Intelligent cushion

Abstract

The application provides a kind of intelligent cushion, it is related to cushion technical field, including cushion body and control module, cushion body is sequentially provided with massage layer, sensing layer, cold and warm layer and base layer from top to bottom, massage layer includes evenly arranged massage air cushion base and the elastic fusion type massage head corresponding with massage air cushion base, sensing layer is used to obtain sensor data, cold and warm layer includes heating sheet and the condensing sheet being arranged below heating sheet, base layer is used to be installed on external equipment;Control module is used to pre-process sensor data based on dynamic noise compensation algorithm, and biomechanical characteristics are extracted, and according to biomechanical characteristics, the operation mode of massage air cushion base, elastic fusion type massage head, heating sheet and condensing sheet is controlled respectively.The application can improve the health problem of existing cushion.

Description

Technical Field

[0001] This invention relates to the field of seat cushion technology, and more specifically, to a smart seat cushion. Background Technology

[0002] Most mainstream seat cushions on the market are made of memory foam, gel, ordinary sponge, or fixed ergonomic shaping. A few have basic heating or targeted vibration massage, and they are widely used in office, home, and study settings.

[0003] In related technologies, existing seat cushions rely on a fixed shape and a single material to achieve pressure cushioning. Their functions are fixed and cannot be adjusted, and they can only provide static physical support. Some seat cushions add manual temperature control and fixed massage levels. They rely on the outer fabric to simply improve the breathability and rely on grooves and hollow structures to slightly reduce pressure on the tailbone. However, after sitting for a long time, there are still problems such as blood stasis in the lower limbs, excessive pressure on the lumbar spine and ischial tuberosities, and local damp heat, which affect physical health. Summary of the Invention

[0004] The present invention aims to solve at least one of the above-mentioned problems.

[0005] To address the aforementioned problems, this invention provides a smart seat cushion, comprising a cushion body and a control module. The cushion body is provided with a massage layer, a sensing layer, a heating / cooling layer, and a base layer from top to bottom. The massage layer includes uniformly arranged massage air cushion bases and elastic fused massage heads corresponding to each massage air cushion base. The sensing layer is used to acquire sensor data, including temperature and humidity sensor data, pressure sensor data, IMU sensor data, and piezoelectric film sensor data. The heating / cooling layer includes a heating element and a condensation element disposed below the heating element. The base layer is used for mounting on an external device. The control module is used to preprocess the sensor data based on a dynamic noise compensation algorithm, extract biomechanical features, and control the operation mode of the massage air cushion base, the elastic fusion massage head, the heating element, and the condensation element according to the biomechanical features.

[0006] Optionally, the control module is specifically used for: Based on the dynamic noise compensation algorithm, the sensor data is subjected to moving average filtering and data validity verification. When the verification result is invalid, the surrounding sensors are activated to perform interpolation and completion, obtain corrected sensor data, and perform feature conversion to obtain the biomechanical characteristics. Based on the total pressure value, pressure distribution characteristics, dispersion, and corresponding preset thresholds in the biomechanical characteristics, the usage state of the seat cushion is determined, and the operation mode of the massage air cushion base, the elastic fusion massage head, the heating plate, and the condensation plate are controlled according to the usage state of the seat cushion.

[0007] Optionally, the control module is specifically used for: The biomechanical features are matched with stored templates to determine the user's identity; Based on the user's identity and personalized settings, the operation modes of the massage air cushion base, the elastic integrated massage head, the heating element, and the condensation element are controlled respectively.

[0008] Optionally, the control module is specifically used for: The usage time context features, response action data, implicit user feedback data, explicit user feedback data, and environmental features are obtained and anonymized to obtain multidimensional features; Based on the aforementioned multidimensional features, an active perturbation mechanism is used to determine high-quality training samples; Based on the high-quality training samples, a weighted moving average and bias integral algorithm is used to learn local rules at the edge. When the cloud learning conditions are met, a cloud deep learning model is used for incremental learning to obtain a personalized decision weight file. Based on the personalized decision weight file, the model is updated using an A / B testing gray-scale update strategy and an anomaly data cleaning strategy to obtain a new personalized adjustment model. Based on the biomechanical characteristics and the new personalized adjustment model, the operation modes of the massage air cushion base, the elastic fusion massage head, the heating element, and the condensation element are controlled respectively.

[0009] Optionally, the control module is specifically used for: Perform hardware self-test, load user configuration, and initialize timing status; Read the pressure sensor data and timestamp data and perform data preprocessing to obtain the pressure characteristics in the biomechanical features; Based on the pressure characteristics, the sitting state is determined according to the timing fault tolerance strategy and sitting posture classification conditions, and the total sitting time and the sitting time of the same type of sitting posture are determined. Based on the total sitting time and the sitting time of the same type of sitting posture, a sitting reminder scheme is determined, and based on the sitting reminder scheme, the operation mode of the elastic integrated massage head is controlled. Data is stored based on storage strategies, and reports are generated. Acquire abnormal status data, generate and store abnormal alert schemes.

[0010] Optionally, the heating element has horizontally arranged elongated vents, the outer shell of the condenser plate adopts a honeycomb structure, and the heating and cooling layer also includes a heating chip connected to the heating element. The outer shell of the heating chip is wrapped with aluminum foil, and a temperature sensor is provided on the side of the heating chip. The control module is specifically used for: The operation of the heating element is controlled according to the heating adjustment command and the heating chip, and the operating mode of the heating element is controlled through the temperature feedback of the temperature sensor. The operation of the condenser is controlled according to the cooling adjustment command.

[0011] Optionally, the control module is specifically used for: Perform hardware self-test, parameter loading, and calibration initialization; Based on the aforementioned biomechanical characteristics, the total weight under force, regional pressure distribution, and pressure center of gravity are determined. Using big data cloud technology, the sitting posture category is determined based on the total weight under force, the regional pressure distribution, and the pressure center of gravity. Based on preset user priority or the biomechanical characteristics, a sitting posture protection mode is determined, and the target air pressure value of the massage air cushion base is determined according to the sitting posture protection mode. The target air pressure value is converted into an air pump control command, and the massage air cushion base is adjusted using PID closed-loop feedback. The system obtains information on the fault conditions of the massage layer and the sensing layer, and controls the operating status of the massage air cushion base based on the fault conditions.

[0012] Optionally, pressure sensors are provided on both sides of the massage air cushion base, and multiple pressure sensors are symmetrically distributed. The massage air cushion base includes an air cushion, an air pump, a solenoid valve, and a buffer valve connected in sequence.

[0013] Optionally, the base layer includes sound and heat insulation cotton, a rigid base plate, and an anti-slip base arranged sequentially from top to bottom. The sound and heat insulation cotton is located below the heating and cooling layer, and the anti-slip base is used for installation on the external equipment.

[0014] Optionally, the seat cushion body further includes a surface contact layer disposed above the massage layer, the surface contact layer being made of skin-friendly and breathable fabric and equipped with a temperature and humidity sensor.

