Warm low potential intelligent cockpit system, control method and storage medium
By integrating a low-potential thermotherapy module and a physiological parameter monitoring module into the intelligent cockpit system, and combining them with reinforcement learning algorithms, intelligent control is achieved, solving the problem that existing cockpit systems cannot adjust in real time, thus improving the driving experience and safety.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- GUANGZHOU OKEWE ELECTRONICS CO LTD
- Filing Date
- 2025-03-18
- Publication Date
- 2026-06-09
AI Technical Summary
Existing intelligent cockpit systems lack low-potential functionality, making it difficult to make real-time adjustments to meet individual needs and failing to meet the therapeutic standards for relieving fatigue.
It employs a low-potential thermotherapy module, a physiological parameter monitoring module, an on-board central control system, and an intelligent control system, combined with multimodal biosensors and reinforcement learning algorithms, to monitor and optimize thermotherapy low-potential parameters in real time and achieve automatic adjustment.
It provides a comprehensive therapeutic environment to enhance the driving experience, ensure driving safety and health, and meet personalized needs through real-time feedback and dynamic optimization.
Smart Images
Figure CN120267254B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent cockpit and health monitoring technology, and more specifically, to a warm, low-potential intelligent cockpit system, control method, and storage medium. Background Technology
[0002] With the continuous development of communication and internet technologies, a high-quality driving environment has become an important part of modern life. Against the backdrop of the rapid development of intelligent driving technology in new energy vehicles, the importance of intelligent driving is becoming increasingly prominent. Its intuitive, convenient, and easily accepted characteristics are gradually making it a mainstream driving trend. In recent years, many new energy vehicle manufacturers have launched a series of convenient driving functions to significantly improve the driving experience. Taking the intelligent cockpit as an example, this innovative function has not only powerfully promoted the widespread adoption of new energy vehicles but also brought consumers a superior driving experience, further expanding driving comfort and convenience. However, with the continuous development of the new energy vehicle industry, the shortcomings of intelligent cockpits are gradually becoming apparent. In terms of health and wellness, most cars are only equipped with basic seat heating functions, requiring manual temperature setting. It is difficult to adjust the temperature in real time during driving to meet individual needs, and the heating effect is insufficient to meet the therapeutic standards for relieving fatigue.
[0003] Existing technology discloses a far-infrared heating therapy car seat and its usage method, including a seat body, a fixing structure on the seat body, and a buttock therapy structure and a back therapy structure on the fixing structure; the fixing structure includes: two telescopic sleeves, two telescopic columns, two positioning seats, a positioning groove, two telescopic straps, and two buckles. However, this technology lacks low-potential functionality, making it difficult to meet therapy standards and failing to provide a more comfortable experience. Summary of the Invention
[0004] The purpose of this invention is to disclose a heated and low-potential intelligent cockpit system and its storage medium that combine heating and low-potential functions.
[0005] To achieve the above objectives, the present invention provides a warm, low-potential intelligent cockpit system, comprising:
[0006] The system includes a low-potential thermotherapy module, a physiological parameter monitoring module, an on-board central control system, an intelligent control system, and a data visualization module. The on-board central control system is connected to the low-potential thermotherapy module, the physiological parameter monitoring module, the intelligent control system, and the data visualization module.
[0007] The low-potential thermal therapy module is integrated into the smart cockpit using embedded technology, and its operating parameters are controlled by the vehicle's central control system.
[0008] The physiological parameter monitoring module consists of a multimodal biosensor array, which acquires multiple physiological indicators of the driver in real time and transmits them to the vehicle central control system.
[0009] The data visualization module displays the processed physiological data on the in-vehicle display screen in real time;
[0010] The intelligent control module acquires physiological data, dynamically optimizes treatment parameters, and feeds back the optimal parameters to the central control system.
[0011] Furthermore, the low-potential hyperthermia therapy module includes: a hyperthermia submodule and a low-potential submodule;
[0012] The heating submodule reads the voltage value of the internal temperature sensor of the low-potential seat cushion and transmits it to the CPU. The CPU obtains the temperature value T0 inside the seat cushion at this time and sends this temperature value T0 to the vehicle central control system via Bluetooth module. The vehicle central control system obtains the temperature value and performs comprehensive analysis and calculation through health monitoring data. It then sends the T1 value, which needs to be maintained at the seat cushion temperature, to the CPU control module of the low-potential controller via Bluetooth module. The CPU control module controls the output voltage at both ends of the heating wire in a PWM manner, thereby controlling the heating power of the heating wire and maintaining the current temperature T0 towards T1.
