the vehicle is detected across different zones, the system dynamically adjusts the weighting coefficients of the third-party player based on the step size of the current particulate matter concentration zone. Specifically, when the particulate matter concentration moves from a low concentration zone to a medium concentration zone, or from a medium concentration zone to a high concentration zone, the system adjusts the weighting coefficients according to the step size corresponding to the current zone. For example, if the outside air quality moves from a low concentration zone to a medium concentration zone, the system increases the energy consumption weighting coefficient of the air conditioner according to a preset step size to enhance air purification capabilities. If the outside air quality continues to deteriorate, the system adjusts again according to the preset step size to ensure that the air conditioner can provide sufficient air quality improvement capabilities. For example, when the concentration rises from an excellent zone to a lightly polluted zone, a step size of 0.02 is used, increasing the weighting coefficient of the third-party player by 0.02; if it falls from a heavily polluted zone to a moderately polluted zone, a step size of 0.03 is used, decreasing the weighting coefficient by 0.03. Simultaneously, to prevent drastic fluctuations in weighting, an upper limit of 0.1 is set for the adjustment range, meaning that a single adjustment cannot exceed 0.1. The adjusted weighting coefficients take effect immediately and are used in the next round of dynamic game calculations. By setting an upper limit on the adjustment range, the system can maintain a smooth transition and ensure that the air conditioning system can adapt to environmental changes in an orderly manner, even when particulate matter concentrations fluctuate significantly.

[0056] By dynamically adjusting the energy consumption of the air conditioning system according to different concentration levels of outside air quality, it is possible to maximize energy efficiency while ensuring comfort inside the vehicle.

[0057] Furthermore, in verifying the adjustment command, step P30 of this application embodiment also includes: P36: Within a preset time window after the odor release device is activated, the driver's nasal airflow voiceprint features are collected through a high-sensitivity microphone array; P37: The airflow rate change rate and respiratory interval variation coefficient in the voiceprint features are extracted and matched with a preset healthy breathing pattern database; P38: When the matching similarity is lower than the third threshold, the output power of the negative ion generator is gradually increased, while the driver's new round of bioneural signal response is recorded.

[0058] Specifically, the effectiveness of the adjustment commands is further verified by analyzing acoustic characteristics to assess changes in the driver's breathing state, ensuring that the driver's physiological responses are effectively improved through adjustment methods such as the odor release device and negative ion generator. Specifically, the driver's nasal airflow acoustic signature is collected using a high-sensitivity microphone array and compared with a healthy breathing pattern database to verify the adjustment effect. If the driver's breathing pattern is found to be inconsistent with their healthy state, the system will gradually increase the output power of the negative ion generator based on the verification results to further adjust the in-vehicle environment and optimize the driver's physiological responses.

[0059] Specifically, the system uses a high-sensitivity microphone array to collect the driver's nasal airflow voiceprint characteristics. When the odor release device is activated, the system starts the microphone array to collect data within a preset time window. This microphone array is positioned appropriately within the vehicle to clearly capture the driver's nasal airflow voiceprint characteristics. Airflow voiceprint characteristics are nasal sound patterns caused by airflow and reflect the driver's breathing state. When the driver breathes normally, the speed and rhythm of nasal airflow maintain a certain regularity, while during discomfort or emotional fluctuations, airflow changes may become irregular. Through the high-sensitivity microphone array, the system can acquire this voiceprint data in real time, providing a basis for subsequent physiological analysis.

[0060] Next, the airflow rate change rate and respiratory interval variation coefficient are extracted from the voiceprint features and matched against a preset healthy breathing pattern database. The airflow rate change rate represents the degree of fluctuation in the driver's breathing airflow rate over a certain period of time; the respiratory interval variation coefficient reflects the irregularity of the driver's breathing interval time. The airflow rate change rate is obtained by calculating the first derivative of the sound signal envelope, reflecting the instantaneous change in breathing force; the respiratory interval variation coefficient is calculated by detecting the duration of continuous breathing cycles and calculating the ratio of the standard deviation to the mean, reflecting the stability of the breathing rhythm. Under normal circumstances, the breathing rate and interval time of a healthy driver maintain a certain regularity, while under stress, anxiety, or discomfort, the airflow rate and breathing interval will fluctuate significantly. Therefore, the system compares the collected airflow rate change rate and respiratory interval variation coefficient with the preset healthy breathing pattern database, which contains parameters of various typical healthy breathing patterns. When the airflow features match the healthy patterns in the database, it indicates that the driver's physiological state is relatively stable; conversely, if the matching similarity is lower than the set third threshold, the system will consider that the driver may be experiencing discomfort or stress, prompting further adjustments to the in-vehicle environment.