[0015] The beneficial effects of the smart seat cushion of the present invention are: The seat cushion is designed as an integrated unit with a control module, arranged sequentially from top to bottom as a massage layer, a sensing layer, a heating / cooling layer, and a base layer. Unlike existing technologies that rely on a fixed shape and a single material to provide only static, non-adjustable physical support, this layered structure integrates massage, multimodal sensing, heating / cooling control, and equipment installation support functions, ensuring that each layer's functions do not interfere with each other. The massage layer features evenly arranged massage air cushion bases and individually matched elastic massage heads. Compared to existing seat cushions that only offer fixed massage levels and cannot adapt to the real-time pressure applied by the human body, this provides a dynamically adjustable massage effect based on the air cushion bases and elastic massage heads. The sensing layer simultaneously collects data from four types of sensors: temperature and humidity, pressure, IMU, and piezoelectric film. This overcomes the shortcomings of existing seat cushions that lack multi-dimensional real-time human body and working condition sensing and rely solely on the outer fabric and perforated structure for passive, slight pressure reduction. It can comprehensively capture raw data related to human body pressure, physiological vibration, environmental temperature and humidity, and vehicle vibration. The heating / cooling layer uses heating elements combined with a condenser located below. The composite structure overcomes the shortcomings of existing seat cushions that only offer simple manual temperature control and have a single temperature control method. It also has the hardware foundation for bidirectional temperature control, including both heating and cooling. The base layer allows the entire seat cushion to be stably mounted on external devices such as automobiles, ensuring the stable operation of the entire hardware structure. The supporting control module uses a dynamic noise compensation algorithm to preprocess data from various sensors collected by the sensing layer. After filtering out data interference caused by driving conditions, it extracts biomechanical characteristics that reflect the real-time sitting posture and pressure state of the human body. Based on these biomechanical characteristics, it synchronously and separately controls the operating modes of the massage air cushion base, elastic integrated massage head, heating plate, and cooling plate. This overcomes the shortcomings of existing seat cushions where all functions are preset and fixed, and cannot actively adapt to the real-time physical condition of drivers and passengers. It can dynamically adjust the support and massage intensity and the output of heat and cold based on the real-time pressure data of the human body. It simultaneously alleviates health problems caused by prolonged sitting, such as blood stasis in the lower limbs, excessive pressure on the lumbar spine and ischial tuberosities, and the accumulation of damp heat in the contact area, from three dimensions: dynamic support, active massage, and bidirectional temperature control. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the structure of the smart cushion provided in an embodiment of the present invention. Detailed Implementation

[0017] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Although some embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the present invention. It should be understood that the accompanying drawings and embodiments of the present invention are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0018] It should be understood that the various steps described in the method embodiments of the present invention may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0019] The term "comprising" and its variations as used herein are open-ended, meaning "including but not limited to"; the term "based on" means "at least partially based on"; the term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments"; and the term "optionally" means "optional embodiments". Definitions of other terms will be given in the following description. It should be noted that the concepts of "first," "second," etc., mentioned in this invention are used only to distinguish different devices, modules, or units, and are not intended to limit the order of functions performed by these devices, modules, or units or their interdependencies.

[0020] It should be noted that the terms "a" and "a plurality of" used in this invention are illustrative rather than restrictive. Those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0021] The names of the messages or information exchanged between the multiple devices in the embodiments of the present invention are for illustrative purposes only and are not intended to limit the scope of these messages or information.

[0022] like Figure 1 As shown in the figure, an embodiment of the present invention provides a smart seat cushion, including a seat cushion body and a control module. The seat cushion body is provided with a massage layer, a sensing layer, a heating and cooling layer and a base layer from top to bottom. The massage layer includes a massage air cushion base evenly arranged and an elastic fused massage head corresponding to the massage air cushion base. The sensing layer is used to acquire sensor data, including temperature and humidity sensor data, pressure sensor data, IMU sensor data and piezoelectric film sensor data. The heating and cooling layer includes a heating element and a condensation element disposed below the heating element. The base layer is used to be installed on an external device.

[0023] Specifically, the seat cushion consists of a massage layer, a sensing layer, a heating / cooling layer, and a base layer, arranged from top to bottom. It can be fitted to external devices such as car seats. The massage layer comprises multiple evenly arranged massage air cushion bases, each with a corresponding elastic, integrated massage head. The massage head adjusts its support height according to changes in air pressure within the air cushion base and outputs vibration to provide a massage function. The sensing layer integrates various sensor units for real-time data acquisition, including temperature and humidity sensors, pressure sensors, IMU sensors, and piezoelectric film sensors. These sensors synchronously collect corresponding physical quantities and transmit them to the control module, providing a data foundation for intelligent adjustment. The heating / cooling layer includes heating and cooling plates. The cooling plate is positioned below the heating plate. The heating plate outputs heat to raise the seat cushion's temperature, while the cooling plate cools it down, working together to regulate the seat cushion's temperature. The base layer, located at the bottom of the seat cushion, provides rigid structural support and a mounting base for the upper functional layers, securing the entire seat cushion to the external device. This embodiment integrates massage, sensing, temperature control, and support functions through a layered modular structure design. This ensures that each functional module operates independently without interference, facilitates production assembly and subsequent maintenance, and provides a complete sensing data foundation for intelligent adjustment through the integrated arrangement of multiple types of sensors, thus realizing a multi-functional integrated seat cushion structure layout.

[0024] The control module is used to preprocess the sensor data based on a dynamic noise compensation algorithm, extract biomechanical features, and control the operation mode of the massage air cushion base, the elastic fusion massage head, the heating element, and the condensation element according to the biomechanical features.

[0025] Specifically, the control module first receives various raw sensor data transmitted from the sensing layer. Based on a dynamic noise compensation algorithm, it preprocesses the raw sensor data to eliminate interference noise during the acquisition process, improving data accuracy and stability. After data preprocessing, the control module extracts biomechanical features from the valid sensor data. These features reflect the pressure distribution, body posture, and related physiological state of the occupants, serving as the basis for subsequent adjustments. Subsequently, using the extracted biomechanical features as the core basis, the control module outputs control signals to the corresponding actuators of the massage layer and the heating / cooling layer: for the massage layer, it controls the air pressure of the massage air cushion base to adjust the support firmness of different areas of the seat cushion, while simultaneously controlling the vibration parameters of the elastic integrated massage head to achieve different massage modes; for the heating / cooling layer, it controls the heating power of the heating element and the cooling power of the cooling element to achieve overall or localized temperature regulation of the seat cushion. Through this process, the control module can adaptively adjust the massage, support, and temperature control states of the seat cushion based on real-time sensed human body status data. This embodiment can improve the reliability of sensor data through dynamic noise compensation algorithm, ensure the accuracy of perception results in complex driving environments, and based on the extracted biomechanical features, link and control multiple actuators such as massage, support and temperature control to realize adaptive intelligent adjustment of seat function, thereby improving the comfort and health protection effect of driving and riding.