[0013] Furthermore, the low-potential submodule includes:
[0014] When the driver's fatigue index rises during health monitoring, the vehicle's central control system sends low potential intensity data to the low potential controller. The controller internally controls the negative potential frequency converter circuit to apply the low potential to the heating wire, thereby outputting an appropriate low potential intensity to alleviate the driver's fatigue. When the driver's mental state recovers, the vehicle's central control system will control the low potential output in real time, forming a closed loop.
[0015] Furthermore, the physiological parameter monitoring module includes:
[0016] Heart rate measurement: Heart rate is measured using a camera-based detection method; the driver's face is captured using a camera in the smart cockpit at 30fps with a resolution of 1980*1240 to capture subtle physiological changes in the face; the face detection algorithm from the dlib library in Python is used to detect faces in each frame, obtaining the bounding box of the face and cropping the face to obtain the driver's facial region; the G channel value of each pixel is extracted from the face, and the average RGB value of all pixels in that region is calculated.
[0017]
[0018] Gi(t) corresponds to the value of the G channel of the i-th pixel at time t, and N is the total number of pixels in the region.
[0019] A bandpass filter is applied to G(t); the bandpass filter is a Butterworth filter with a sampling frequency of 30 Hz, a low cutoff frequency of 0.5 Hz, and a high cutoff frequency of 3 Hz to remove high-frequency noise and remaining low-frequency interference; the filter order is set to 4 to balance the filtering effect and computational complexity.
[0020] G f (t) = filter(G(t))
[0021] The filtered signal is normalized to eliminate the influence of amplitude changes, and then a fast Fourier transform is performed to obtain the spectrum; the peak value fpeak, i.e. the spectrum corresponding to the heart rate, is found based on the spectrum.
[0022]
[0023] X(f) = FFT(G) n (t))
[0024] HR = 60 * f peak
[0025] Furthermore, the physiological parameter monitoring module also includes:
[0026] Heart rate variability: Heart rate variability is detected based on photoplethysmography (PPG). A PPG sensor measures the light absorption changes caused by subcutaneous blood flow through the emission of green light into the skin tissue, generating a pulse wave signal x(t). A bandpass filter is then used to remove low-frequency drift and high-frequency noise, and a moving average filter is used to smooth the signal.
[0027]
[0028] After preprocessing, pulse wave features are extracted. First, the peak-to-peak value of the pulse wave in the PPG signal is examined. The peak sequence is denoted as T = {t1, t2, t3, ..., tn}. The interval between adjacent peaks is calculated to generate the PPI sequence. The SDNN (standard deviation during RR) is calculated, where RR is the mean of the intervals during the peak period.
[0029] PPI i =t i+1 -t i (i = 1, 2, ..., n-1)
[0030] RR i =PPI i (i = 1, 2, ..., n-1)
[0031]
[0032] Furthermore, the physiological parameter monitoring module also includes:
[0033] Blood oxygen saturation: The skin is irradiated with red and infrared light, and the collected PPG signals include red light PPG signals and infrared light PPG signals. The two are separated according to different wavelengths. The DC and AC components of the two are decomposed using a low-pass filter. The characteristic value R of blood oxygen is then obtained by linear regression, and blood oxygen saturation is calculated by formula.
[0034] Regarding red light:
[0035] I red,DC =LPF(I red (t))
[0036] I red,AC =I red (t)-I red,DC
[0037] For infrared light:
[0038] I IR,DC =LPF(I IR (t))
[0039] I IR,AC =I IR (t)-I IR,DC
[0040] Calculate the ratio R of the DC component to the AC component of the red and infrared PPG signals:
[0041]
[0042] The R value can be converted into blood oxygen saturation using an empirical formula:
[0043] SpO2 = 110-25R.
[0044] Furthermore, the intelligent control module includes:
[0045] Prediction objective: Based on the human body index parameters at time t, predict the two control parameters for thermal low potential at time t+1, namely voltage and temperature, in order to achieve intelligent regulation of the parameters; and prevent abnormal physical conditions of the driver due to improper temperature or voltage settings.
[0046] Data collection: human body indicators, voltage and temperature values at time t; model prediction of voltage and temperature at time t+1; human body indicators at time t+1 after control parameter changes.
[0047] The setup for reinforcement learning training:
[0048] Environment: An external system that interacts with the predictive model, providing status and reward feedback.
[0049] Status: Human body parameters at time t, such as heart rate ht, heart rate variability Vt, blood oxygen saturation SOt, etc.
[0050] Action: Predicted voltage Ut+1 and temperature value Tt+1.