[0061] When the detected matching similarity is below the third threshold, the output power of the negative ion generator is gradually increased. The negative ion generator improves the air quality inside the vehicle by releasing negative ions, which have the effects of improving air quality, promoting smooth breathing, and reducing fatigue and stress. If the system detects that the driver's breathing pattern deviates from the healthy mode, indicating that the driver may be under stress or discomfort, the output power of the negative ion generator will be gradually increased to promote the driver's physiological recovery. Each power increase is based on the current physiological signal feedback, ensuring that the adjustment process is smooth and gradually optimized. At the same time, the system will simultaneously record the driver's new round of bio-neural signal responses, including facial micro-expressions, blood oxygen fluctuations, and dynamic changes in pupil size. These signals will provide further data support for subsequent system adjustments and optimizations. For example, when the matching similarity is below the third threshold of 0.7, it is determined that the adjustment effect has not met expectations. The system gradually increases the output power of the negative ion generator, with an initial power of 0.5W, increasing by 0.2W each time, with a maximum of 2.0W. Simultaneously, within 10 seconds after each power adjustment, the driver's steering wheel blood oxygen fluctuation characteristics and seat pressure distribution characteristics are re-acquired, recording a new round of bio-neural signal responses. If the similarity remains below the threshold after three consecutive power increases, an alarm signal is triggered and a fault code is recorded.

[0062] Furthermore, step P30 in this embodiment also includes a personalized adaptation process: P39: Establish a time-series-based driver sensitivity feature matrix, with dimensions including the reaction threshold of each volatile organic compound component, physiological response delay time, and the attenuation curve of the regulatory effect; P310: When the system detects that the time interval between multiple consecutive manual interventions shows a shortening trend, automatically increase the weight of the pupil dynamic change feature in the payoff function of the second player.

[0063] Optionally, a personalized adaptation process can be further implemented to optimize model parameters by continuously learning the driver's sensitivity characteristics. These processes help to dynamically adjust the system's regulation strategies based on the driver's individual needs and reaction patterns, thereby providing more personalized and precise environmental regulation.

[0064] First, a time-series-based driver sensitivity feature matrix is ​​established. The rows of this matrix correspond to different volatile organic compounds (VOCs) such as formaldehyde, benzene, and toluene, while the columns contain three feature parameters: the reaction threshold, which records the concentration of the component at which the driver first experiences a physiological response (e.g., when the formaldehyde concentration reaches 0.06 ppm and the pupillary constriction rate exceeds the baseline by 20%, the formaldehyde reaction threshold is updated to 0.06 ppm); the physiological response delay time, which records the time difference from the concentration exceeding the threshold to a significant change in the physiological signal, with an accuracy of 0.1 seconds; and the adjustment effect decay curve, which records the time required for the physiological discomfort score to drop to 50% after the purification device is activated, fitted with an exponential decay function. The matrix data is automatically updated every 24 hours, smoothed using a moving average method, and retains historical data from the most recent 30 days.

[0065] To better adapt to the individual needs of different drivers, the system updates the values ​​in the sensitivity feature matrix based on historical data and real-time driver responses. By recording drivers' reaction patterns to different environmental conditions through time-series analysis, the system adjusts the sensitivity parameters in the matrix in a timely manner, enabling it to dynamically optimize its adjustment strategies based on each driver's specific circumstances. For example, if the system detects an increased sensitivity of a driver to a certain VOC component, or a change in their physiological reaction delay time, it can automatically adjust the priority and response speed of air purification or other regulatory measures by updating the sensitivity feature matrix.

