A pressure-sensing based membrane switch system

Abstract

This invention relates to the fields of smart home and flexible sensing technology, and discloses a pressure-sensing-based thin-film switch system, including a flexible thin-film sensor array, a pressure characteristic parameter conversion module, a central control unit, an actuation and adjustment mechanism, and a wireless communication module. The flexible thin-film sensor array responds to external loads, generating changes in electrical parameters. The conversion module converts these changes into raw pressure matrix data. The central control unit uses a discrete Laplace convolution kernel to decouple the data from tension interference, generating true pressure matrix data. It then calculates the local cumulative fatigue index by combining a sliding time window and a microcirculation occlusion threshold. When the fatigue index exceeds the limit, the central control unit drives the actuation and adjustment mechanism to perform coordinated adjustment of sinking unloading and surrounding upward movement based on a neighborhood compensation mechanism. This invention eliminates tension interference in flexible substrates, achieves quantitative assessment and active morphological reconstruction of accumulated tissue damage, and maintains the mechanical balance of the support surface.

Description

Technical Field

[0001] This invention relates to the fields of smart home and flexible sensing technology, and in particular to a pressure-sensing membrane switch system. Background Technology

[0002] With the rapid development of IoT technology and the smart home industry, users' demands for home furnishings such as sofas and mattresses have shifted from basic functionality to comfort and health management. Acquiring interaction data between the human body and the support surface has become the foundation for realizing intelligent upgrades of products.

[0003] Traditional membrane switches typically function only as binary input devices to control the on / off state or function switching of electronic devices. They cannot provide quantitative data on pressure distribution or provide feedback adjustments based on the user's real-time pressure. To improve comfort and health protection, existing technologies are beginning to explore the introduction of flexible sensor arrays to monitor pressure distribution.

[0004] However, existing pressure monitoring and regulation systems based on flexible films have limitations in practical applications. Since most flexible film sensors are constructed from continuous elastic substrates, when a specific area is subjected to vertical loads and deforms, the lateral stretching of the substrate transmits tension to adjacent unloaded areas, causing the sensor output to contain distorted signals due to tension interference. Simultaneously, existing comfort assessment and health warning mechanisms typically only focus on instantaneous pressure amplitudes, lacking consideration of the duration of pressure application, making it difficult to establish integral models that conform to human physiological characteristics to quantify the cumulative ischemic risk of soft tissues caused by continuous pressure. Regarding actuator adjustment, existing actuators often employ single-position sinking avoidance control logic, which can easily lead to local support collapse and fails to consider the impact of load transfer on surrounding areas, making it difficult to maintain the overall mechanical balance of the support surface. Summary of the Invention

[0005] The purpose of this invention is to provide a pressure-sensing membrane switch system that solves the problem of pressure data distortion caused by tension interference from the deformation of flexible materials in existing pressure monitoring devices, as well as the problem that existing support devices lack quantitative assessment and active adjustment mechanisms for cumulative tissue damage.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] This invention provides a pressure-sensing-based thin-film switch system, comprising a flexible thin-film sensor array, a pressure characteristic parameter conversion module, a central control unit, an actuation and adjustment mechanism, and a wireless communication module. The flexible thin-film sensor array serves as the sensing front-end and switch triggering medium, responding to changes in electrical characteristic parameters caused by external loads. The pressure characteristic parameter conversion module is electrically connected to the flexible thin-film sensor array, collecting the changes in electrical characteristic parameters and converting them into digitized raw pressure matrix data. The central control unit receives the raw pressure matrix data, performs tension interference decoupling on the raw pressure matrix data to generate true pressure matrix data, calculates a local cumulative fatigue index based on the true pressure matrix data, and generates an adjustment command based on the local cumulative fatigue index. The actuation and adjustment mechanism is disposed below the flexible thin-film sensor array, performing active morphological adjustment in response to the adjustment command. The wireless communication module transmits the true pressure matrix data and the local cumulative fatigue index to a terminal display device.

[0008] Preferably, the flexible thin-film sensor array adopts a multi-layer composite structure, which sequentially comprises a panel layer, a circuit sheet layer, and an adhesive backing layer from top to bottom. The circuit sheet layer has an array of conductive lines printed on it, comprising multiple sets of sensing nodes arranged in a row and column matrix. A pressure-sensing functional layer is covered on the side of the panel layer facing the circuit sheet layer, and this pressure-sensing functional layer is suspended above the array of conductive lines. When an external load is applied to the panel layer, it drives the pressure-sensing functional layer to make physical contact with the array of conductive lines, thereby generating a change in the electrical characteristic parameters at the contact point.

[0009] Preferably, when processing the original pressure matrix data, the central control unit performs a sliding window convolution operation on the original pressure matrix data using a discrete Laplacian convolution kernel, and extracts the local curvature features at each node location by calculating the spatial second derivative. The central control unit then performs a weighted calculation on the local curvature features and the original pressure matrix data, removes the lateral tension interference component from the original pressure matrix data, and outputs the true pressure matrix data.

[0010] Preferably, the central control unit constructs a sliding time window, mapping the continuously acquired time-varying sequence of real pressure matrix data into the sliding time window. The central control unit introduces an ischemic injury acceleration coefficient and a microcirculation occlusion pressure threshold, performs spatiotemporal integration on the real pressure matrix data within the sliding time window, and generates a local cumulative fatigue index by determining the relationship between the pressure value and the microcirculation occlusion pressure threshold and applying a nonlinear weighting. The local cumulative fatigue index characterizes the degree of ischemic injury risk at the sensing node location due to continuous pressure. The microcirculation occlusion pressure threshold refers to the critical pressure value that causes blockage of microcirculatory blood flow in human skin.

[0011] Preferably, the system further includes a user feedback interface located on the terminal display device. The central control unit receives subjective evaluation instructions input by the user through the user feedback interface and inputs the subjective evaluation instructions into a machine learning-based parameter adaptation module. The parameter adaptation module dynamically fine-tunes the ischemic injury acceleration coefficient or the microcirculation occlusion pressure threshold used to calculate the local cumulative fatigue index based on the subjective evaluation instructions.