[0051] Reward: Based on the predicted voltage and temperature, the set values are changed, and the human body indicators at time t+1 are obtained through the sensor. The difference from the normal and stable values is used as the reward function; if the human body indicators after adjusting the voltage and temperature deviate from the normal values, it is penalized, and vice versa.
[0052] Model Architecture
[0053] Encoder: An LSTM-based encoder that can capture long-term dependencies in time-series data. It stacks three LSTM layers, with the output being the hidden states of the LSTMs, encoding human body indicator parameters into a 256-dimensional feature vector.
[0054] Predictor: Based on three linear fully connected layers, with output dimensions of 64, 16 and 2 respectively; the fully connected layers learn the mapping relationship between human feature vectors and setting parameters, mapping the encoder's 256-dimensional feature vectors into voltage and temperature values;
[0055] The encoder and predictor are initialized using the deep learning PyTorch framework. An adaptive momentum optimizer is chosen to update the model parameters. Simultaneously, a learning rate preheating algorithm is used to update the learning rate. This involves gradually increasing the learning rate in the early stages of training to accelerate model convergence, then gradually decreasing it after a certain number of iterations to help the model converge to the optimal solution. The initial learning rate is set to 0.001, and the weight decay is set to 0.9 to prevent overfitting. The learning rate preheating algorithm uses a cosine increment strategy, increasing the learning rate from 0 to the initial learning rate using a cosine function.
[0056] Training is performed based on the reinforce algorithm.
[0057] Furthermore, training based on the reinforce algorithm includes:
[0058] The reinforce algorithm is a gradient-based reinforcement learning algorithm that directly updates parameters to maximize the cumulative expected reward. It is suitable for parameter prediction tasks with continuous values.
[0059] The core idea of the reinforce algorithm is to calculate the policy gradient by sampling trajectories and update the policy network parameters to maximize the expected cumulative reward.
[0060] In a driving environment, sampling is performed every 5 minutes as a sampling period. The voltage U and temperature T values for the next second are predicted every second. After obtaining the set parameters, the human body indicators are denoted as Xi = {Ht, Vt, SOt}. The normal indicator x... normal ={h=80, V=150, SO=0.96}.
[0061] Reward for a single prediction:
[0062]
[0063] Collect data from the environment over a period of time and calculate the expected cumulative return:
[0064] G t =τ t +γτ t+1 +γ 2 τ t+2 +…+γ T-1 τ T
[0065] Where t is the time step, τ t The immediate reward is given by γ, which is the discount factor (γ = 0.95), and T is the termination time step of the sampling trajectory. Let the parameters of the policy network be π(α). t \s t ), where a is the predicted voltage and temperature values, s is the human body's state, and θ is the neural network parameters. The policy gradient update formula is:
[0066]
[0067] Where logπ(α) t \s t G represents the probability of the LSTM model's output value. t The cumulative reward weight parameter is α = 0.001, which is the learning rate.
[0068] Furthermore, this invention provides a control method for a thermally heated, low-potential intelligent cockpit system, applied to the aforementioned thermally heated, low-potential intelligent cockpit system, comprising:
[0069] The low-potential thermal therapy module is integrated into the smart cockpit using embedded technology, and the operating parameters of the low-potential thermal therapy module are adjusted through the vehicle central control system.
[0070] The physiological parameter monitoring module acquires multiple physiological indicators of the driver in real time and transmits them to the vehicle's central control system.
[0071] The processed physiological data is displayed in real time on the vehicle's screen through the data visualization module;
[0072] Physiological data is acquired through the intelligent control module, treatment parameters are dynamically optimized, and the optimal parameters are fed back to the central control system.
[0073] Furthermore, the present invention also provides a thermally heated, low-potential intelligent cockpit storage medium, characterized in that it stores a computer program thereon, and when the computer program is executed by a processor, it implements the thermally heated, low-potential intelligent cockpit system as described above.
[0074] Compared with the prior art, the beneficial effects of the technical solution of the present invention are:
[0075] This invention integrates thermal and low-potential therapeutic functions into a smart cockpit via a low-potential thermal therapy module, creating a comprehensive therapeutic environment that allows users to enjoy professional services while driving, achieving a fusion of driving and wellness, and enhancing their health experience. A physiological parameter monitoring module and a data visualization module introduce real-time feedback functionality. Utilizing high-precision sensors and rapid transmission technology, the driver's heart rate, blood pressure, and other physical indicators are displayed in real-time on the central control screen, allowing the driver to easily monitor their physical condition and accurately determine whether they meet safe driving requirements, thus ensuring driving safety. An intelligent control module automatically and precisely adjusts the thermal and low-potential parameters, achieving real-time dynamic optimization to ensure the therapeutic effect. Attached Figure Description
[0076] Figure 1 This is a diagram of the intelligent cockpit system with warm and low electrical potential described in Example 1;
[0077] Figure 2 This is a flowchart of the intelligent cockpit method for generating heat and low potential as described in Example 3; Detailed Implementation
[0078] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the scope of this patent.