[0066] When the system detects that the driver has made multiple manual interventions, such as adjusting the air conditioning or fragrance system, with a shortening time interval, it automatically increases the weight of pupil dynamic change features in the payoff function of the second player. Manual interventions typically reflect driver dissatisfaction or discomfort with the current in-vehicle environment. Multiple adjustments to system settings may indicate increasing dissatisfaction. To better respond to this need, the system increases the weight of pupil dynamic change features, thereby enhancing its focus on driver emotions and comfort. Pupil dynamic change features reflect driver emotional fluctuations, especially the speed and frequency of pupil constriction and dilation, helping the system determine whether the driver is experiencing discomfort or stress. When the weight of pupil dynamic change features increases, the system adjusts the in-vehicle environment more sensitively, especially when the driver exhibits dissatisfaction or discomfort. For example, if the standard deviation of the time interval between the most recent five interventions is less than 15% of the average interval, and the mean interval shows a decreasing trend (e.g., the interval decreases from 120 minutes to 90 minutes and then to 70 minutes), it is determined that the driver's sensitivity has increased. At this point, the system automatically increases the weight of the pupil dynamic change feature in the payoff function of the second player, gradually increasing it from an initial weight of 0.3 to 0.4, while correspondingly reducing the weights of other features to ensure the sum is 1. After adjustment, the system continuously monitors for 3 driving cycles. If the frequency of manual intervention decreases, the new weight is locked; otherwise, the original weight is restored and a sensitivity review process is triggered. The review process includes recalibrating the reaction threshold and delay time.

[0067] This mechanism adaptively adjusts the weights in the payoff function of the game players, enabling the system to respond more flexibly to the driver's physiological and emotional needs when faced with multiple manual interventions, thereby enhancing the personalization of the in-vehicle environment. By dynamically adjusting the weights, the system can pay closer attention to the driver's physiological responses and promptly optimize environmental variables such as in-vehicle air quality, temperature, or humidity, thus ensuring a continuous improvement in driver comfort.

[0068] Furthermore, in updating the personalized adaptation process, step P30 of this embodiment also includes: P311: Locally encrypted storage of environmental parameters, execution instructions, and final scores for each regulation event; P312: When the number of accumulated regulation events reaches the batch processing threshold, extract the volatile organic compound concentration feature vectors of each event, add random noise perturbation, and upload them to the cloud aggregation server; P313: The server generates population fitness parameters through feature vector clustering analysis and distributes them to each terminal device for model fine-tuning.

[0069] It should be understood that further optimization of the personalized adaptation process, by adding data storage, feature extraction and cloud analysis, enables the system to continuously accumulate experience after each adjustment event, and to perform data aggregation and analysis in the cloud, thereby optimizing the group adaptation parameters and ultimately improving the personalized adjustment capabilities of the entire vehicle series.

[0070] First, the system locally encrypts and stores the environmental parameters, executed commands, and final scores for each adjustment event. Specifically, for each adjustment operation, such as adjusting air quality, temperature, humidity, or activating the fragrance / purification system, the system records the environmental parameters before and after the adjustment, the specific commands executed, and the user's feedback score, such as assessing driver comfort through physiological signals like facial micro-expressions and breathing patterns. This data will be encrypted and stored locally to protect user privacy. During data storage, the system encrypts all sensitive data to ensure data security and confidentiality and prevent data leakage.

[0071] When the number of accumulated adjustment events reaches a preset batch processing threshold, the system extracts volatile organic compound (VOC) component concentration feature vectors for each event. These feature vectors represent the in-vehicle air quality characteristics of each adjustment event, helping the system understand the impact of each air quality state on the driver's physiological response. After extracting the feature vectors, random noise perturbation is added to them to increase data diversity and prevent overfitting. This perturbation effectively enhances the model's generalization ability, allowing the system to adapt more flexibly to different driver and environmental conditions. Subsequently, the perturbed feature vectors are uploaded to a cloud aggregation server for further processing and analysis, with an upload frequency not exceeding once per day.

[0072] Finally, the cloud aggregation server generates population adaptation parameters by performing cluster analysis on the uploaded feature vectors. These parameters are obtained by analyzing data collected from multiple terminal devices, revealing commonalities in drivers' general air quality needs, comfort responses, and adjustment strategies. Through cluster analysis, the server can identify differences in adaptability to the in-vehicle environment among different groups and optimize the adjustment strategies for the entire vehicle series based on these differences. For example, the system may find that a certain type of driver is particularly sensitive to specific VOC components, or that changes in air quality under certain conditions cause most drivers to experience discomfort. Based on these findings, the server generates a set of optimized adaptation parameters for model fine-tuning across all terminal devices. For instance, K-means cluster analysis is used to generate population adaptation parameters. The number of clusters is set to 5, representing low sensitivity, low-to-medium sensitivity, medium sensitivity, medium-to-high sensitivity, and high sensitivity groups, respectively. The server calculates the feature centroids for each group and generates corresponding model fine-tuning parameters; for example, the pupil change feature weight for the high-sensitivity group is increased to 0.45, and the reaction threshold is reduced by 20%. The fine-tuning parameters are sent to each terminal device after being digitally signed.