[0026] In this embodiment, an integrated seat cushion body is provided, consisting of a massage layer, a sensing layer, a heating / cooling layer, and a base layer arranged sequentially from top to bottom, along with a control module. Unlike existing technologies that rely on a fixed shape and a single material to provide only static, non-adjustable physical support, this layered structure integrates massage, multimodal sensing, heating / cooling temperature control, and equipment installation support functions, ensuring that each layer's functions do not interfere with each other. The massage layer features evenly arranged massage air cushion bases and individually matched elastic massage heads. Compared to existing seat cushions that only offer fixed massage levels and cannot adapt to real-time human pressure, this layer provides a dynamically adjustable massage effect based on the air cushion bases and elastic massage heads. The sensing layer simultaneously collects data from four types of sensors: temperature and humidity, pressure, IMU, and piezoelectric film. This overcomes the shortcomings of existing seat cushions that lack multi-dimensional real-time human and working condition sensing and rely solely on the outer fabric and hollow structure for passive, slight pressure reduction. It can comprehensively capture raw data related to human pressure, physiological vibration, environmental temperature and humidity, and vehicle vibration. The heating / cooling layer uses heating elements located below it. The unique condenser plate combination structure overcomes the shortcomings of existing seat cushions that only offer simple manual temperature control and a single temperature control method. It also has the hardware foundation for bidirectional temperature control, including both heating and cooling. The base layer allows the entire seat cushion to be stably mounted on external devices such as automobiles, ensuring the stable operation of the entire hardware structure. The supporting control module uses a dynamic noise compensation algorithm to preprocess data from various sensors collected by the sensing layer. After filtering out data interference caused by driving conditions, it extracts biomechanical characteristics that reflect the real-time sitting posture and pressure state of the human body. Based on these biomechanical characteristics, it synchronously and separately controls the operating modes of the massage air cushion base, elastic integrated massage head, heating plate, and condenser plate. This overcomes the shortcomings of existing seat cushions where all functions are preset and fixed, and cannot actively adapt to the real-time physical condition of drivers and passengers. It can dynamically adjust the support massage intensity and hot and cold output based on the real-time pressure data of the human body, and simultaneously alleviate the health problems caused by prolonged sitting, such as blood stasis in the lower limbs, excessive pressure on the lumbar spine and ischial tuberosities, and accumulation of damp heat in the contact area, from three dimensions: dynamic support, active massage, and bidirectional temperature control.

[0027] Optionally, the control module is specifically used for: Based on the dynamic noise compensation algorithm, the sensor data is subjected to moving average filtering and data validity verification. When the verification result is invalid, the surrounding sensors are activated to perform interpolation and completion, obtain corrected sensor data, and perform feature conversion to obtain the biomechanical characteristics. Based on the total pressure value, pressure distribution characteristics, dispersion, and corresponding preset thresholds in the biomechanical characteristics, the usage state of the seat cushion is determined, and the operation mode of the massage air cushion base, the elastic fusion massage head, the heating plate, and the condensation plate are controlled according to the usage state of the seat cushion.

[0028] Specifically, firstly, the control module synchronously acquires the raw pressure data P_raw output from the pressure sensor and the triaxial acceleration data A_acc output from the IMU sensor at a fixed sampling period. Combined with a preset vehicle motion model, it calculates the additional dynamic pressure P_dyn generated by the vehicle motion based on the triaxial acceleration data. Then, it calculates the true pressure value P_true = P_raw - P_dyn using the formula, thus eliminating pressure data interference caused by vehicle acceleration / deceleration and vibration. After dynamic pressure stripping, the control module performs a moving average filtering process on the true pressure data. The moving window length N is set to 5 to 8, and the average value of continuous data within the window is taken to remove instantaneous spikes and outliers, further eliminating the noise effects of electromagnetic interference and minor vehicle vibrations. After filtering, the control module performs data validity verification to determine whether the output data of each sensor is within the normal pressure range. When the verification result is invalid, i.e., when the data from a single or a few sensors exceeds the reasonable range, the control module activates a peripheral sensor interpolation completion mechanism. Based on the data from the normal sensors adjacent to the faulty sensor, it completes the data through interpolation and averaging to form corrected sensor data, preventing overall function interruption due to single-point sensor failure. After obtaining the corrected sensor data, the control module performs feature conversion to extract biomechanical characteristic parameters, including the total pressure value F_total, pressure distribution characteristics in each region, and pressure dispersion. The total pressure value is the sum of the detection values ​​from all effective pressure sensors. The pressure distribution characteristics are represented by the regional pressure percentages of the ischial, thigh, and lateral regions. The pressure dispersion is calculated using the standard deviation of the pressure values ​​in each region. Subsequently, the control module compares these biomechanical characteristics with multiple preset thresholds to determine the current usage status of the seat cushion. The preset thresholds include a first threshold T1 corresponding to an unoccupied state, a second threshold T2 corresponding to children / small-sized users, and a third threshold T3 corresponding to instantaneous high pressure: when the total pressure value F_total is less than threshold T1, it is determined to be unoccupied or with items placed on it; when the total pressure value is between T1 and T2, and the pressure distribution matches the characteristics of a small-sized user, it is determined to be a child-occupied state; when the pressure distribution is extremely biased towards the edge areas, or the instantaneous peak pressure exceeds threshold T3, it is determined to be an improper usage state; all other situations are determined to be normal usage states. For different usage states, the control module outputs corresponding control strategies: In the unattended state, the massage air cushion base is kept at zero pressure, the elastic fusion massage head stops working, the heating element and the cooling element are both powered off and shut down, and the system enters low-power standby mode; when a child is riding in the system, the maximum heating temperature of the heating element is limited, the vibration intensity of the elastic fusion massage head is reduced, and the air pressure of the massage air cushion base is limited to a safe threshold; in the case of improper use, all active adjustment functions are immediately suspended, the air pump power is cut off, and the exhaust valve is opened to release pressure to avoid damage to the equipment or discomfort to the user caused by abnormal use; in the normal use state, the normal adaptive operation of each functional module is maintained.This embodiment effectively eliminates the interference of driving conditions on pressure sensor data through dual processing of vehicle dynamic pressure stripping and moving average filtering, improving the accuracy of data acquisition in complex environments; it enhances the system's fault tolerance through peripheral sensor interpolation completion mechanism, avoiding the impact of single-point sensor failure on functional operation; and it achieves graded control of unattended standby, child safety protection, and abnormal use protection through multi-level threshold usage status recognition, improving the safety and reliability of product use and reducing ineffective energy consumption.

[0029] Optionally, the control module is specifically used for: The biomechanical features are matched with stored templates to determine the user's identity; Based on the user's identity and personalized settings, the operation modes of the massage air cushion base, the elastic integrated massage head, the heating element, and the condensation element are controlled respectively.

[0030] Specifically, the control module extracts biomechanical features from the preprocessed sensor data, including parameters such as the relative position and spacing of the ischial tuberosities, the bilateral pressure ratio of the buttocks, the center of gravity of the thigh pressure distribution, total pressure, and contact area. These parameters are combined to form the user's pressure feature vector, or "pressure fingerprint." The control module performs similarity matching between the real-time collected pressure feature vector and the stored templates, i.e., the locally stored registered user identity templates. Euclidean distance is used to calculate the similarity value between the real-time feature vector and each template. When the similarity value exceeds a preset confidence threshold, a successful match is determined, and the current user's identity is identified. After identification, the control module automatically retrieves the user's pre-stored personalized settings, including the hardness parameters of each area of ​​the massage air cushion base, the vibration mode and intensity of the elastic integrated massage head, the temperature settings of the heating and cooling plates, and the automatic start / stop thresholds, and controls the operation of each actuator according to this personalized setting. If no matching template is found, the system enters guest mode and prompts the user to create a new identity template. Stored templates can be entered upon first use or via the app. This embodiment can achieve seamless user identification through similarity matching of pressure feature vectors. It can automatically call up personalized configurations without manual operation by the user, improving ease of use. In scenarios where multiple people share a vehicle, it can quickly adapt to the usage preferences of different drivers and passengers, optimizing the user experience.