[0079] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.
[0080] Example 1:
[0081] This embodiment provides, as follows: Figure 1 The illustrated warm, low-potential smart cockpit includes:
[0082] The system includes a low-potential thermotherapy module, a physiological parameter monitoring module, an on-board central control system, an intelligent control system, and a data visualization module. The on-board central control system is connected to the low-potential thermotherapy module, the physiological parameter monitoring module, the intelligent control system, and the data visualization module.
[0083] The low-potential thermal therapy module is integrated into the smart cockpit using embedded technology, and its operating parameters are controlled by the vehicle's central control system.
[0084] The physiological parameter monitoring module consists of a multimodal biosensor array, which acquires multiple physiological indicators of the driver in real time and transmits them to the vehicle central control system.
[0085] The data visualization module displays the processed physiological data on the in-vehicle display screen in real time;
[0086] The intelligent control module acquires physiological data, dynamically optimizes treatment parameters, and feeds back the optimal parameters to the central control system.
[0087] This embodiment integrates thermal and low-potential therapeutic functions into the smart cockpit through a low-potential thermal therapy module, creating a comprehensive therapeutic environment that allows users to enjoy professional services while driving, achieving a fusion of driving and wellness, and enhancing their health experience. A physiological parameter monitoring module and a data visualization module introduce real-time feedback functionality. Utilizing high-precision sensors and rapid transmission technology, the driver's heart rate, blood pressure, and other physical indicators are displayed in real-time on the central control screen, allowing them to easily monitor their physical condition and accurately determine whether they meet safe driving requirements, thus ensuring driving safety. An intelligent control module automatically and precisely adjusts the thermal and low-potential parameters, achieving real-time dynamic optimization to ensure the therapeutic effect.
[0088] Example 2:
[0089] This embodiment further discloses information based on Embodiment 1:
[0090] Furthermore, the low-potential hyperthermia therapy module includes: a hyperthermia submodule and a low-potential submodule;
[0091] The heating submodule reads the voltage value of the internal temperature sensor of the low-potential seat cushion and transmits it to the CPU. The CPU obtains the temperature value T0 inside the seat cushion at this time and sends this temperature value T0 to the vehicle central control system via Bluetooth module. The vehicle central control system obtains the temperature value and performs comprehensive analysis and calculation through health monitoring data. It then sends the T1 value, which needs to be maintained at the seat cushion temperature, to the CPU control module of the low-potential controller via Bluetooth module. The CPU control module controls the output voltage at both ends of the heating wire in a PWM manner, thereby controlling the heating power of the heating wire and maintaining the current temperature T0 towards T1.
[0092] Furthermore, the low-potential submodule includes:
[0093] When the driver's fatigue index rises during health monitoring, the vehicle's central control system sends low potential intensity data to the low potential controller. The controller internally controls the negative potential frequency converter circuit to apply the low potential to the heating wire, thereby outputting an appropriate low potential intensity to alleviate the driver's fatigue. When the driver's mental state recovers, the vehicle's central control system will control the low potential output in real time, forming a closed loop.
[0094] Furthermore, the physiological parameter monitoring module includes:
[0095] Heart rate measurement: Heart rate is measured using a camera-based detection method; the driver's face is captured using a camera in the smart cockpit at 30fps with a resolution of 1980*1240 to capture subtle physiological changes in the face; the face detection algorithm from the dlib library in Python is used to detect faces in each frame, obtaining the bounding box of the face and cropping the face to obtain the driver's facial region; the G channel value of each pixel is extracted from the face, and the average RGB value of all pixels in that region is calculated.
[0096]
[0097] Gi(t) corresponds to the value of the G channel of the i-th pixel at time t, and N is the total number of pixels in the region.
[0098] A bandpass filter is applied to G(t); the bandpass filter is a Butterworth filter with a sampling frequency of 30 Hz, a low cutoff frequency of 0.5 Hz, and a high cutoff frequency of 3 Hz to remove high-frequency noise and remaining low-frequency interference; the filter order is set to 4 to balance the filtering effect and computational complexity.