[0073] Finally, the generated group adaptability parameters will be distributed to each terminal device for model fine-tuning. Each terminal device will fine-tune its local model based on the parameters distributed from the cloud, thereby improving its responsiveness to the needs of individual drivers. The fine-tuning process not only improves the system's personalized adjustment effect but also optimizes the in-vehicle environment adjustment strategy based on group data, enabling each terminal device to optimize the adjustment effect by referring to group trends while ensuring personalized needs.

[0074] P40: According to the adjustment command, when the concentration of a specific component in the volatile organic compound concentration data exceeds the dynamic threshold and the brain electrical inhibition characteristics in the biological neural signal reach the preset intensity, the hidden purification device is triggered to unfold.

[0075] Specifically, the system manages in-vehicle air purification in real time according to adjustment commands and optimizes air quality control strategies by monitoring the efficiency of the purification device. Specifically, when the concentration of volatile organic compounds (VOCs) in the vehicle exceeds a dynamic threshold, and the driver's bio-neural signals show brainwave inhibition characteristics reaching a preset intensity, the system activates the hidden purification device and monitors its effectiveness in real time. If the system detects insufficient purification in certain areas, it can automatically adjust the angle of the air vent deflectors and increase the fan speed to ensure effective improvement of in-vehicle air quality.

[0076] First, when the concentration of certain harmful gases exceeds a dynamic threshold based on volatile organic compound (VOC) concentration data, and the driver's bio-neural signals (such as brainwaves) show signs of stress or discomfort, the concealed air purification device is triggered to deploy. Once deployed, the system immediately begins real-time monitoring of the device's operation, including airflow and particulate matter filtration efficiency. Using sensors placed in multiple locations within the cabin, the system calculates the settling rate of particulate matter within the vehicle to assess the overall effectiveness of the purification device. This data is then used to generate a heatmap of purification efficiency, showing the improvement in air quality at different locations. The heatmap reveals areas where pollutant concentrations decrease slowly, or areas where purification is ineffective—areas where the purification effect is minimal.

[0077] When the heat map shows that the rate of pollutant concentration decrease in certain purification dead zones falls below the preset fifth threshold, the system will automatically adjust. Specifically, the system will adjust the angle of the air outlet deflectors to make airflow more uniform, reduce the formation of airflow dead zones, and thus improve air purification efficiency. At the same time, the system will also increase the speed of the air conditioning fan as needed to enhance air circulation and particulate matter adsorption capacity. These adjustments ensure that the air inside the vehicle is effectively purified in all areas, reducing harmful substances in the air and improving the comfort, health, and safety of the driver and passengers.

[0078] Furthermore, the system records the parameter combinations and actual effects of each purification operation, forming an optimal strategy knowledge base. During each purification process, data including VOC concentration, fan speed, deflector angle, and purification effect are recorded, and the most effective purification strategy is identified through data analysis. These records will serve as a reference for future decision-making, helping the system continuously optimize in-vehicle air purification strategies and make intelligent adjustments based on actual in-vehicle conditions.

[0079] Through the above process, not only can the purification device be activated when the air quality inside the vehicle is detected to be substandard, but the purification effect can also be automatically optimized based on real-time monitoring results to ensure that the air quality in all areas of the vehicle is comprehensively improved.

[0080] P50: When the correlation between the concentration data of the volatile organic compounds and the dynamic change characteristics of the pupil is lower than the critical value, the tactile feedback device and the odor release device are activated to work together.

[0081] Optionally, the system can further optimize the in-vehicle environment by intelligently triggering the coordinated operation of the haptic feedback device and the odor release device based on the correlation between the concentration of volatile organic compounds (VOCs) and the dynamic changes in the driver's pupils. Specifically, when the correlation between the VOC concentration and the driver's dynamic pupil changes is detected to be lower than a set threshold, the system will activate the coordinated operation of the haptic feedback device and the odor release device. By adjusting the intensity of the haptic feedback and the concentration of the odor release, the system can enhance the driver's perception of changes in in-vehicle air quality while avoiding excessive stimulation.