[0031] Optionally, the control module is specifically used for: The usage time context features, response action data, implicit user feedback data, explicit user feedback data, and environmental features are obtained and anonymized to obtain multidimensional features; Based on the aforementioned multidimensional features, an active perturbation mechanism is used to determine high-quality training samples; Based on the high-quality training samples, a weighted moving average and bias integral algorithm is used to learn local rules at the edge. When the cloud learning conditions are met, a cloud deep learning model is used for incremental learning to obtain a personalized decision weight file. Based on the personalized decision weight file, the model is updated using an A / B testing gray-scale update strategy and an anomaly data cleaning strategy to obtain a new personalized adjustment model. Based on the biomechanical characteristics and the new personalized adjustment model, the operation modes of the massage air cushion base, the elastic fusion massage head, the heating element, and the condensation element are controlled respectively.

[0032] Specifically, firstly, the control module continuously collects multi-dimensional operational data, including usage time context features such as usage period and continuous usage duration; response action data such as target air pressure value, temperature level, and massage mode in the decision output; implicit user feedback data such as the difference and frequency of manual parameter adjustments and posture adjustment frequency; explicit user feedback data such as user ratings and mode favorites after use; and environmental characteristic data such as seat cushion temperature and humidity. All data to be uploaded to the cloud is anonymized, with only the pressure distribution feature vector and action records uploaded, not the original pressure image matrix. Simultaneously, the user's identity ID is locally hashed and encrypted before transmission to ensure user data privacy. In the initial stage of user use, the control module employs an active perturbation mechanism to obtain high-quality training samples: around the decision target value given by the basic general model, random small-amplitude jitters of ±2kPa are added. If the user does not manually intervene, this direction is determined as a positive preference and recorded; if the user manually adjusts parameters to correct the bias, this combination is marked as a negative preference and quickly reverts to the original value, accumulating effective preference samples through active trial and error. After completing sample accumulation, the system adopts a hierarchical incremental self-learning architecture. First, it performs local rule learning at the edge: the control module uses a weighted moving average and deviation integral algorithm to calculate a regional preference deviation vector δ_vector for each adjustment region. Whenever a user manually intervenes in the air pressure of a certain region, the adjustment difference Δ is recorded, and the local deviation value is updated using the exponential moving average formula: δ_new=α*Δ+(1-α)*δ_old, where the learning rate α is set to 0.3. The updated δ_new is stored in the deviation vector. When calculating the target air pressure subsequently, the control module directly superimposes this deviation vector to achieve rapid local response to user preferences. When the accumulated effective manual intervention records reach 50 sets, or every 7 days, the locally anonymized data is uploaded to the cloud. The cloud-based deep learning model performs incremental training based on the uploaded data, generating personalized decision weight files and distributing them to the seat cushion control module unit to replace the original decision layer parameters, completing cloud-based deep optimization. A dual protection mechanism is employed during model updates to prevent model degradation: First, A / B testing-based gray-scale updates are used, where the new model generated in the cloud is first activated during certain usage periods, while the old model is retained for the remaining periods. The frequency of manual user intervention is compared between the two periods; if the intervention rate significantly decreases under the new model, the application scope is gradually expanded. Second, an abnormal data cleaning strategy is used, marking improper sitting postures identified by the abnormal usage identification module, as well as all sensor data and adjustment actions generated during abnormal use by children, as dirty data and completely excluding them from the training dataset to prevent abnormal data from contaminating the preference model. The final optimized new personalized adjustment model is obtained, and the control module outputs control commands based on real-time biomechanical characteristics and the new model to drive the massage air cushion base, elastic integrated massage head, heating element, and cooling element.This embodiment utilizes a layered architecture that enables rapid learning at the edge and deep optimization in the cloud, balancing local response speed with model optimization accuracy. It rapidly acquires user preference samples through an active perturbation mechanism, improving model adaptation efficiency. Simultaneously, it ensures model stability through gray-scale updates and abnormal data cleaning, avoiding overfitting and data pollution. This allows the adjustment strategy to continuously align with users' actual usage preferences, thereby continuously improving driving comfort.

[0033] Optionally, the control module is specifically used for: Perform hardware self-test, load user configuration, and initialize timing status; Read the pressure sensor data and timestamp data and perform data preprocessing to obtain the pressure characteristics in the biomechanical features; Based on the pressure characteristics, the sitting state is determined according to the timing fault tolerance strategy and sitting posture classification conditions, and the total sitting time and the sitting time of the same type of sitting posture are determined. Based on the total sitting time and the sitting time of the same type of sitting posture, a sitting reminder scheme is determined, and based on the sitting reminder scheme, the operation mode of the elastic integrated massage head is controlled. Data is stored based on storage strategies, and reports are generated. Acquire abnormal status data, generate and store abnormal alert schemes.

[0034] Specifically, this embodiment describes a control implementation method for sedentary monitoring and posture health management. The first step is system initialization: the control module first performs a hardware self-test to verify the synchronization status of the RTC real-time clock. If it is not synchronized, it obtains the network time from the associated terminal via Bluetooth to complete the calibration. Then, it sends a test vibration command to the elastic fusion massage head and confirms that the vibration module is normal after receiving feedback. At the same time, it reads the baseline value of the pressure sensor and performs a self-test. After the self-test is completed, the control module loads the user-configured parameters. If there is no custom configuration, it loads the factory default parameters, including a default sedentary reminder threshold of 60 minutes, a default unhealthy sitting posture reminder threshold of 30 minutes, a default daily report generation cycle, and a default reminder method of APP reminder plus seat cushion-specific vibration. At the same time, it clears the temporary timing cache, marks the current sitting state as unoccupied, and completes the timing state initialization. The second step is data acquisition and preprocessing: the control module reads the pressure value data of the pressure sensor array, reuses the processed result after moving average filtering to avoid double calculation, and reads the timestamp data of the RTC clock, with the time accuracy controlled between 0.5 and 0.9 seconds. The control module sets a minimum threshold of 1000 kPa for total pressure after deducting the weight of the seat cushion as the minimum threshold for someone to sit. It also extracts the pressure percentages of each area—the ischial tuberosity, front of the thigh, back of the thigh, and left and right sides of the buttocks—from the pressure data as pressure characteristics for posture determination. The third step is sitting status recognition and timing statistics: the control module determines the sitting status based on the total pressure value and duration: if the total pressure is less than 1000 kPa and lasts for more than 1 minute, it is determined to be an empty seat, all timing is paused, and the time stamp of leaving the seat is recorded; if the total pressure is greater than 1000 kPa and lasts for more than 5 seconds, it is determined to be occupied, timing is started or resumed, and the time stamp of sitting start is recorded; if the total pressure is less than 800 kPa and lasts for less than 10 seconds, it is determined to be a brief absence from the seat, and timing is not interrupted. During timing, both total duration and duration of each sitting posture are statistically analyzed. The total prolonged sitting time is accumulated every second from the moment someone is determined to be seated, while brief absences are not interrupted; simultaneously, the start time, posture type, and duration are recorded for each type of sitting posture. The timing process incorporates a fault-tolerant mechanism: when the change in the proportion of the center of gravity and pressure distribution of the sitting posture characteristics exceeds 20% and lasts for more than 5 seconds, it is determined as a posture change, the current stage of timing ends, and a new stage of timing begins; if the sensor experiences a brief malfunction, timing continues according to the sitting posture type before the malfunction, and the duration is corrected after the malfunction is resolved. The fourth step is sitting posture health recognition and reminder triggering: the control module determines the sitting posture health status based on pressure characteristics. A healthy sitting posture must simultaneously meet the following conditions: the pressure in the ischial region accounts for 40%-60% of the total pressure, the center of gravity is located in the central area, the pressure difference between the left and right buttocks is less than 100kPa, and the pressure proportion on the front of the thigh is less than 25%.Unhealthy sitting postures are categorized into four types: excessive forward leaning (more than 25% pressure on the front of the thigh), excessive backward leaning (more than 70% pressure on the back of the buttocks and ischial tuberosities), unilateral lateral leaning (more than 60% pressure on one buttock with a pressure difference greater than 0.3 between the two sides), and crossing one's legs (a sudden increase of more than 50% pressure on one thigh and ischial tuberosities while the pressure on the other side drops sharply to less than 30%). When the total sitting time exceeds the user-set threshold, or when the continuous duration of a certain type of unhealthy sitting posture exceeds the set threshold, the control module triggers an alert, controls the elastic fusion massage head to output corresponding vibrations according to the preset mode, and simultaneously pushes an app alert. The same type of alert will not be triggered repeatedly within 10 minutes to avoid frequent interference. The fifth step is data storage and report generation: the control module adopts a real-time caching strategy, writing timing data to Flash storage every minute. By default, it retains 30 days of daily report data, 60 days of weekly report data, and 90 days of monthly report data. Data exceeding the threshold is automatically overwritten, and breakpoint resume is also supported. The control module generates reports according to a set cycle, summarizing the total duration, duration of each sitting posture type, and time period. It matches corresponding health hazard descriptions and correction suggestions for unhealthy sitting postures, converts this into structured text, and pushes it to the terminal. The generated reports are cached locally for repeated viewing. Finally, there are exception handling mechanisms: when an RTC clock failure occurs, timing is paused and a time calibration reminder is pushed; timing data is corrected after recovery. When three or more pressure sensors fail, posture recognition is paused, only sitting time is retained, and a fault notification is pushed. When the vibration module fails, it automatically switches to APP-only notification and marks the exception. When Bluetooth connection is lost, the notification information is cached locally and pushed again after reconnection. This embodiment achieves accurate statistics of prolonged sitting time and posture through precise sitting status determination and timing fault tolerance mechanisms. It achieves refined health monitoring through multi-dimensional posture classification and recognition, and guides users to correct poor posture in a timely manner through multi-channel reminder mechanisms. Simultaneously, it provides users with complete sedentary health data references through structured reports, comprehensively enhancing the health protection value of the cushion.