[0099] G f (t) = filter(G(t))
[0100] The filtered signal is normalized to eliminate the influence of amplitude changes, and then a fast Fourier transform is performed to obtain the spectrum; the peak value fpeak, i.e. the spectrum corresponding to the heart rate, is found based on the spectrum.
[0101]
[0102] X(f) = FFT(G) n (t))
[0103] HR = 60 * f peak
[0104] Furthermore, the physiological parameter monitoring module also includes:
[0105] Heart rate variability: Heart rate variability is detected based on photoplethysmography (PPG). A PPG sensor measures the light absorption changes caused by subcutaneous blood flow through the emission of green light into the skin tissue, generating a pulse wave signal x(t). A bandpass filter is then used to remove low-frequency drift and high-frequency noise, and a moving average filter is used to smooth the signal.
[0106]
[0107] After preprocessing, pulse wave features are extracted. First, the peak-to-peak value of the pulse wave in the PPG signal is examined. The peak sequence is denoted as T = {t1, t2, t3, ..., tn}. The interval between adjacent peaks is calculated to generate the PPI sequence. The SDNN (standard deviation during RR) is calculated, where RR is the mean of the intervals during the peak period.
[0108] PPI i =t i+1 -t i (i = 1, 2, ..., n-1)
[0109] RR i =PPI i (i = 1, 2, ..., n-1)
[0110]
[0111] Furthermore, the physiological parameter monitoring module also includes:
[0112] Blood oxygen saturation: The skin is irradiated with red and infrared light, and the collected PPG signals include red light PPG signals and infrared light PPG signals. The two are separated according to different wavelengths. The DC and AC components of the two are decomposed using a low-pass filter. The characteristic value R of blood oxygen is then obtained by linear regression, and blood oxygen saturation is calculated by formula.
[0113] Regarding red light:
[0114] I red,DC =LPF(I red (t))
[0115] I red,AC =I red (t)-I red,DC
[0116] For infrared light:
[0117] I IR,DC =LPF(I IR (t))
[0118] I IR,AC =I IR (t)-I IR,DC
[0119] Calculate the ratio R of the DC component to the AC component of the red and infrared PPG signals:
[0120]
[0121]
[0122] The R value can be converted into blood oxygen saturation using an empirical formula:
[0123] SpO2 = 110-25R.
[0124] Furthermore, the intelligent control module includes:
[0125] Prediction objective: Based on the human body index parameters at time t, predict the two control parameters for thermal low potential at time t+1, namely voltage and temperature, in order to achieve intelligent regulation of the parameters; and prevent abnormal physical conditions of the driver due to improper temperature or voltage settings.
[0126] Data collection: human body indicators, voltage and temperature values at time t; model prediction of voltage and temperature at time t+1; human body indicators at time t+1 after control parameter changes.
[0127] The setup for reinforcement learning training:
[0128] Environment: An external system that interacts with the predictive model, providing status and reward feedback.
[0129] Status: Human body parameters at time t, such as heart rate ht, heart rate variability Vt, blood oxygen saturation SOt, etc.
[0130] Action: Predicted voltage Ut+1 and temperature value Tt+1.
[0131] Reward: Based on the predicted voltage and temperature, the set values are changed, and the human body indicators at time t+1 are obtained through the sensor. The difference from the normal and stable values is used as the reward function; if the human body indicators after adjusting the voltage and temperature deviate from the normal values, it is penalized, and vice versa.
[0132] Model Architecture
[0133] Encoder: An LSTM-based encoder that can capture long-term dependencies in time-series data. It stacks three LSTM layers, with the output being the hidden states of the LSTMs, encoding human body indicator parameters into a 256-dimensional feature vector.
[0134] Predictor: Based on three linear fully connected layers, with output dimensions of 64, 16 and 2 respectively; the fully connected layers learn the mapping relationship between human feature vectors and setting parameters, mapping the encoder's 256-dimensional feature vectors into voltage and temperature values;
[0135] The encoder and predictor are initialized using the deep learning PyTorch framework. An adaptive momentum optimizer is chosen to update the model parameters. Simultaneously, a learning rate preheating algorithm is used to update the learning rate. This involves gradually increasing the learning rate in the early stages of training to accelerate model convergence, then gradually decreasing it after a certain number of iterations to help the model converge to the optimal solution. The initial learning rate is set to 0.001, and the weight decay is set to 0.9 to prevent overfitting. The learning rate preheating algorithm uses a cosine increment strategy, increasing the learning rate from 0 to the initial learning rate using a cosine function.
[0136] Training is performed based on the reinforce algorithm.