[0082] First, a corresponding tactile vibration pattern coding library is generated based on the concentration gradient of various volatile organic compounds (VOCs) in the vehicle. This library sets different vibration patterns according to different VOC concentration gradients. For example, when the concentration of high-concentration irritant gases (such as formaldehyde or benzene) rises rapidly, the system generates a high-frequency vibration pattern through the tactile feedback device. High-frequency vibration can quickly transmit stimulation signals, alerting the driver to the rapid change in the air quality inside the vehicle and drawing their attention. The tactile feedback device applies vibrations of different frequencies and intensities through locations such as the seat, steering wheel, or headrest, allowing the driver to directly feel intuitive feedback on changes in the air quality inside the vehicle.

[0083] Simultaneously, an odor-tactile cross-adaptation model is established. This model ensures that the odor release device and tactile feedback device work together to avoid excessive sensory stimulation. Specifically, when the concentration of certain VOC components consistently exceeds a set fourth threshold, the system activates a progressive tactile intensity enhancement mode. Specifically, when the concentration of certain volatile organic compounds is excessively high and continues to exceed the standard, the system gradually increases the intensity of tactile feedback, providing the driver with a stronger physical stimulus to alert them that the in-vehicle air quality has reached a level requiring intervention. At the same time, to avoid overstimulating the driver's sense of smell, the release concentration of the odor release device is correspondingly reduced. By adjusting the odor release concentration, the system can maintain a balance in the overall stimulation, ensuring that the perceived intensity of both odor and tactile feedback is within an appropriate range to avoid causing discomfort to the driver.

[0084] This cross-adaptation mode of scent and touch makes in-vehicle air quality regulation more personalized and intelligent. By using tactile feedback and odor release devices in tandem, it not only provides multimodal environmental feedback but also optimizes adjustments based on actual air quality changes and the driver's physiological responses. When VOC concentration rises rapidly, tactile feedback can quickly attract the driver's attention, while the odor release device adjusts appropriately during this process to balance sensory stimulation.

[0085] The coordinated adjustment of touch and smell provides drivers with a more intuitive and personalized driving experience, helping them to react promptly when the in-vehicle environment deteriorates, thus improving driver comfort and health.

[0086] In summary, the embodiments of this application have at least the following technical effects: This application achieves precise capture of volatile organic compound (VOC) concentration in the occupant's breathing zone through a three-dimensional spatial layout of an electronic nose sensor array, eliminating detection blind spots caused by single sensor placement. By fusing VOC component concentration data with multi-dimensional bio-neural signals from the driver, a dynamic mapping relationship between objective chemical indicators and subjective comfort evaluation is established, overcoming perceptual biases caused by individual differences. Through multimodal weight adaptive allocation of a dynamic game-theoretic evaluation model, the adaptability and decision-making accuracy of the evaluation algorithm under different pollution scenarios are improved. A dual threshold judgment mechanism enables refined collaborative control of the purification device, tactile feedback device, and odor release device, balancing health protection and comfort experience. Through the deployable structure design of the concealed purification device, spatial concealment in non-triggered states and rapid response in triggered states are achieved, optimizing cabin space utilization efficiency.

[0087] It achieves the technical effect of improving the accuracy of comfort management by using multimodal sensing and dynamic game to coordinate the assessment of volatile organic compound concentration and driver physiological signals.

[0088] Example 2, based on the same inventive concept as the intelligent testing and evaluation method for automotive parts in the foregoing examples, such as... Figure 2 As shown, this application provides an intelligent testing and evaluation device for automotive parts. The device and method embodiments in this application are based on the same inventive concept. The device includes: The organic component concentration monitoring module 11 is used to collect real-time data on the concentration of volatile organic components in the cabin through an electronic nose sensor array arranged in the seat headrest, instrument panel and ceiling.

[0089] The driver biosignal acquisition module 12 is used to synchronously acquire the driver's bioneural signals, including the fingertip blood oxygen fluctuation characteristics monitored by the steering wheel photoplethysmography sensor, the pupil dynamic change characteristics captured by the infrared camera, and the sitting posture adjustment characteristics detected by the seat pressure distribution matrix.