[0035] Optionally, the heating element has horizontally arranged elongated vents, the outer shell of the condenser plate adopts a honeycomb structure, and the heating and cooling layer also includes a heating chip connected to the heating element. The outer shell of the heating chip is wrapped with aluminum foil, and a temperature sensor is provided on the side of the heating chip. The control module is specifically used for: The operation of the heating element is controlled according to the heating adjustment command and the heating chip, and the operating mode of the heating element is controlled through the temperature feedback of the temperature sensor. The operation of the condenser is controlled according to the cooling adjustment command.

[0036] Specifically, the heating element shell of the heating layer features horizontally arranged elongated vents. This structure improves heat conduction efficiency, accelerates airflow after heating is turned off, achieving rapid heat dissipation and cooling, while simultaneously increasing the fault tolerance of wiring connections while ensuring structural strength. The condenser fin is located below the heating element, suspended between the sound and heat insulation structures. The upper and lower shells of the condenser fin adopt a honeycomb structure, composed of multiple evenly and densely arranged equilateral hexagonal holes, the overall shape of which is also equilateral hexagonal. The air pockets formed by the honeycomb holes act as insulators, reducing heat transfer. Combined with the honeycomb structure of other layers of the cushion, this gradually reduces operating noise, achieving a quiet operation. The heating element is equipped with a heating chip, whose shell is wrapped in aluminum foil to insulate and reflect the heat from the heating element, preventing high temperatures from affecting the chip's operational stability. An independent temperature sensor is fixed to the side of the heating chip, directly connected to it, for real-time monitoring of the average temperature and rate of temperature change in the cushion area. The control flow of the heating function is as follows: The control module receives the heating adjustment command sent by the user through the terminal. After parsing the parameters in the command, such as the heating temperature level, custom temperature threshold, and intermittent heating rules, it sends a PWM pulse width modulation signal with a specific duty cycle to the heating chip. After parsing the command, the heating chip starts the heating element to begin heating. During the heating process, the temperature sensor detects the seat cushion temperature data in real time and feeds it back to the heating chip. The heating chip synchronously transmits the temperature data to the control module. The terminal interface can display the real-time temperature and the estimated time to reach the set temperature. When the seat cushion temperature reaches the set threshold, the heating chip sends a signal back to the control module. The control module sends a constant temperature command, and the heating chip adjusts the PWM signal duty cycle to reduce the heating element power to maintain a constant temperature. The heating temperature can achieve precise temperature control within the range of 35℃ to 42℃ ±0.5℃, supporting global or local area heating. The system has independent overheat protection logic: when the temperature sensor detects that the temperature exceeds the set threshold, the user-defined threshold, or the maximum safe temperature limit, the temperature sensor immediately transmits an abnormal signal to the heating chip. After analyzing the signal, the heating chip can directly cut off the power supply to the heating element without the control module's instruction, stopping heating. Simultaneously, the heating chip transmits the overheat protection signal to the control module, which then pushes a notification to the user terminal, informing the user that the product has entered overheat protection mode and will be restarted after the heating element has completely cooled down. The control flow for the condensation function is as follows: the control module receives the cooling adjustment command from the user, analyzes parameters such as the cooling level and intermittent cooling rules, and sends a control signal to the condenser's drive module to start the condenser and adjust the cooling power to cool the seat cushion. The condensation function's lower temperature limit can reach 10℃, supporting global or localized cooling.This embodiment achieves both temperature control efficiency and quiet operation through the ventilation structure of the heating element and the honeycomb structure of the condenser. The heating chip design with aluminum foil insulation improves the reliability of components in high-temperature environments. It adopts PWM precise temperature control to achieve stable temperature regulation, and provides dual safety protection with an independent overheat protection mechanism. At the same time, it supports a variety of custom temperature control rules to meet the temperature preferences and usage scenarios of different users.

[0037] Optionally, the control module is specifically used for: Perform hardware self-test, parameter loading, and calibration initialization; Based on the aforementioned biomechanical characteristics, the total weight under force, regional pressure distribution, and pressure center of gravity are determined. Using big data cloud technology, the sitting posture category is determined based on the total weight under force, the regional pressure distribution, and the pressure center of gravity. Based on preset user priority or the biomechanical characteristics, a sitting posture protection mode is determined, and the target air pressure value of the massage air cushion base is determined according to the sitting posture protection mode. The target air pressure value is converted into an air pump control command, and the massage air cushion base is adjusted using PID closed-loop feedback. The system obtains information on the fault conditions of the massage layer and the sensing layer, and controls the operating status of the massage air cushion base based on the fault conditions.