[0012] Preferably, the central control unit invokes a logical partition mapping table to divide the flexible thin-film sensor array into several logical sensing zones and establishes a mapping relationship, so that each logical sensing zone uniquely corresponds to and controls a corresponding physical execution unit at its geometric center. Based on the mapping relationship, the central control unit aligns the coordinates of the sensing nodes of the flexible thin-film sensor array with the spatial positions of the physical execution units. Furthermore, the central control unit obtains multiple local cumulative fatigue indices contained within each logical sensing zone, and based on an extreme value priority strategy, extracts the maximum value from the multiple local cumulative fatigue indices to generate the regional aggregate fatigue degree corresponding to the logical sensing zone.

[0013] Preferably, the central control unit determines whether the regional aggregate fatigue degree corresponding to the current physical execution unit exceeds a preset safety threshold. When the regional aggregate fatigue degree exceeds the preset safety threshold, the central control unit generates the adjustment command based on a neighborhood compensation mechanism. The adjustment command is used to drive the current physical execution unit to sink and drive the physical execution units located in the neighborhood of the current physical execution unit to rise, so as to maintain the mechanical balance of the support surface. When the regional aggregate fatigue degree does not exceed the preset safety threshold, the morphological parameters of the current physical execution unit remain unchanged, and no adjustment command is generated. The preset safety threshold refers to the pressure accumulation limit that causes deep tissue damage.

[0014] Preferably, the wireless communication module uses a low-power Bluetooth, Wi-Fi, or ZigBee communication module. When the terminal display device receives the real pressure matrix data and performs display, it uses a bicubic interpolation algorithm to spatially reconstruct the real pressure matrix data, generating a high-resolution three-dimensional data stream, and renders a three-dimensional pressure distribution image based on the high-resolution three-dimensional data stream.

[0015] Preferably, the pressure characteristic parameter conversion module acquires the changes in electrical characteristic parameters at the sensing nodes in the flexible thin-film sensor array through a physical addressing mechanism of row scanning and column gating, and obtains the digitized measured digital voltage value. The pressure characteristic parameter conversion module uses a voltage-pressure linearization calibration strategy to process the measured digital voltage value, eliminate zero-point drift and range error, and output the original pressure matrix data.

[0016] In summary, the present invention has at least one of the following beneficial technical effects:

[0017] 1. This invention uses a central control unit to extract local curvature features using a discrete Laplacian convolution kernel, thereby removing the transverse tension interference component from the original pressure matrix data. This solves the problem of signal entanglement in non-stressed areas caused by substrate connectivity in flexible thin-film sensor arrays, eliminates the interference of tension on vertical pressure readings, ensures the accuracy of the real pressure matrix data, and provides a reliable data foundation for subsequent fatigue assessment.

[0018] 2. This invention constructs a spatiotemporal integral evaluation model based on a sliding time window, and introduces a microcirculation occlusion pressure threshold and an ischemic injury acceleration coefficient to perform nonlinear weighted calculation of pressure data. This changes the limitation of traditional equipment that only monitors instantaneous pressure amplitude, and can quantitatively characterize the cumulative ischemic injury risk of soft tissue caused by continuous pressure. This makes the system's early warning mechanism conform to the physiological characteristics of human microcirculation and improves the effectiveness of pressure ulcer risk identification.

[0019] 3. This invention adopts a collaborative control strategy based on a neighborhood compensation mechanism. When the regional aggregate fatigue level is detected to be excessive, the corresponding physical execution unit is driven to sink and unload, and the neighboring unit is driven to rise to take over the load. This solves the problem of concentrated tangential tension of the surrounding skin caused by single-point sinking adjustment. The mechanical balance of the support surface is maintained through active load transfer, and effective decompression of high-risk areas and closed-loop optimization of support form are achieved. Attached Figure Description

[0020] Figure 1 A block diagram of the overall structure of a pressure-sensing membrane switch system provided in an embodiment of the present invention;

[0021] Figure 2 A system overall workflow diagram provided for embodiments of the present invention;

[0022] Figure 3 This is a comparison chart of experimental verification data from one embodiment of the present invention.

[0023] Among them, 10 is a flexible thin-film sensor array; 20 is a pressure characteristic parameter conversion module; 30 is a central control unit; 40 is a wireless communication module; 50 is a terminal display device; and 60 is an actuation and adjustment mechanism. Detailed Implementation

[0024] See attached document Figure 1 This invention provides a pressure-sensing membrane switch system configured to acquire pressure distribution data of a stressed surface and perform active morphological adjustment. The pressure-sensing membrane switch system mainly includes: a flexible thin-film sensor array 10, a pressure characteristic parameter conversion module 20, a central control unit 30, a wireless communication module 40, a terminal display device 50, and an adjustment mechanism 60.

[0025] The flexible thin-film sensor array 10, serving as the sensing front end of the system, employs a multi-layer composite structure. From top to bottom, the flexible thin-film sensor array 10 sequentially includes a panel layer, a circuit layer, and an adhesive backing layer, with each layer fixed together by adhesive. The circuit layer contains an array of conductive lines, which includes multiple sets of sensing nodes arranged in a row and column matrix. A pressure-sensing functional layer, made of carbon paste, silver paste, or a pressure-sensitive composite material, covers the side of the panel layer facing the circuit layer. A gap exists between the pressure-sensing functional layer and the array of conductive lines. When an external load is applied to the surface of the panel layer, the pressure-sensing functional layer comes into contact with the array of conductive lines, causing a change in the resistance or capacitance at that location.

[0026] The pressure characteristic parameter conversion module 20 is electrically connected to the flexible thin-film sensor array 10. The pressure characteristic parameter conversion module 20 is configured to acquire data from each sensing node in the flexible thin-film sensor array 10 via row scanning and column gating. The pressure characteristic parameter conversion module 20 integrates an analog front-end signal conditioning circuit and an analog-to-digital converter to convert detected resistance or capacitance changes into digitized raw pressure matrix data.