[0090] The dynamic game evaluation module 13 is used to input the volatile organic compound concentration data and the biological neural signals into the dynamic game evaluation model and output adjustment instructions.

[0091] The hidden purification module 14 is used to trigger the deployment of the hidden purification device according to the adjustment command when the concentration of a specific component in the volatile organic compound concentration data exceeds the dynamic threshold and the brain electrical inhibition characteristics in the biological neural signal reach a preset intensity.

[0092] The collaborative working module 15 is used to activate the collaborative working of the tactile feedback device and the odor release device when the correlation between the concentration data of the volatile organic compounds and the dynamic change characteristics of the pupil is lower than a critical value.

[0093] Furthermore, the driver biosignal acquisition module 12 is also used to perform the following steps: The fingertip blood oxygen fluctuation characteristics are decomposed at multiple scales to extract specific frequency band components related to autonomic nervous system activity and eliminate motion artifacts caused by steering wheel grip force. A dual-channel evaluation model for pupil dynamic changes is established to process the pupil dynamic change characteristics. The dual-channel evaluation model includes a brightness adaptation channel and an emotion response channel. The brightness adaptation channel compensates for changes in ambient light intensity through iris texture features, and the emotion response channel identifies aversion response characteristics through the mapping relationship between pupil oscillation frequency and a known emotion database.

[0094] Furthermore, in the dynamic game evaluation module 13: The first game-playing payoff function generates an air quality deviation score by calculating the Euclidean distance between the concentrations of each volatile organic compound component and the preset standard concentration curve; the second game-playing payoff function generates a physiological discomfort score by weightedly fusing the frequency domain energy distribution of the fingertip blood oxygen fluctuation characteristics, the contraction rate of the pupil dynamic change characteristics, and the frequency of sitting posture adjustment characteristics; the third game-playing payoff function generates an energy efficiency score by establishing a linear relationship matrix between the external particulate matter concentration gradient and the power consumption of the air conditioning system.

[0095] Furthermore, the dynamic game evaluation module 13 is also used to perform the following steps: In a closed testing environment, volatile organic compounds of different concentration gradients are generated using a controlled release device. Simultaneously, changes in the driver's facial micro-expressions, finger electromyography signals, and respiratory rhythm are recorded as test data. Based on the recorded test data, a three-stage parameter optimization mechanism is established, including: The first stage uses gradient descent to adjust the initial weights of each player, so that the model output matches the expert evaluation results to the first threshold. The second stage introduces a time decay factor to exponentially decay the contribution of historical data. The third stage establishes an online learning mechanism that triggers weight recalibration when the actual adjustment effect deviates from the predicted value by more than the second threshold.

[0096] Furthermore, the dynamic game evaluation module 13 is also used to perform the following steps: A vehicle exterior air quality classification table is established, dividing particulate matter concentration into multiple continuous intervals, with each interval corresponding to a different energy consumption weight coefficient adjustment step size. When a change in vehicle exterior particulate matter concentration across intervals is detected, the weight coefficient of the third party is dynamically adjusted according to the step size corresponding to the current interval, while limiting the adjustment range to not exceed a preset upper limit value.

[0097] Furthermore, the dynamic game evaluation module 13 is also used to perform the following steps: Within a preset time window after the odor release device is activated, the driver's nasal airflow voiceprint features are collected through a high-sensitivity microphone array; the airflow rate change rate and respiratory interval variation coefficient in the voiceprint features are extracted and matched with a preset healthy breathing pattern database; when the matching similarity is lower than the third threshold, the output power of the negative ion generator is gradually increased, and the driver's new round of bioneural signal response is recorded.

[0098] Furthermore, the dynamic game evaluation module 13 also includes a personalized adaptation process: A time-series-based driver sensitivity feature matrix is ​​established, with dimensions including the reaction threshold of each volatile organic compound component, physiological response delay time, and the decay curve of the regulatory effect. When the system detects that the time interval between multiple consecutive manual interventions shows a shortening trend, the weight of the pupil dynamic change feature in the payoff function of the second player is automatically increased.