[0038] Specifically, this embodiment describes a control implementation method for adaptive adjustment of seat cushion firmness. The first step is system initialization: the control module first performs a hardware self-test, reads the baseline values ​​of each pressure sensor to determine if the sensor is abnormal; sends test commands to the air pump, solenoid valve, and buffer valve to verify the on / off status of the air circuit components; and simultaneously checks the Bluetooth communication status to confirm normal connection with the terminal. After the self-test is completed, the operating parameters are loaded, prioritizing the reading of stored user-defined configurations. If no user-defined configuration is available, the factory default parameters are loaded, including an initial air pressure of 0.08 MPa and a default mode of ischial comfort mode. Finally, calibration initialization is performed, reading the sensor calibration coefficients, subtracting the seat cushion's own weight, and correcting the sensor offset to complete the baseline calibration of the pressure value. The second step is human feature recognition: based on the pre-processed biomechanical characteristics, the control module calculates the total weight under force, regional pressure distribution, and pressure center of gravity. The total weight is calculated using the formula: Weight (kg) = Total pressure (kPa) × Contact area (m²) ÷ Gravitational acceleration (9.8 m / s²), with the seat cushion's weight already deducted. The regional pressure distribution represents the average pressure of each region, equal to the total regional pressure divided by the regional contact area. The regional contact area is estimated by multiplying the number of triggered sensors by the coverage area of ​​a single sensor. The pressure center of gravity is calculated using the pressure difference between each sensor, yielding the planar coordinates of the pressure center of gravity. Based on these parameters, the control module, combined with big data cloud technology, determines the sitting posture category: a normal sitting posture is characterized by 40%-60% pressure in the ischial tuberosity region, uniform pressure on the front and back of the thighs, and the center of gravity located in the center; a forward-leaning posture is characterized by greater than 45% pressure in the front thigh region, less than 30% in the ischial tuberosity region, and a forward-leaning center of gravity; a backward-leaning posture is characterized by greater than 45% pressure in the back of the buttocks region, less than 30% in the ischial tuberosity region, and a backward-leaning center of gravity; and a side-leaning posture is characterized by greater than 50% pressure in one buttock region, less than 20% on the other, and a shift in the center of gravity towards the higher pressure side. The third step is mode decision-making and target air pressure calculation: The control module first reads the user-set mode priority list, with custom modes having higher priority than those automatically recommended by the algorithm; if no priority is set, the optimal mode is automatically recommended based on human characteristics, for example, the ischial protection mode is recommended first when the pressure in the ischial region is too high. The system includes three core adjustment modes, and the target air pressure of each mode is dynamically calculated by combining human body weight, regional pressure, and sitting posture: one is the ischial protection mode, and the base air pressure calculation formula is: ischial region air cushion pressure = base value + (human body weight - preset weight) ÷ preset coefficient; at the same time, sitting posture adaptation correction is added, for example, when sitting in a forward-leaning posture, the ischial region air pressure is increased by αkPa, and the corresponding front thigh region is decreased by αkPa; if the ischial region pressure is detected to be greater than 70kPa, the corresponding air cushion pressure is continuously increased until the ischial region pressure is lower than 70kPa, and the local pressure is reduced by increasing the contact area of ​​the air cushion.Secondly, there's the blood flow maximization mode. Its core logic is to adjust the air pressure in all areas according to the pressure uniformity command, ensuring that the pressure values ​​in each area are basically consistent. Simultaneously, it makes small adjustments to the air pressure in each area every 30 to 35 seconds, preventing prolonged localized ischemia through pressure fine-tuning. When the sitting posture is tilted to the side, if the pressure on one side is too high, the corresponding air cushion on that side is increased to balance the pressure distribution. Thirdly, there's the comfort mode. It calculates the pressure difference in all areas. When the pressure difference exceeds a set threshold, it increases the air pressure in the low-pressure area and decreases the air pressure in the high-pressure area until the difference is within the threshold range. It also makes fine-tuning based on the user's historical usage preferences. On top of the basic mode, there's an automatic posture adaptation rule: in a forward-leaning posture, the air cushion on the front of the thighs automatically deflates to eliminate the feeling of localized impact; in a backward-leaning posture, the air cushion in the ischial tuberosity area inflates to increase the contact area and reduce ischial tuberosity pressure; in a side-leaning posture, the air cushion on the pressure-bearing side inflates to balance the pressure on both sides. When switching postures, a gradual pressure adjustment is used, with each adjustment a small value at 0.5-second intervals to avoid sudden changes in sensation. The fourth step is air pressure execution and closed-loop feedback: The control module converts the target air pressure value into an air pump control command. When the target air pressure is higher than the current air pressure, the air pump operates at the corresponding PWM duty cycle to inflate the system. After reaching the target value, the solenoid valve and air pump are shut off. When the target air pressure is lower than the current air pressure, the air pump is shut off, and air is slowly released through the solenoid valve in conjunction with the buffer valve. When the target air pressure is equal to the current air pressure, the air valve remains closed to maintain the air pressure. During the adjustment process, a lightweight PID algorithm is used to achieve closed-loop feedback: the proportional term P adjusts the adjustment amplitude according to the difference between the target pressure and the actual pressure; the larger the difference, the larger the adjustment amplitude. The integral term I is used to eliminate long-term small deviations. The derivative term D is used to suppress the overshoot of air pressure regulation and avoid excessive adjustment amplitude. The rate of a single pressure adjustment does not exceed a preset threshold to ensure a smooth and comfortable adjustment process. The fifth step is safety protection: When a single pressure sensor fails, the data is supplemented using the differential values ​​of surrounding sensors, and the system continues to operate normally. When three or more sensors fail, the system switches to safety mode, and all air cushions are slowly released to normal pressure, while a fault notification is pushed out. When the air pump malfunctions and cannot inflate, the solenoid valve closes to maintain the current air pressure and pushes the faulty valve out. When the solenoid valve fails and cannot deflate, the backup valve is activated for emergency pressure relief. If no effective pressure is detected for five consecutive minutes, the system enters automatic sleep mode, reduces the sampling frequency, disconnects power to the air pump and solenoid valve, and retains only Bluetooth communication. This embodiment can achieve adaptive adjustment of the firmness to suit different sitting postures and user needs through precise human body force and posture recognition, combined with multi-mode air pressure adjustment strategies. The accuracy of air pressure adjustment is ensured through PID closed-loop feedback, and the progressive pressure adjustment improves the user experience. At the same time, the comprehensive multi-level safety protection mechanism can effectively avoid equipment failure and usage risks, comprehensively improving the support comfort and safety of the seat cushion.

[0039] Optionally, pressure sensors are provided on both sides of the massage air cushion base, and multiple pressure sensors are symmetrically distributed. The massage air cushion base includes an air cushion, an air pump, a solenoid valve, and a buffer valve connected in sequence.