[0027] The central control unit 30 is connected to the pressure characteristic parameter conversion module 20 and the wireless communication module 40. The central control unit 30 is electrically connected to the actuator adjustment mechanism 60 via a drive circuit module. The central control unit 30 receives raw pressure matrix data from the pressure characteristic parameter conversion module 20 and is configured to perform tension disturbance decoupling, spatiotemporal fatigue calculation, and curvature following control operations. The wireless communication module 40 transmits the processed data to the terminal display device 50. The terminal display device 50 renders a three-dimensional pressure distribution image based on the received data.

[0028] The actuation and adjustment mechanism 60 is located below the flexible thin-film sensor array 10. The actuation and adjustment mechanism 60 includes multiple physical actuation units arranged in an array. The physical actuation units are selected from pneumatic adjustment components or electromechanical actuator components. The spatial distribution density of the physical actuation units is lower than the distribution density of the sensing nodes in the flexible thin-film sensor array 10. The central control unit 30 stores a logical partition mapping table, dividing the flexible thin-film sensor array 10 into several logical sensing areas, each logical sensing area corresponding to the control of one physical actuation unit.

[0029] See attached document Figure 2 This invention provides a control method for a pressure-sensing membrane switch system, comprising the following steps:

[0030] S1, the pressure characteristic parameter conversion module 20 scans and acquires the flexible thin film sensor array 10, obtains the changes in electrical characteristic parameters of the sensing nodes caused by pressure, and converts the changes in electrical characteristic parameters into digital voltage signals to generate original pressure matrix data;

[0031] S2, the central control unit 30 receives the original pressure matrix data output by S1, uses the Laplace operator to perform spatial convolution operation on the original pressure matrix data, calculates the spatial second derivative to identify tension interference areas and remove tension components, and generates real pressure matrix data.

[0032] S3, the central control unit 30 performs spatiotemporal integration calculation based on the real pressure matrix data output by S2, calculates the weighted integral value of each sensing node within the sliding time window, and generates a local cumulative fatigue index that reflects the risk of cumulative damage to the stressed surface.

[0033] S4, the central control unit 30 determines whether the local cumulative fatigue index output by S3 exceeds the preset safety threshold. When it exceeds the preset safety threshold, it generates an adjustment command according to the logical partition mapping relationship and drives the execution adjustment mechanism 60 to perform curvature following adjustment action through the drive circuit module.

[0034] The following will, in conjunction with the accompanying drawings and specific embodiments, provide a detailed explanation of the signal conversion principle of the Pressure Characteristic Parameter Acquisition (PCPA) involved in steps S1 to S4, the tension decoupling algorithm, the fatigue assessment model, and the dynamic load transfer strategy.

[0035] The specific implementation of the data acquisition and processing process (corresponding to the aforementioned step S1) performed by the pressure characteristic parameter conversion module 20 in conjunction with the flexible thin-film sensor array 10 is as follows:

[0036] S101, the pressure characteristic parameter conversion module 20 acquires the electrical characteristic parameters of the sensing nodes based on a physical addressing mechanism of row scanning and column gating. During system power-on operation, the pressure characteristic parameter conversion module 20 establishes an electrical connection with the array-type conductive lines of the flexible thin-film sensor array 10 through a multiplexer. In this connection architecture, the array-type conductive lines are configured as a grid topology with orthogonal row and column lines, and the physical intersections of the row and column lines are defined as sensing nodes. In response to the physical action of an external load acting on the panel layer, the pressure-sensing functional layer undergoes microscopic deformation at the sensing nodes. As a preferred embodiment, when using a piezoresistive material, the reduced spacing between conductive particles in the pressure-sensing functional layer leads to a non-linear decrease in contact resistance that is negatively correlated with the pressure value; when using a capacitor architecture, the reduced electrode spacing between the panel layer and the circuit layer leads to a positive increase in inter-electrode capacitance that is positively correlated with the pressure value. The pressure characteristic parameter conversion module 20 achieves a traversal of the physical state of the entire array of sensing nodes by sequentially applying excitation signals to the row lines and synchronously reading the electrical responses from the column lines.

[0037] S102 performs signal conditioning and analog-to-digital conversion to generate a digitized voltage signal. The built-in analog front-end circuit of the pressure characteristic parameter conversion module 20 processes the weak electrical signal read from the column lines. For changes in contact resistance, the analog front-end circuit uses a constant current source drive or a Wheatstone bridge circuit to convert resistance fluctuations into fluctuations in analog voltage amplitude; for changes in inter-electrode capacitance, the analog front-end circuit uses a capacitance-to-voltage conversion circuit or a charge amplifier to convert capacitance fluctuations into an analog voltage signal. This continuous analog voltage signal is then input to the analog-to-digital converter, and after sampling, holding, quantization, and encoding, the output is the sensing node. Measured digital voltage value The sensing node The measured digital voltage value directly maps to the sensing node. The numerical accuracy of the current stress state depends on the number of quantization bits of the analog-to-digital converter.

[0038] S103, construct and normalize the original pressure matrix data. Central control unit 30 receives data from sensing nodes. The measured digital voltage value is obtained based on the physical spatial coordinates of the sensing node. This data is reconstructed into a two-dimensional data structure. Given the discrete differences in substrate characteristics among different batches of flexible thin-film sensor arrays 10, and the slight fluctuations in the power supply voltage of the driving circuit, directly using measured values ​​would lead to inconsistent benchmarks for subsequent algorithms. Therefore, this embodiment employs a voltage-pressure linearization calibration strategy to eliminate zero-point drift and range errors at the hardware level, mapping the physical signal to a unified dimensionless numerical space. The central control unit 30 calculates the normalized pressure value of each node according to the following formula, generating the original pressure matrix data:

[0039] ;

[0040] in:

[0041] Indicates the first Line 1 The normalized pressure intensity of the column sensing nodes has a range of values. This value is used to characterize the degree of pure force after removing hardware differences;

[0042] This indicates the sensing node output in step S102. The measured digital voltage value;

[0043] Indicates sensing node The zero-point drift voltage value under no-load conditions, which is automatically calibrated by the system during the power-on initialization phase to eliminate the inherent zero drift of the sensor;

[0044] The effective range voltage span of the system is determined by both the reference voltage of the analog-to-digital converter and the sensitivity of the sensor.