[0137] Furthermore, training based on the reinforce algorithm includes:
[0138] The reinforce algorithm is a gradient-based reinforcement learning algorithm that directly updates parameters to maximize the cumulative expected reward. It is suitable for parameter prediction tasks with continuous values.
[0139] The core idea of the reinforce algorithm is to calculate the policy gradient by sampling trajectories and update the policy network parameters to maximize the expected cumulative reward.
[0140] In a driving environment, sampling is performed every 5 minutes as a sampling period. The voltage U and temperature T values for the next second are predicted every second. After obtaining the set parameters, the human body indicators are denoted as Xi = {Ht, Vt, SOt}. The normal indicator x... normal ={h=80, V=150, SO=0.96}.
[0141] Reward for a single prediction:
[0142]
[0143] Collect data from the environment over a period of time and calculate the expected cumulative return:
[0144] G t =τ t +γτ t+1 +γ 2 τ t+2 +…+γ T-1 τ T
[0145] Where t is the time step, τ t The immediate reward is given by γ, which is the discount factor (γ = 0.95), and T is the termination time step of the sampling trajectory. Let the parameters of the policy network be π(α). t \s t), where a is the predicted voltage and temperature values, s is the human body's state, and θ is the neural network parameters. The policy gradient update formula is:
[0146]
[0147] Where logπ(α) t \s t G represents the probability of the LSTM model's output value. t The cumulative reward weight parameter is α = 0.001, which is the learning rate.
[0148] This embodiment integrates thermal and low-potential therapeutic functions into the smart cockpit through a low-potential thermal therapy module, creating a comprehensive therapeutic environment that allows users to enjoy professional services while driving, achieving a fusion of driving and wellness, and enhancing their health experience. A physiological parameter monitoring module and a data visualization module introduce real-time feedback functionality. Utilizing high-precision sensors and rapid transmission technology, the driver's heart rate, blood pressure, and other physical indicators are displayed in real-time on the central control screen, allowing them to easily monitor their physical condition and accurately determine whether they meet safe driving requirements, thus ensuring driving safety. An intelligent control module automatically and precisely adjusts the thermal and low-potential parameters, achieving real-time dynamic optimization to ensure the therapeutic effect.
[0149] Example 3:
[0150] This embodiment provides, as follows: Figure 2 The method for controlling a thermally low-potential smart cockpit system, as shown, is applied to the aforementioned thermally low-potential smart cockpit system and includes:
[0151] The low-potential thermal therapy module is integrated into the smart cockpit using embedded technology, and the operating parameters of the low-potential thermal therapy module are adjusted through the vehicle central control system.
[0152] The physiological parameter monitoring module acquires multiple physiological indicators of the driver in real time and transmits them to the vehicle's central control system.
[0153] The processed physiological data is displayed in real time on the vehicle's screen through the data visualization module;
[0154] Physiological data is acquired through the intelligent control module, treatment parameters are dynamically optimized, and the optimal parameters are fed back to the central control system.
[0155] This embodiment integrates thermal and low-potential therapeutic functions into the smart cockpit through a low-potential thermal therapy module, creating a comprehensive therapeutic environment that allows users to enjoy professional services while driving, achieving a fusion of driving and wellness, and enhancing their health experience. A physiological parameter monitoring module and a data visualization module introduce real-time feedback functionality. Utilizing high-precision sensors and rapid transmission technology, the driver's heart rate, blood pressure, and other physical indicators are displayed in real-time on the central control screen, allowing them to easily monitor their physical condition and accurately determine whether they meet safe driving requirements, thus ensuring driving safety. An intelligent control module automatically and precisely adjusts the thermal and low-potential parameters, achieving real-time dynamic optimization to ensure the therapeutic effect.
[0156] Example 4:
[0157] Furthermore, the embodiment also provides a thermally low-potential smart cockpit storage medium, characterized in that it stores a computer program, which, when executed by a processor, implements the thermally low-potential smart cockpit system as described above.
[0158] This embodiment integrates thermal and low-potential therapeutic functions into the smart cockpit through a low-potential thermal therapy module, creating a comprehensive therapeutic environment that allows users to enjoy professional services while driving, achieving a fusion of driving and wellness, and enhancing their health experience. A physiological parameter monitoring module and a data visualization module introduce real-time feedback functionality. Utilizing high-precision sensors and rapid transmission technology, the driver's heart rate, blood pressure, and other physical indicators are displayed in real-time on the central control screen, allowing them to easily monitor their physical condition and accurately determine whether they meet safe driving requirements, thus ensuring driving safety. An intelligent control module automatically and precisely adjusts the thermal and low-potential parameters, achieving real-time dynamic optimization to ensure the therapeutic effect.