[0040] Specifically, this embodiment is an implementation of the sensing and air circuit structure for a massage air cushion base. The pressure sensors matching the massage air cushion base are arranged in a symmetrical distribution: specifically, pressure sensors are respectively arranged on the sides of two sets of air bags parallel to the two side directions of the cushion, and four additional pressure sensors are added in an indented square arrangement, totaling eight pressure sensors which are respectively distributed at the four corner positions and the inner four corner positions of the cushion, presenting a symmetrical layout as a whole. This distribution mode can realize pressure coverage detection of the entire pressure-bearing area, and the user's center of gravity distribution and position change can be quickly calculated through the pressure difference values of the sensors at each corner: when the user leans forward in sitting posture, the pressure detection values at the two front corners of the cushion rise, and the pressure values at the two rear corners decrease; when the user sits sideways, the pressure detection values at the two corners on one side rise, and the corresponding values on the other side decrease. This symmetrical distribution structure can accurately capture subtle changes in human posture, reduce monitoring errors caused by local body movement, and provide accurate data support for sedentary protection and softness and hardness adjustment; at the same time, it can improve the accuracy of sitting recognition: when three or more pressure sensors detect a pressure value corresponding to a weight of 25kg or more, it is determined that a person is sitting on the cushion, and the preset intelligent functions can be started; when three or more sensors detect a pressure value corresponding to a weight of less than 10kg, it is determined that the user has left the seat, and the corresponding intelligent functions are suspended. This determination logic can effectively avoid false triggering of functions caused by placed items. In addition, the four corners of the cushion are key areas for structural load-bearing, and the pressure sensors installed at these positions can be directly fixed to the cushion support, featuring higher physical stability and being less prone to displacement and falling off due to sitting posture changes and operation of internal components. The air circuit system of the massage air cushion base is formed by sequentially connecting the air cushion, the air pump, the solenoid valve and the buffer valve. The air pump provides power for the entire air cushion system, adopts a positive-displacement micro diaphragm pump, generates air pressure by physically capturing and moving air, and delivers compressed air to the air cushion; the air pump and the solenoid valve group are integrally arranged at the rear or side of the cushion, which can adapt to installation space constraints such as those of automobile seats, and is simultaneously equipped with a silencing cavity structure to reduce operating noise. The solenoid valve adopts an electric coil-driven plunger actuator, which receives electric signals to control the opening and closing of the valve, and is used to precisely control the on-off of the air circuit between the air pump and the air cushion; priority rules are set for the operation sequence of the solenoid valve: when opening, the air pump starts first and the solenoid valve opens later, so that the air pump accumulates a certain amount of compressed air in advance to achieve a fast inflation response; when closing, the solenoid valve closes first and the air pump stops later, so as to avoid component wear caused by residual pressure in the air circuit. The buffer valve is a regulating and buffering element of the air circuit system. A honeycomb-shaped buffer throttling structure is arranged inside the buffer valve, which is a circular rubber sheet with multiple uniformly arranged hexagonal holes. This structure can limit the airflow velocity, reduce the generation of vortexes, make the airflow and air pressure output more stable, simultaneously consume the kinetic energy of the airflow, slow down the inflation speed, make the air cushion expand smoothly, and avoid physical discomfort caused by instantaneous expansion of the air cushion due to airflow impact.The buffer valve also incorporates a channel adjustment structure located at the front end of the buffer throttling structure. By controlling the cross-sectional area of ​​the channel within the valve, it works in conjunction with the buffer throttling structure to achieve pressure relief and flow diversion, reducing pressure peaks and preventing instantaneous high pressure from damaging the venting pipes and air cushion, as well as preventing airflow impact from causing component loosening. This embodiment utilizes eight symmetrically distributed pressure sensors to achieve precise and comprehensive pressure sensing, improving the accuracy of posture recognition and body detection, while also enhancing structural stability. Through the sequential coordination of the air pump, solenoid valve, and buffer valve, smooth and precise air cushion pressure regulation is achieved, reducing operating noise and airflow impact, improving comfort during pressure regulation, and extending the service life of the air circuit system.

[0041] Optionally, the base layer includes sound and heat insulation cotton, a rigid base plate, and an anti-slip base arranged sequentially from top to bottom. The sound and heat insulation cotton is located below the heating and cooling layer, and the anti-slip base is used for installation on the external equipment.

[0042] Specifically, the base layer consists of sound and heat insulation cotton, a hard base plate, and an anti-slip base, arranged from top to bottom, providing structural support, sound and heat insulation, and installation and fixing functions for the seat cushion.

[0043] The sound and heat insulation cotton is placed below the heating and cooling layers, serving as the top layer of the base layer and directly supporting the heating and cooling functional modules above. The cotton employs a honeycomb structure design combining natural porous fibers and an MLV damping layer. The numerous fiber gaps within the cotton trap air, forming an insulating barrier that effectively blocks heat conduction. This reduces heat loss from the heating layer, improving heating efficiency, and also prevents cold air from the condensation layer from transferring downwards, enhancing cooling performance. Simultaneously, it prevents temperature changes from affecting the control module circuitry and air pump components below. The outer surface of the cotton is coated with an aluminum foil layer, reflecting 80%-90% of radiant heat. When heat is transferred downwards from the upper layer, the aluminum foil layer reflects most of the heat energy, with the remaining small amount absorbed by the internal fiber insulation material. Through the triple insulation effect of the honeycomb structure, fiber air layer, and aluminum foil reflection, excellent heat insulation performance is achieved, protecting the underlying electronic components from temperature shocks. Meanwhile, the porous honeycomb structure of the sound and heat insulation cotton, combined with the MLV damping layer, effectively absorbs and reduces noise generated by the operation of the air pump and airflow. Combined with the honeycomb noise reduction structure of each functional layer above, it achieves layer-by-layer noise reduction, ultimately reaching a near-silent operating effect, preventing noise from the seat cushion from disturbing passengers. A rigid base plate is located beneath the sound and heat insulation cotton, providing rigid structural support for the entire seat cushion. Made of high-strength rigid materials, it can bear the weight of the human body and each functional module, ensuring the overall structural strength of the seat cushion and preventing deformation during use. The rigid base plate has pre-drilled mounting positions for various components, including mounting grooves and fixing structures for components such as the air pump, solenoid valve assembly, and control module circuit board. This allows for secure mounting of functional components on the base plate, preventing component displacement, collisions, abnormal noises, or damage during vehicle operation. The base plate also has pre-drilled wiring channels and air passages, ensuring neat internal wiring and air pipe layout, improving assembly efficiency and operational reliability. The anti-slip base is located beneath the rigid base plate, forming the bottom layer of the base layer. It's used to secure the seat cushion to external devices such as car seats. The bottom surface of the anti-slip base uses a high-friction coefficient anti-slip material, increasing friction with the car seat surface and preventing the cushion from sliding or shifting due to user movement or vehicle acceleration / deceleration, thus improving safety and stability. The anti-slip base is compatible with various car seat surfaces, allowing for stable placement without complex installation, and will not damage the original seat surface material. This embodiment utilizes a multi-layered composite base structure to simultaneously achieve multiple functions: heat insulation, noise suppression, structural support, and anti-slip fixation. The triple heat insulation design effectively improves temperature control efficiency and protects underlying components, the honeycomb noise reduction structure optimizes the user experience, the rigid base plate ensures structural strength, and the anti-slip base enhances ease of installation and stability, comprehensively improving the overall reliability and quality of the seat cushion.

[0044] Optionally, such as Figure 1As shown, the seat cushion body also includes a surface contact layer disposed above the massage layer. The surface contact layer is made of skin-friendly and breathable fabric and is equipped with a temperature and humidity sensor.