[0045] This represents the temperature compensation coefficient, which is used to correct the effect of ambient temperature changes on the conductivity of semiconductor materials. Its value is derived from the readings of the system's built-in temperature sensor and is obtained by looking up a table.

[0046] This indicates the calculation of the robustness threshold, which takes the value of a very small positive number (e.g., The purpose of introducing this decision logic is to prevent overflow errors in the division operation when the effective range voltage span approaches zero due to sensor open circuit or power failure, thereby ensuring the robustness of the algorithm under hardware failure conditions.

[0047] Based on the above calculations, the generated original pressure matrix data for:

[0048] ;

[0049] in, This indicates the row number of the flexible thin-film sensor array 10. This indicates the column number of the flexible thin-film sensor array 10. This is the normalized and cleaned raw pressure matrix data. This will serve as a reliable input source for the tension decoupling process in the subsequent step S2.

[0050] The specific implementation of the spatial filtering and signal reconstruction process (corresponding to the aforementioned step S2) performed by the central control unit 30 in conjunction with the flexible thin-film sensor array 10 on the original pressure matrix data is as follows:

[0051] S201, Construct a physical model of the lateral tension interference of the flexible thin film and define the tension interference region. Given the physical continuity of the substrate of the flexible thin film sensor array 10, when a specific sensing node is subjected to a vertical load and undergoes downward deformation, this deformation will inevitably be transmitted to the surrounding areas through the elastic substrate, causing lateral stretching of adjacent non-stressed nodes. This lateral stretching alters the microscopic contact state of the pressure sensing functional layer at the non-stressed nodes, thereby introducing a false pressure signal, i.e., a "tension artifact," into the original pressure matrix data. Based on the principle of spatial frequency analysis, this embodiment utilizes the physical difference between the high-frequency abrupt changes at the actual stress point and the low-frequency smooth transition at the tension traction region, introducing a spatial second-order derivative operator to accurately separate the signal components into the vertical pressure principal component and the lateral tension interference component.

[0052] In step S202, a discrete Laplacian operator is used to perform spatial convolution to extract local curvature features. To quantify the degree of numerical abrupt change of each sensing node relative to its neighboring nodes, the central control unit 30 calls the discrete Laplacian convolution kernel stored in its internal register to process the raw pressure matrix data output in step S103. Perform sliding window convolution operation. Before execution, the system pads the matrix edges with zeros to ensure consistent dimensionality at boundary nodes. The central control unit 30 obtains the eigenvalues ​​of each node using the following formula:

[0053] ;

[0054] in:

[0055] Indicates sensing node The local curvature characteristic value reflects the degree of concavity and steepness of the pressure distribution surface at that node. Physically, a larger positive value represents the center of the pressure peak, while a negative value or a value close to zero represents a smooth transition zone under tension.

[0056] Indicates that in the sensing node Within the convolution window centered on the center, the relative coordinates are The neighboring node of Line 1 Normalized pressure intensity of column sensing nodes (Note: This refers to the physical quantity defined in S1, with only the coordinate index changing).

[0057] Represents the discrete Laplacian convolution kernel in relative coordinates Weighting coefficients at each location;

[0058] This represents the summation operator.

[0059] As a preferred implementation, in order to highlight the extreme value characteristics of the center node and suppress isotropic background tension, this embodiment selects a four-neighbor sharpening template with positive center weights, namely:

[0060] ;

[0061] S203 executes a tension artifact correction algorithm based on local curvature characteristics to generate realistic pressure matrix data. The central control unit 30 then uses the calculated sensing nodes... The local curvature feature values ​​are used to perform nonlinear correction on the original pressure matrix data. This correction logic follows the principle of "enhancing peak values ​​and suppressing sidelobes," using the curvature feature as a discriminator to dynamically remove the estimated lateral tension interference component from the original signal, thereby restoring the true contact pressure distribution. The central control unit 30 calculates the true pressure value using the following formula:

[0062] ;

[0063] in:

[0064] Indicates the first Line 1 The actual pressure values ​​of the sensing nodes, which are the effective pressure data after removing tension interference, constitute the final true pressure matrix data. ;

[0065] Indicates the first Line 1 Normalized pressure intensity of the column sensing nodes (i.e., the raw data output in step S1).

[0066] This represents the tension suppression coefficient, which ranges from [0.5, 2.0]. The selection of this coefficient is based on the elastic modulus of the material of the flexible thin-film sensor array 10: the softer the substrate, the stronger the tension transmission effect, and a larger tension suppression coefficient is required to enhance the decoupling effect;

[0067] This indicates the sensing node calculated in step S202. The local curvature characteristic value;

[0068] The non-negative rectification and limiting function is mathematically defined as follows: The physical purpose of introducing this function is twofold: firstly, the corrected pressure value cannot be physically negative, and it is forced to be zero when the calculation result is less than zero; secondly, the upper limit of the normalized pressure value is 1 to prevent numerical overflow caused by Laplace sharpening overshoot, ensuring that the data is always within the valid physical domain of [0,1].

[0069] After the above steps, the central control unit 30 will contain all True pressure matrix data of elements The data is output to subsequent modules as a precise data basis for fatigue assessment and morphological control.

[0070] The specific implementation method of the central control unit 30, in coordination with the wireless communication module 40 and the terminal display device 50, performing time-varying cumulative damage assessment based on the real pressure matrix data output in step S2 (corresponding to the aforementioned step S3) is as follows:

[0071] S301, a high-resolution 3D pressure topography map is constructed based on a bicubic interpolation algorithm. To address the limited monitoring resolution caused by the discrete distribution of the physical sensing nodes in the flexible thin-film sensor array 10, and to provide medical personnel with a continuous and smooth pressure distribution view, the central control unit 30 processes the actual pressure matrix data output in step S2. Perform spatial reconstruction. In this embodiment, the system uses a bicubic interpolation algorithm to select the surrounding data for any point to be interpolated. 16th in the neighborhood Line 1 The actual pressure values ​​of the sensing nodes are used as a benchmark for weighted calculation.