[0159] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art can make other variations or modifications based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.
Claims
1. A smart cockpit system with low electrical potential and high temperature, characterized in that, include: Low-potential thermal therapy module, physiological parameter monitoring module, vehicle-mounted central control system and intelligent control module, and data visualization module; The vehicle-mounted central control system is connected to the low-potential thermotherapy module, the physiological parameter monitoring module, the intelligent control system, and the data visualization module, respectively. The low-potential thermal therapy module is integrated into the smart cockpit using embedded technology, and its operating parameters are controlled by the vehicle's central control system. The physiological parameter monitoring module consists of a multimodal biosensor array, which acquires multiple physiological indicators of the driver in real time and transmits them to the vehicle central control system. The data visualization module displays the processed physiological data on the in-vehicle display screen in real time; The intelligent control module acquires physiological data, dynamically optimizes treatment parameters, and feeds back the optimal parameters to the central control system; The intelligent control module includes: Prediction objective: Based on the human body index parameters at time t, predict the two control parameters for thermal low potential at time t+1, namely voltage and temperature, in order to achieve intelligent regulation of the parameters; and prevent abnormal physical conditions of the driver due to improper temperature or voltage settings. Data collection: human body indicators, voltage and temperature values at time t; model prediction of voltage and temperature at time t+1; human body indicators at time t+1 after control parameter changes. The setup for reinforcement learning training: Environment: External systems that interact with the predictive model, providing status and reward feedback; Status: Human body parameters at time t, including heart rate. Heart rate variability blood oxygen saturation ; Action: Predicted voltage and temperature value ; Reward: Based on the predicted voltage and temperature, the set values are changed, and the human body indicators at time t+1 are obtained through the sensor. The difference from the normal and stable values is used as the reward function; if the human body indicators after adjusting the voltage and temperature deviate from the normal values, it is penalized, and vice versa. Model Architecture Encoder: An LSTM-based encoder that can capture long-term dependencies in time series data; three LSTM layers are stacked, and the output is the hidden state of the LSTM, encoding human body indicator parameters into a 256-dimensional feature vector. Predictor: Based on three linear fully connected layers, with output dimensions of 64, 16 and 2 respectively; the fully connected layers learn the mapping relationship between human feature vectors and setting parameters, mapping the encoder's 256-dimensional feature vectors into voltage and temperature values; The encoder and predictor are initialized according to the deep learning PyTorch framework. An adaptive momentum optimizer is selected to update the model parameters. At the same time, a learning rate preheating algorithm is used to update the learning rate. That is, in the early stage of training, the learning rate is gradually increased to speed up the convergence of the model. After a certain number of iterations, the learning rate is gradually reduced to help the model converge to the optimal solution. The initial learning rate is set to 0.001 and the weight decay is set to 0.9 to prevent overfitting. The learning rate preheating algorithm adopts a cosine increase strategy, increasing the learning rate from 0 to the initial learning rate according to the cosine function. Training is performed based on the reinforce algorithm; Training based on the reinforce algorithm includes: The reinforce algorithm is a gradient-based reinforcement learning algorithm that directly updates parameters to maximize the cumulative expected reward. It is suitable for parameter prediction tasks with continuous values. The core idea of the reinforce algorithm is to calculate the policy gradient by sampling the trajectory and update the policy network parameters to maximize the expected cumulative reward. In a driving environment, sampling is performed every 5 minutes as a sampling period. The voltage U and temperature T values for the next second are predicted every second. The resulting human body indicators after setting parameters are denoted as Xi = { , }, normal indicators ; Reward for a single prediction: Collect data from the environment over a period of time and calculate the expected cumulative return: Where t is the time step, It is an instant reward, and γ is a discount factor ( = 0.95), T is the termination time step of the sampling trajectory; let the parameters of the policy network be ,in, Let s represent the predicted voltage and temperature values, s represent the human body's state indicators, and θ represent the neural network parameters; the policy gradient update formula is: in The probability representing the output value of the LSTM model. For cumulative return weighting parameters, = 0.001 is the learning rate.