[0045] Specifically, the seat cushion also features a surface contact layer, located on top of the massage layer, which is the surface structure that directly contacts the human body. This surface contact layer is made of skin-friendly and breathable fabric, which has a soft, skin-friendly feel, enhancing comfort and preventing discomfort and skin irritation from prolonged contact. The fabric employs a honeycomb hexagonal mesh structure; the numerous evenly distributed hexagonal holes significantly improve the fabric's breathability, allowing air to circulate between the body and the seat cushion, effectively dissipating heat and moisture from the skin and preventing the accumulation of heat and moisture in areas such as the perineum caused by prolonged sitting, thus reducing the risk of skin problems such as eczema and folliculitis. The mesh structure also has a certain degree of deformation capability, slightly deforming under body pressure to improve fit and comfort, optimizing the riding experience. Integrated within the surface contact layer are temperature and humidity sensors, positioned against the inside of the fabric, which can monitor the temperature and humidity data of the area where the seat cushion contacts the body in real time and transmit the data to the control module. Temperature and humidity data serve two main purposes: firstly, they provide feedback parameters for the temperature control function, aiding in the precise adjustment of heating and condensation functions to ensure the seat cushion temperature is more closely aligned with the actual comfort level; secondly, they monitor the humidity and heat status of the contact area. When humidity exceeds a set threshold, the system adjusts ventilation or cooling strategies to improve the dryness of the contact area. The surface contact layer is both wear-resistant and easy to clean. The fabric surface is treated with stain resistance, allowing for easy wiping and cleaning of stains during daily use, extending the seat cushion's lifespan. The honeycomb mesh structure of the fabric works in conjunction with the underlying massage heads and air cushion structure to evenly distribute the massage vibrations and support adjustment forces to the body, without weakening the massage and support effects due to the surface structure. Furthermore, the surface fabric is removable, allowing it to be removed from the seat cushion for washing and replacement, maintaining hygiene and meeting long-term usage needs. In terms of functional linkage, when the temperature and humidity sensor detects excessively high temperature or humidity in the contact area, the control module automatically activates the condensation function or adjusts the cooling level to lower the temperature and accelerate moisture dissipation; when the temperature is detected to be too low, the heating function is automatically activated to enhance seating comfort. Temperature and humidity data are simultaneously incorporated into the environmental characteristic data of the self-learning system to optimize the temperature control model, making the temperature regulation strategy more aligned with the actual usage environment and user comfort needs. This embodiment utilizes skin-friendly and breathable honeycomb mesh fabric to enhance contact comfort and breathability, mitigating the problem of damp heat from prolonged sitting. The integrated temperature and humidity sensor enables real-time sensing of the environment in the contact area, providing accurate ergonomic feedback data for temperature control. Furthermore, the detachable design facilitates cleaning and maintenance, comprehensively improving the surface experience and hygiene of the cushion.

[0046] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc. In this application, the units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of the present invention according to actual needs. Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated units can be implemented in hardware or as software functional units.

[0047] While the present invention has been disclosed above, its scope of protection is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention, and all such changes and modifications will fall within the scope of protection of the present invention.

Claims

1. A smart seat cushion, characterized in that, The device includes a seat cushion body and a control module. The seat cushion body is provided with a massage layer, a sensing layer, a heating and cooling layer and a base layer from top to bottom. The massage layer includes a massage air cushion base with uniformly arranged massage air cushion bases and elastic integrated massage heads that correspond one-to-one with the massage air cushion bases. The sensing layer is used to acquire sensor data, including temperature and humidity sensor data, pressure sensor data, IMU sensor data and piezoelectric film sensor data. The heating and cooling layer includes a heating element and a condensation element disposed below the heating element. The base layer is used for mounting on external devices. The control module is used to preprocess the sensor data based on a dynamic noise compensation algorithm, extract biomechanical features, and control the operation mode of the massage air cushion base, the elastic fusion massage head, the heating element, and the condensation element according to the biomechanical features.

2. The smart seat cushion according to claim 1, characterized in that, The control module is specifically used for: Based on the dynamic noise compensation algorithm, the sensor data is subjected to moving average filtering and data validity verification. When the verification result is invalid, the surrounding sensors are activated to perform interpolation and completion, obtain corrected sensor data, and perform feature conversion to obtain the biomechanical characteristics. Based on the total pressure value, pressure distribution characteristics, dispersion, and corresponding preset thresholds in the biomechanical characteristics, the usage state of the seat cushion is determined, and the operation mode of the massage air cushion base, the elastic fusion massage head, the heating plate, and the condensation plate are controlled according to the usage state of the seat cushion.

3. The smart seat cushion according to claim 1, characterized in that, The control module is specifically used for: The biomechanical features are matched with stored templates to determine the user's identity; Based on the user's identity and personalized settings, the operation modes of the massage air cushion base, the elastic integrated massage head, the heating element, and the condensation element are controlled respectively.

4. The smart seat cushion according to claim 1, characterized in that, The control module is specifically used for: The usage time context features, response action data, implicit user feedback data, explicit user feedback data, and environmental features are obtained and anonymized to obtain multidimensional features; Based on the aforementioned multidimensional features, an active perturbation mechanism is used to determine high-quality training samples; Based on the high-quality training samples, a weighted moving average and bias integral algorithm is used to learn local rules at the edge. When the cloud learning conditions are met, a cloud deep learning model is used for incremental learning to obtain a personalized decision weight file. Based on the personalized decision weight file, the model is updated using an A / B testing gray-scale update strategy and an anomaly data cleaning strategy to obtain a new personalized adjustment model. Based on the biomechanical characteristics and the new personalized adjustment model, the operation modes of the massage air cushion base, the elastic fusion massage head, the heating element, and the condensation element are controlled respectively.

5. The smart seat cushion according to claim 1, characterized in that, The control module is specifically used for: Perform hardware self-test, load user configuration, and initialize timing status; Read the pressure sensor data and timestamp data and perform data preprocessing to obtain the pressure characteristics in the biomechanical features; Based on the pressure characteristics, the sitting state is determined according to the timing fault tolerance strategy and sitting posture classification conditions, and the total sitting time and the sitting time of the same type of sitting posture are determined. Based on the total sitting time and the sitting time of the same type of sitting posture, a sitting reminder scheme is determined, and based on the sitting reminder scheme, the operation mode of the elastic integrated massage head is controlled. Data is stored based on storage strategies, and reports are generated. Acquire abnormal status data, generate and store abnormal alert schemes.

6. The smart seat cushion according to claim 1, characterized in that, The heating element has horizontally arranged elongated vents, the outer shell of the condenser plate adopts a honeycomb structure, and the heating and cooling layer also includes a heating chip connected to the heating element. The outer shell of the heating chip is wrapped with aluminum foil, and a temperature sensor is provided on the side of the heating chip. The control module is specifically used for: The operation of the heating element is controlled according to the heating adjustment command and the heating chip, and the operating mode of the heating element is controlled through the temperature feedback of the temperature sensor. The operation of the condenser is controlled according to the cooling adjustment command.

7. The smart seat cushion according to claim 1, characterized in that, The control module is specifically used for: Perform hardware self-test, parameter loading, and calibration initialization; Based on the aforementioned biomechanical characteristics, the total weight under force, regional pressure distribution, and pressure center of gravity are determined. Using big data cloud technology, the sitting posture category is determined based on the total weight under force, the regional pressure distribution, and the pressure center of gravity. Based on preset user priority or the biomechanical characteristics, a sitting posture protection mode is determined, and the target air pressure value of the massage air cushion base is determined according to the sitting posture protection mode. The target air pressure value is converted into an air pump control command, and the massage air cushion base is adjusted using PID closed-loop feedback. The system obtains information on the fault conditions of the massage layer and the sensing layer, and controls the operating status of the massage air cushion base based on the fault conditions.

8. The smart seat cushion according to claim 1, characterized in that, The massage air cushion base is equipped with pressure sensors on both sides, and multiple pressure sensors are symmetrically distributed. The massage air cushion base includes an air cushion, an air pump, a solenoid valve, and a buffer valve connected in sequence.

9. The smart seat cushion according to claim 1, characterized in that, The base layer includes, from top to bottom, sound and heat insulation cotton, a rigid base plate, and an anti-slip base. The sound and heat insulation cotton is located below the heating and cooling layer, and the anti-slip base is used for installation on the external equipment.

10. The smart seat cushion according to claim 1, characterized in that, The seat cushion body also includes a surface contact layer disposed above the massage layer. The surface contact layer is made of skin-friendly and breathable fabric and is equipped with a temperature and humidity sensor.

Images