[0072] The algorithm's technical advantage lies in simultaneously constraining the continuity of function values, first derivatives, and second derivatives at the grid boundaries, thereby constructing a continuous pressure surface that approximates the stress state of real skin. The wireless communication module 40 transmits the interpolated high-resolution 3D data stream to the terminal display device 50 in real time, rendering a 3D pressure terrain map that dynamically reflects the peak gradient and morphological characteristics of the stress.

[0073] S302, construct a spatiotemporal integral accumulation model within a sliding time window. The pathological evolution of deep tissue injuries such as pressure ulcers depends not only on the magnitude of instantaneous pressure but also on the duration of pressure application (i.e., the "pressure-time" integral effect). Therefore, the central control unit 30 constructs a circular buffer in system memory, establishing a length of... The system employs a sliding time window. It synchronously records the real pressure matrix data at a fixed sampling frequency for each historical moment and performs strict timestamp alignment, forming a three-dimensional data tensor containing spatial coordinates and a temporal dimension. This model aims to quantify the cumulative ischemic effect of soft tissue under continuous load by performing discrete integral calculations on the stress history of each sensing node along the time axis.

[0074] S303 executes a nonlinear weighted strategy and fatigue calculation based on the microcirculation occlusion threshold. According to biomechanical and microcirculation physiology principles, when the externally applied vertical pressure is lower than the capillary perfusion pressure, the tissue can maintain oxygen supply through its own vasodilation regulation; however, when the pressure exceeds a certain critical value, vascular occlusion leads to severe tissue ischemia and hypoxia, at which point the rate of damage accumulation will increase nonlinearly and rapidly. To accurately simulate this physiological process, the central control unit 30 introduces a microcirculation occlusion pressure threshold as an inflection point, applying differentiated damage weights to pressure values ​​in different ranges.

[0075] The central control unit 30 calculates the cumulative fatigue index of each sensing node using the following formula:

[0076] ;

[0077] in:

[0078] Indicates sensing node The local cumulative fatigue index is a quantitative indicator with clear physical meaning, which directly reflects the degree of ischemic damage risk to the skin and subcutaneous tissue at that location due to continuous pressure.

[0079] This represents the total number of sampling points within the sliding time window, and its value is obtained by dividing the preset time window length (e.g., 30 minutes) by the system sampling period;

[0080] Indicates the first time within the sliding time window The sampling time of the first sampling moment Line 1 The actual pressure value of the sensor node (Note: This value is derived from the output of step S2). In the A snapshot of a historical moment, and this value has been mapped to the physical pressure space through normalization).

[0081] This represents the microcirculation occlusion pressure threshold, which is set based on the statistical characteristics of the average perfusion pressure of human capillaries (corresponding to a specific threshold in the normalization space, such as 0.3 to 0.45), and serves as a non-zero constant denominator to ensure the stability of the division operation.

[0082] This represents the unit step function, defined as taking a value of 1 when the input variable is greater than or equal to 0, and a value of 0 otherwise. The physical purpose of introducing this function is to establish a "damage threshold," ensuring that additional nonlinear weighted penalties are triggered only when the pressure exceeds the occlusion threshold.

[0083] This represents the ischemic injury acceleration factor, with a value range of [value missing]. This coefficient is used to adjust the sensitivity of fatigue accumulation under high pressure. The larger the coefficient, the more sensitive the tissue is to overpressure.

[0084] It represents a nonlinear damage index, usually with a value of 2 or 3, which characterizes the power-law growth of tissue necrosis risk with the magnitude of overpressure.

[0085] This represents the system sampling time interval, used to convert the pressure intensity values ​​of a discrete sequence into integrals over the continuous time domain.

[0086] S304, output the local cumulative fatigue index and execute the trigger judgment. The central control unit 30 updates the local cumulative fatigue index of all sensing nodes in the entire array in real time. The calculation result is presented on the terminal display device 50 in the form of a color-coded heat map through a visualization interface (for example, high-risk areas are rendered as a dark red warning color), and serves as a decision input variable for the closed-loop control system. When the local cumulative fatigue index of any one or more adjacent nodes exceeds the system's preset safety threshold, the central control unit 30 determines that the current stress state has a risk of pressure sores and then triggers the active adjustment mechanism. The safety threshold is the tissue tolerance limit value calibrated based on clinical trial data. This judgment logic ensures the timeliness of system intervention and provides clear data support for the subsequent step S4 to generate adjustment instructions based on the logical partition mapping relationship and drive the execution adjustment mechanism 60.

[0087] The specific implementation of the active morphological reconstruction (corresponding to the aforementioned step S4) performed by the central control unit 30 in coordination with the adjustment mechanism 60 through the drive circuit module, based on the local cumulative fatigue index output in step S3, is as follows:

[0088] S401, establish the logical partition mapping relationship between the sensor array and the physical execution units. Given that the spatial density of the sensing nodes in the flexible thin-film sensor array 10 is much higher than the distribution density of the physical execution units in the execution adjustment mechanism 60, the system cannot directly adopt a "point-to-point" control mode. To achieve spatial alignment between the microscopic sensing layer data and the macroscopic execution layer actions, the central control unit 30 calls the pre-stored logical partition mapping table during the initialization phase. This mapping mechanism is based on the principle of spatial geometric projection, dividing the high-resolution sensing plane into... A rectangular logic sensing area (of which) Each logical sensing region uniquely corresponds to a physical execution unit at its geometric center. In terms of data structure, each logical sensing region contains a specific set of sensing nodes, denoted as... ,in The array coordinate index representing the physical execution unit ( Through this mapping mechanism, the system reduces the dimensionality of the fine pressure distribution data collected by the flexible thin-film sensor array 10 to the mechanical control space, ensuring that subsequent control commands have precise spatial orientation.