[0099] Furthermore, the dynamic game evaluation module 13 is also used to perform the following steps: The system encrypts and stores the environmental parameters, execution instructions, and final scores of each regulation event locally. When the number of accumulated regulation events reaches the batch processing threshold, it extracts the volatile organic compound concentration feature vectors of each event, adds random noise perturbation, and uploads them to the cloud aggregation server. The server generates population fitness parameters through feature vector clustering analysis and distributes them to each terminal device for model fine-tuning.

[0100] Example 3: Exemplary electronic device.

[0101] The following is for reference. Figure 3 The electronic device described in the embodiments of this application is used to illustrate this application.

[0102] Based on the same inventive concept as the intelligent testing and evaluation method for automotive parts in the foregoing embodiments, this application also provides an intelligent testing and evaluation device for automotive parts, comprising: a processor coupled to a memory for storing a program, wherein when the program is executed by the processor, the device performs the steps of the method described in Embodiment 1.

[0103] The electronic device 300 includes a processor 302, a communication interface 303, and a memory 301. Optionally, the electronic device 300 may also include a bus architecture 304. The communication interface 303, processor 302, and memory 301 can be interconnected via the bus architecture 304; the bus architecture 304 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus architecture 304 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0104] Processor 302 may be a CPU, microprocessor, ASIC, or one or more integrated circuits used to control the execution of programs according to the present application.

[0105] Communication interface 303 uses any transceiver-like device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area network (WLAN), wired access network, etc.

[0106] Memory 301 can be ROM or other types of static storage devices that can store static information and instructions, RAM or other types of dynamic storage devices that can store information and instructions, or it can be electrically erasable programmable read-only memory. EEPROM (Electronic EPROM) and Compact Discrete (CD-ROM) are read-only memory systems. Only memory, CD ROM or other optical disc storage, optical disk storage (including compressed optical discs, laser discs, optical discs, universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited to these. The memory may exist independently and be connected to the processor via bus architecture 304. The memory may also be integrated with the processor.

[0107] The memory 301 stores computer execution instructions for implementing the scheme of this application, and the processor 302 controls the execution. The processor 302 executes the computer execution instructions stored in the memory 301, thereby realizing the intelligent testing and evaluation method for automotive parts provided in the above embodiments of this application.

[0108] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0109] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0110] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A method for intelligent testing and evaluation of automotive parts, characterized in that, The method includes: The electronic nose sensor array, located in the seat headrests, instrument panel, and ceiling, collects real-time data on the concentration of volatile organic compounds in the cabin. Simultaneously acquire driver bioneural signals, including fingertip blood oxygen fluctuation characteristics monitored by steering wheel photoplethysmography sensor, pupil dynamic change characteristics captured by infrared camera, and posture adjustment characteristics detected by seat pressure distribution matrix. The concentration data of the volatile organic compounds and the biological neural signals are input into a dynamic game evaluation model, which outputs adjustment instructions. According to the adjustment command, when the concentration of a specific component in the volatile organic compound concentration data exceeds the dynamic threshold and the brain electrical inhibition characteristics in the biological neural signal reach a preset intensity, the hidden purification device is triggered to unfold. When the correlation between the concentration data of the volatile organic compounds and the dynamic changes in the pupil is lower than a critical value, the tactile feedback device and the odor release device are activated to work together.

2. The intelligent testing and evaluation method for automotive parts as described in claim 1, characterized in that, In the dynamic game evaluation model: The first player's payoff function generates an air quality deviation score by calculating the Euclidean distance between the concentrations of each volatile organic compound component and the preset standard concentration curves. The second game's payoff function generates a physiological discomfort score by weighting and fusing the frequency domain energy distribution of the fingertip blood oxygen fluctuation characteristics, the contraction rate of the pupil dynamic change characteristics, and the frequency of the sitting posture adjustment characteristics. The third game's payoff function generates an energy efficiency score by establishing a linear relationship matrix between the external particulate matter concentration gradient and the power consumption of the air conditioning system.

3. The intelligent testing and evaluation method for automotive parts as described in claim 1, characterized in that, The construction of the dynamic game evaluation model includes: In a closed testing environment, volatile organic compounds of different concentration gradients are generated through a controlled release device, while the driver's facial micro-expression changes, finger electromyography signals, and respiratory rhythm changes are recorded as test data. Based on the recorded test data, a three-stage parameter optimization mechanism is established, including: In the first stage, the initial weights of each player are adjusted using the gradient descent method so that the model output matches the expert evaluation results to the first threshold. The second stage introduces a time decay factor to exponentially decay the contribution of historical data. The third stage establishes an online learning mechanism, which triggers weight recalibration when the actual adjustment effect deviates from the predicted value by more than the second threshold.