2. The intelligent cockpit system with low electrical potential and high temperature as described in claim 1, characterized in that, The low-potential hyperthermia module includes a hyperthermia submodule and a low-potential submodule. The heating submodule reads the voltage value of the internal temperature sensor of the low-potential seat cushion and transmits it to the CPU. The CPU obtains the temperature value T0 inside the seat cushion at this time and sends this temperature value T0 to the vehicle central control system via Bluetooth module. The vehicle central control system obtains the temperature value and performs comprehensive analysis and calculation through health monitoring data. It then sends the T1 value, which needs to be maintained at the seat cushion temperature, to the CPU control module of the low-potential controller via Bluetooth module. The CPU control module controls the output voltage at both ends of the heating wire in a PWM manner, thereby controlling the heating power of the heating wire and maintaining the current temperature T0 towards T1.
3. The intelligent cockpit system with low electrical potential and high temperature according to claim 2, characterized in that, The low-potential submodule includes: When the driver's fatigue index rises during health monitoring, the vehicle's central control system sends low potential intensity data to the low potential controller. The controller internally controls the negative potential frequency converter circuit to apply the low potential to the heating wire, thereby outputting an appropriate low potential intensity to alleviate the driver's fatigue. When the driver's mental state recovers, the vehicle's central control system will control the low potential output in real time, forming a closed loop.
4. The intelligent cockpit system with low electrical potential and high temperature according to claim 1, characterized in that, The physiological parameter monitoring module includes: Heart rate measurement: Heart rate is measured using a camera-based detection method; the driver's face is captured using a camera in the smart cockpit at 30fps with a resolution of 1980*1240 to capture subtle physiological changes in the face; the face detection algorithm from the dlib library in Python is used to detect faces in each frame, obtaining the bounding box of the face and cropping the face to obtain the driver's facial region; the G channel value of each pixel is extracted from the face, and the average RGB value of all pixels in that region is calculated. The value of the G channel of the i-th pixel at time t, where N is the total number of pixels in the region; A bandpass filter is applied to G(t); the bandpass filter is a Butterworth filter with a sampling frequency of 30 Hz, a low cutoff frequency of 0.5 Hz, and a high cutoff frequency of 3 Hz to remove high-frequency noise and remaining low-frequency interference; the filter order is set to 4 to balance the filtering effect and computational complexity. The filtered signal is normalized to eliminate the influence of amplitude variations, and then a Fast Fourier Transform is performed to obtain the spectrum; peak values are then identified based on the spectrum. That is, the spectrum corresponding to heart rate; 。 5. The intelligent cockpit system with low electrical potential and high temperature according to claim 1, characterized in that, The physiological parameter monitoring module also includes: Heart rate variability: Heart rate variability is detected based on photoplethysmography (PPG). A PPG sensor measures the light absorption changes caused by subcutaneous blood flow through the emission of green light into the skin tissue, generating a pulse wave signal x(t). A bandpass filter is then used to remove low-frequency drift and high-frequency noise, and a moving average filter is used to smooth the signal. After preprocessing, pulse wave features are extracted. First, the peak-to-peak value of the pulse wave in the PPG signal is examined, and the peak sequence is denoted as T = { , , ,…, } Calculate the interval between adjacent peaks, generate the PPI sequence, and calculate the SDNN (standard deviation during RR). The interval between peaks; The mean of the intervals between peaks 。 6. The intelligent cockpit system with low electrical potential and high temperature according to claim 1, characterized in that, The physiological parameter monitoring module also includes: Blood oxygen saturation: The skin is irradiated with red and infrared light, and the collected PPG signals include red light PPG signals and infrared light PPG signals. The two are separated according to different wavelengths. The DC and AC components of the two are decomposed using a low-pass filter. The characteristic value R of blood oxygen is then obtained by linear regression, and blood oxygen saturation is calculated by formula. Regarding red light: For infrared light: Calculate the ratio R of the DC component to the AC component of the red and infrared PPG signals: The R value can be converted into blood oxygen saturation using an empirical formula: 。 7. A control method for a thermally heated, low-potential intelligent cockpit system, applied to the thermally heated, low-potential intelligent cockpit system as described in claim 1, characterized in that, include: The low-potential thermal therapy module is integrated into the smart cockpit using embedded technology, and the operating parameters of the low-potential thermal therapy module are adjusted through the vehicle central control system. The physiological parameter monitoring module acquires multiple physiological indicators of the driver in real time and transmits them to the vehicle's central control system. The processed physiological data is displayed in real time on the vehicle's screen through the data visualization module; Physiological data is acquired through the intelligent control module, treatment parameters are dynamically optimized, and the optimal parameters are fed back to the central control system.
8. A warm, low-potential intelligent cockpit storage medium, characterized in that, It stores a computer program, which, when executed by a processor, implements the intelligent cockpit system with low thermal potential as described in any one of claims 1 to 7.