[0089] S402, calculate the regional aggregate fatigue degree and aggregate curvature modulus. To determine whether a specific physical execution unit needs intervention and adjustment, the central control unit 30 needs to extract features of the state within each logical sensing area. Considering that human tissue damage often begins with single-point high pressure (i.e., the shortest plank in the "barrel effect"), if only the regional average value is used for evaluation, it is easy to smooth out local extreme value risks, leading to misjudgment. Therefore, this embodiment adopts an "extreme value priority" strategy to calculate fatigue degree features, while retaining the average feature of curvature to maintain deformation smoothness.

[0090] The central control unit 30 extracts the control variables using the following formula:

[0091] ;

[0092] in:

[0093] Indicates the first Line 1 The region aggregate fatigue value corresponding to the physical execution unit is taken as the maximum value of the local cumulative fatigue index among all sensing nodes in the region, so as to ensure that the state of the most dangerous point is captured first and used as the direct criterion for triggering adjustment.

[0094] This indicates the sensing node calculated in step S303. Local cumulative fatigue index;

[0095] This indicates the maximum value selection operation;

[0096] Indicates the first Line 1 The region aggregate curvature modulus corresponding to the physical execution unit reflects the current overall concavity and convexity of the region (positive value indicates convexity, negative value indicates concavity), and serves as a reference for subsequent adjustment actions;

[0097] Indicates the first Line 1 The total number of sensing nodes contained in the column logic sensing area. This value is an integer that is always greater than zero, ensuring the validity of the division operation.

[0098] This indicates the sensing node calculated in step S202. The local curvature characteristic value.

[0099] S403 executes a curvature-following adjustment strategy based on a neighborhood compensation mechanism. When the regional aggregate fatigue of any physical execution unit exceeds a preset safety threshold, it indicates that the supporting surface corresponding to that region has caused the human tissue to approach its damage limit. At this time, if only the unit is controlled to perform a simple sinking avoidance action, although it can alleviate local vertical pressure, it will cause the human body to sink deeper, which will intensify the tangential tension of the surrounding skin, forming a "hammock effect." To resolve this contradiction, the central control unit 30 executes a "central unloading-neighborhood compensation" collaborative control strategy. The core logic of this strategy follows the principles of "local volume conservation" and "stress redistribution": while driving the affected area to sink and unload, it forces the adjacent physical execution units around it to rise moderately to actively take over the transferred load. This mechanism not only eliminates high-pressure points but also prevents shear force concentration caused by single-point collapse, maintaining the continuity of the supporting surface.

[0100] The central control unit 30 generates action commands for each execution unit using the following formula:

[0101] ;

[0102] in:

[0103] Indicates the first Line 1 The physical execution unit sends an adjustment stroke command, the unit of which is millimeters (mm). A positive value represents upward lifting and a negative value represents downward retraction. The central control unit 30 converts the command into a drive pulse and sends it to the drive circuit module.

[0104] This represents the active unloading gain coefficient, with a value range of [value missing]. mm / unit is used to define the sinking depth corresponding to a unit fatigue deviation. The larger the coefficient, the more drastic the adjustment response.

[0105] Indicates the first Line 1 The fatigue deviation value of the physical execution unit is calculated using the following formula: .in As a safety threshold, This is the aforementioned non-negative rectification function. This term is only non-zero when the fatigue level exceeds the limit, representing the active sinking requirement of this unit;

[0106] Indicates the first Line 1 The effective neighborhood set of a physical execution unit (containing 3, 5, or 8 adjacent units, depending on whether the unit is located at a corner, edge, or center).

[0107] Representing neighborhood units The fatigue deviation value is positive when the fatigue value of neighboring elements exceeds the limit, and it drives the current element through the summation term. To rise in order to compensate;

[0108] Representing neighborhood units The effective neighborhood number is used as the normalization denominator to ensure that neighborhood units The total load that needs to be unloaded is evenly distributed to the surrounding auxiliary units, thereby maintaining the overall mechanical balance of the support surface.

[0109] S404, full-field curvature variance feedback and closed-loop convergence verification. The central control unit 30 sends data in parallel through the drive circuit module, containing all... The control signal of the command drives the action of the adjustment mechanism 60. After the action is completed, the system control pressure characteristic parameter conversion module 20 re-acquires a frame of original pressure matrix data. Steps S2 to S402 are repeated to calculate the new full-field curvature distribution and fatigue state. The central control unit 30 calculates the full-field curvature variance. If the variance value shows a decreasing trend and the regional aggregate fatigue of all areas falls below the safety threshold, the dynamic load transfer control is determined to have converged, and the system enters steady-state monitoring mode; otherwise, the system will start the next round of fine-tuning iteration based on the updated data. This closed-loop feedback mechanism ensures a smooth and stable deformation adjustment process, avoiding system oscillations or secondary mechanical damage to the patient caused by over-adjustment.

[0110] Specific application examples:

[0111] This embodiment constructs a smart nursing mattress based on a pressure-sensing membrane switch system. The flexible membrane sensor array 10 has 64 rows. Column (i.e.) The physical dimensions cover the core pressure zones of the mattress. The adjustment mechanism consists of 60... Composed of a pneumatic adjustment component (i.e.) According to the logical partition mapping table, each pneumatic control component corresponds to the control of 4. The area with 4 sensing nodes.

[0112] The following is a single closed-loop adjustment process for the sacral and coccygeal pressure area of ​​a long-term bedridden patient:

[0113] S1, the pressure characteristic parameter conversion module 20 performs a scan on the flexible thin-film sensor array 10. Assuming at time... The sensing nodes (32,16) located in the central region of the sacrococcygeal region and their adjacent nodes (32,17) on the right were collected.

[0114] For the center node (32,16), the analog-to-digital converter outputs the measured digital voltage value. Given the zero-point drift voltage Effective range voltage span Temperature compensation coefficient Substituting into the normalization formula:

[0115] ;

[0116] Meanwhile, due to the lateral tension of the flexible film, adjacent nodes (32,17) that are not directly compressed generate false signals, which are measured to be... At this point, there is a noticeable "tension artifact" in the original pressure matrix data.