4. The intelligent testing and evaluation method for automotive parts as described in claim 3, characterized in that, Optimize the weight allocation relationship of the revenue function, including: Establish an outdoor air quality classification table, dividing particulate matter concentration into multiple continuous intervals, with each interval corresponding to a different energy consumption weighting coefficient adjustment step size; When a change in the concentration of particulate matter outside the vehicle is detected across intervals, the weight coefficient of the third player is dynamically adjusted according to the step size corresponding to the current interval, while limiting the adjustment range to no more than a preset upper limit.

5. The intelligent testing and evaluation method for automotive parts as described in claim 1, characterized in that, The verification of the adjustment command includes: Within a preset time window after the odor release device is activated, the driver's nasal airflow voiceprint characteristics are collected using a high-sensitivity microphone array. Extract the airflow rate change rate and respiratory interval variation coefficient from the voiceprint features and match them with a preset healthy breathing pattern database; When the matching similarity is lower than the third threshold, the output power of the negative ion generator is gradually increased, while the driver's new round of biological neural signal response is recorded.

6. The intelligent testing and evaluation method for automotive parts as described in claim 1, characterized in that, The method also includes a personalized adaptation process: A time-series-based driver sensitivity feature matrix was established, with dimensions including the response threshold, physiological response delay time, and modulation effect decay curve for each volatile organic compound component. When the system detects that the time interval between multiple consecutive manual interventions is showing a shortening trend, it automatically increases the weight of the pupil dynamic change feature in the payoff function of the second player.

7. The intelligent testing and evaluation method for automotive parts as described in claim 6, characterized in that, Updating the personalization process includes: The environment parameters, execution instructions, and final scores for each adjustment event are stored locally in encrypted form. When the number of accumulated regulation events reaches the batch processing threshold, the volatile organic compound component concentration feature vector of each event is extracted, random noise perturbation is added, and then uploaded to the cloud aggregation server. The server generates population adaptation parameters through feature vector clustering analysis and distributes them to each terminal device for model fine-tuning.

8. The intelligent testing and evaluation method for automotive parts as described in claim 1, characterized in that, Processing the driver's biological neural signals includes: The fingertip blood oxygen fluctuation characteristics are decomposed into multiple scales to extract specific frequency band components related to the activity of the autonomic nervous system and eliminate motion artifacts caused by steering wheel grip force. A dual-channel evaluation model for dynamic pupil changes was established to process the characteristics of these dynamic pupil changes. The dual-channel evaluation model includes a brightness adaptation channel and an emotion response channel. The brightness adaptation channel compensates for changes in ambient light intensity through iris texture features, while the emotion response channel identifies aversion response features through the mapping relationship between pupil oscillation frequency and a known emotion database.

9. An intelligent testing and evaluation device for automotive parts, characterized in that, The device includes: The organic component concentration monitoring module is used to collect real-time data on the concentration of volatile organic components in the cabin through an electronic nose sensor array arranged in the seat headrest, instrument panel and ceiling; The driver biosignal acquisition module is used to synchronously acquire the driver's bioneural signals, including the fingertip blood oxygen fluctuation characteristics monitored by the steering wheel photoplethysmography sensor, the pupil dynamic change characteristics captured by the infrared camera, and the sitting posture adjustment characteristics detected by the seat pressure distribution matrix. The dynamic game evaluation module is used to input the volatile organic compound concentration data and the biological neural signals into the dynamic game evaluation model and output adjustment instructions. The hidden purification module is used to trigger the deployment of the hidden purification device according to the adjustment command when the concentration of a specific component in the volatile organic compound concentration data exceeds the dynamic threshold and the brain electrical inhibition characteristics in the biological neural signal reach a preset intensity. The collaborative working module is used to activate the tactile feedback device and the odor release device to work together when the correlation between the concentration data of the volatile organic compounds and the dynamic change characteristics of the pupil is lower than a critical value.

10. An electronic device, characterized in that, include: A processor coupled to a memory for storing a program, which, when executed by the processor, causes the apparatus to perform the steps of the method as claimed in any one of claims 1 to 8.

Images