[0117] S2, the central control unit 30 receives data and performs tension disturbance decoupling. The tension suppression coefficient is set. For the spurious signal node (32,17) affected by tension interference, the original pressure values ​​of its four neighboring areas (top, bottom, left, and right) are as follows: the peak value on the left is 0.818, and the other three are 0 (assuming no background pressure). Using a Laplacian convolution kernel... Calculate the local curvature eigenvalues:

[0118] ;

[0119] Substitute the values ​​collected in S1:

[0120] ;

[0121] Since the node is located in the tension zone next to the center of force, its value is lower than the neighborhood peak, resulting in a negative second derivative.

[0122] Based on the actual pressure calculation formula:

[0123] ;

[0124] via nonnegative rectifier function After processing:

[0125] ;

[0126] The calculation results show that by introducing spatial curvature features, the system successfully identified and eliminated spurious signals caused by tension (i.e., corrected from the original 0.150 to 0), achieving precise signal decoupling.

[0127] In the algorithm logic of this embodiment, a more typical case is to utilize the center point. Enhancement. For the central node (32,16), assume that there is a certain degree of tension disturbance around it (average value 0.150):

[0128] ;

[0129] Calculate the enhanced actual pressure value:

[0130] ;

[0131] S3 performs spacetime integration. A sliding time window is set. min, microcirculation occlusion pressure threshold .

[0132] To demonstrate a typical working condition for the fatigue accumulation process, the stable pressure value output by S2 is used for simulation. It is assumed that node (32,16), after position adjustment, has maintained a typical high-pressure state for the past 30 minutes. The value is kept constant at 0.8 (after normalization). The damage accumulation term at a single sampling time is calculated (assuming...). , ): Weighting Item ;

[0133] If sampling interval The local cumulative fatigue index after 30 minutes At this point, the value far exceeds the preset safety threshold (e.g., ).

[0134] S4, generate adjustment instructions. This corresponds to the physical execution unit. Regional aggregate fatigue Exceeding the limit deviation value Set the active unloading gain coefficient. Central Unit Action instructions:

[0135] (sink);

[0136] Its neighborhood units (e.g.) Assuming no fatigue risk and having 8 valid neighborhoods, based on the neighborhood compensation mechanism:

[0137] (rise);

[0138] Ultimately, the center point of the actuator 60 sinks by 11mm, while the surrounding units rise by approximately 1.4mm, forming a smooth stress transfer groove.

[0139] Experimental verification and effect comparison:

[0140] To verify the technical effectiveness of this invention, a system simulation model was established using MATLAB software, and a comparative experiment was conducted with existing technologies (traditional airbag adjustable mattresses without tension decoupling and neighborhood compensation). The experimental test object was set as a simulated human hip force model, with a constant load applied for 60 minutes. The control group (existing technology) was configured to perform threshold judgment based solely on the original resistance value, and only single-point sinking was performed when adjustment was triggered; the experimental group (this invention) was configured to enable the Laplace tension decoupling algorithm and neighborhood compensation control strategy.

[0141] in conclusion:

[0142] See attached document Figure 3 , Figure 3(a) illustrates the comparison of pressure distribution signals. Existing technologies obtain pressure distribution curves that are distorted by amplitude attenuation and spatial dispersion due to interference from the lateral tension of the flexible substrate (as shown by the dotted line). This signal crosstalk leads to ambiguity in locating high-pressure risk points. In contrast, this invention introduces a tension decoupling algorithm based on the Laplace operator, generating a true pressure distribution curve (as shown by the dashed line) that closely matches the actual physical pressure value (as shown by the solid line). Data shows that the decoupling process significantly narrows the half-width at half-maximum (WHM) of the stressed area, improving the effective spatial resolution of the pressure peak by approximately 20% and effectively suppressing the interference of lateral mechanical crosstalk on the data.

[0143] Reference Figure 3 (b) The spatiotemporal integral fatigue assessment model proposed in this invention can accurately reflect the cumulative process of tissue damage. Unlike traditional methods that rely solely on instantaneous pressure thresholds, the local cumulative fatigue index of this system increases non-linearly over time. The data point marked as data1 (data recording point 1) in the figure indicates the critical moment when the system triggers regulation: when subjected to continuous pressure for approximately 21 minutes (calculated value: 50.0 / 2.4≈20.8 minutes), the cumulative fatigue index intersects with the safety threshold (value 50.0), and the system immediately triggers the regulation command. This mechanism conforms to the pathological characteristic of accelerated accumulation of soft tissue damage risk over time under ischemic conditions, avoiding malfunctions caused by short-term pressure fluctuations and preventing the formation of latent cumulative damage under long-term low-load conditions.

[0144] Reference Figure 3 (c) During morphological adjustment, the support surface morphology differs significantly between the two technical approaches. The existing single-point avoidance strategy (shown by the dashed line diagram with square markings), while releasing central pressure, creates a steep step between the 8th unit and the adjacent 7th and 9th units, leading to local shear force concentration. In contrast, the neighborhood compensation mechanism employed in this invention (shown by the solid line diagram with circular markings) simultaneously drives the 8th unit to sink and unload (stroke -11.0 mm) while simultaneously controlling the adjacent 7th and 9th units to perform a slight upward movement (precise value of 1.375 mm). This coordinated action constructs a smoothly transitioning stress redistribution gradient on the support surface, ensuring effective removal of central high pressure while maintaining the overall mechanical continuity of the support surface, preventing the accumulation of tangential tension due to local collapse.

[0145] In summary, this invention solves the technical problems of low accuracy, lack of time dimension assessment, and single adjustment method of traditional flexible thin-film switch systems by using three core technologies: tension interference decoupling, cumulative damage quantitative assessment, and neighborhood compensation collaborative control. It achieves high-fidelity monitoring and closed-loop adjustment of the human body's stress state.

Claims

1. A pressure-sensing membrane switch system, characterized in that, include: A flexible thin-film sensor array (10) serves as a sensing front end and a switch triggering medium, used to respond to changes in electrical characteristic parameters caused by external loads. The pressure characteristic parameter conversion module (20) is electrically connected to the flexible thin film sensor array (10) and is used to collect the changes in the electrical characteristic parameters and convert them into digital raw pressure matrix data; The central control unit (30) is used to receive the original pressure matrix data, perform tension disturbance decoupling on the original pressure matrix data, generate real pressure matrix data, calculate the local cumulative fatigue index based on the real pressure matrix data, and generate adjustment instructions based on the local cumulative fatigue index. An adjustment mechanism (60) is disposed below the flexible thin-film sensor array (10) and is used to perform active shape adjustment in response to the adjustment command; The wireless communication module (40) is used to send the real pressure matrix data and the local cumulative fatigue index to the terminal display device (50).

2. The pressure-sensing membrane switch system according to claim 1, characterized in that, The flexible thin-film sensor array (10) has a structure for responding to changes in electrical characteristic parameters caused by external loads, comprising: The flexible thin-film sensor array (10) adopts a multi-layer composite structure, which is provided with a panel layer, a circuit sheet layer and a backing film layer from top to bottom. The circuit sheet is printed with an array of conductive lines, which includes multiple sets of sensing nodes arranged in a row and column matrix. The panel surface is covered with a pressure-sensing functional layer on one side facing the circuit sheet layer, and the pressure-sensing functional layer is suspended above the array of conductive lines. When an external load is applied to the panel layer, it drives the pressure sensing functional layer to make physical contact with the array of conductive lines, thereby generating a change in the electrical characteristic parameters at the contact location.

3. A pressure-sensing-based membrane switch system according to claim 1, characterized in that, When the central control unit (30) receives the original pressure matrix data and performs tension disturbance decoupling on the original pressure matrix data to generate real pressure matrix data, it performs the following steps: The original pressure matrix data is received, and a sliding window convolution operation is performed on the original pressure matrix data using a discrete Laplacian convolution kernel to extract the local curvature features at each node position. The local curvature features are weighted and calculated with the original pressure matrix data. The transverse tension interference component is removed from the original pressure matrix data, and the true pressure matrix data is output.

4. A pressure-sensing membrane switch system according to claim 3, characterized in that, When calculating the local cumulative fatigue index based on the real pressure matrix data, the central control unit (30) performs the following steps: A sliding time window is constructed to map the real pressure matrix data under continuously collected time-varying sequences into the sliding time window; By introducing the ischemic injury acceleration coefficient and the microcirculation occlusion pressure threshold, spatiotemporal integration is performed on the real pressure matrix data within the sliding time window to generate a local cumulative fatigue index. The local cumulative fatigue index is output, which is used to characterize the risk of ischemic injury caused by continuous pressure at the sensing node location. The microcirculation occlusion pressure threshold refers to the critical pressure value that causes the blockage of blood flow in the human skin microcirculation.

5. A pressure-sensing membrane switch system according to claim 4, characterized in that, The system also includes a user feedback interface disposed on the terminal display device (50); When calculating the local cumulative fatigue index based on the real pressure matrix data, the central control unit (30) performs the following steps: The user feedback interface receives subjective evaluation instructions from the user. The subjective evaluation command is input into the parameter adaptation module based on machine learning, and the ischemic injury acceleration coefficient or the microcirculation occlusion pressure threshold used to calculate the local cumulative fatigue index is dynamically fine-tuned.

6. A pressure-sensing membrane switch system according to claim 1, characterized in that, The execution adjustment mechanism (60) includes a plurality of physical execution units arranged in an array; When generating adjustment commands based on the local cumulative fatigue index, the central control unit (30) performs the following steps to construct the spatial basis for generating the adjustment commands: Call the logical partition mapping table to divide the flexible thin film sensor array (10) into several logical sensing areas; Establish a mapping relationship so that each of the logical sensing areas uniquely corresponds to and controls the corresponding physical execution unit at the geometric center; Based on the mapping relationship, the coordinates of the sensing nodes of the flexible thin-film sensor array (10) are aligned with the spatial position of the physical execution unit, thereby constructing the spatial basis for generating the adjustment command.

7. A pressure-sensing membrane switch system according to claim 6, characterized in that, When generating adjustment commands based on the local cumulative fatigue index, the central control unit (30) performs the following steps: Obtain the multiple local cumulative fatigue indices contained within each of the said logical sensing regions; Based on the extreme value priority strategy, the maximum value is extracted from multiple local cumulative fatigue indices to generate the regional aggregate fatigue degree corresponding to the logical sensing area.

8. A pressure-sensing membrane switch system according to claim 7, characterized in that, When generating adjustment commands based on the local cumulative fatigue index, the central control unit (30) performs the following steps: Determine whether the aggregated fatigue degree of the region corresponding to the current physical execution unit, obtained based on the local cumulative fatigue index, exceeds a preset safety threshold. When the fatigue level of the region exceeds the preset safety threshold, the adjustment command is generated based on the neighborhood compensation mechanism. The adjustment command is used to drive the current physical execution unit to sink and drive the physical execution units located in the neighborhood of the current physical execution unit to rise; When the fatigue level of the region does not exceed the preset safety threshold, the morphological parameters of the current physical execution unit remain unchanged, and the adjustment command is not generated. The preset safety threshold refers to the maximum cumulative pressure value that can cause deep tissue damage.

9. A pressure-sensing-based membrane switch system according to claim 1, characterized in that, The wireless communication module (40) is selected from low-power Bluetooth, Wi-Fi or ZigBee communication modules; When the terminal display device (50) receives the real pressure matrix data and performs the display, it performs the following steps: Receive the actual pressure matrix data; The real pressure matrix data is spatially reconstructed using a bicubic interpolation algorithm to generate a high-resolution three-dimensional data stream. A three-dimensional pressure distribution image is rendered based on the high-resolution three-dimensional data stream.

10. A pressure-sensing membrane switch system according to claim 1, characterized in that, When the pressure characteristic parameter conversion module (20) collects the changes in the electrical characteristic parameters and converts them into digitized raw pressure matrix data, it performs the following steps: The changes in electrical characteristic parameters at the sensing nodes in the flexible thin-film sensor array (10) are collected through a physical addressing mechanism of row scanning and column gating, and the digitized measured digital voltage value is obtained. The measured digital voltage value is processed using a voltage-pressure linearization calibration strategy to eliminate zero drift and range error, and the original pressure matrix data is output.

Images