A wearable gout early warning sock and early warning system thereof

Abstract

The application discloses a wearable gout early warning sock and an early warning system thereof, and relates to the technical field of gout. The early warning sock is made of elastic and breathable fabric, and is locally reinforced in the first metatarsophalangeal joint area. A plurality of arc-shaped temperature and pressure integrated micro sensor assemblies are embedded in the early warning sock. The assemblies are packaged with skin-friendly and heat-conducting silica gel, and are detachably connected through magnetic attraction. A passive micro environment control structure is arranged on the periphery of the early warning sock. A data processing unit in the sock integrates a multi-modal deep space-time fusion early warning algorithm, a personalized risk trajectory prediction and uncertainty quantification algorithm, and a causal relationship reasoning personalized cause identification algorithm. The processed data are sent to external intelligent devices through a wireless communication module, so that the life quality of gout patients is improved and the attack risk is effectively reduced.

Description

Technical Field

[0001] This invention relates to the field of gout technology, and more specifically to a wearable gout warning sock and its warning system. Background Technology

[0002] Gout is a common metabolic disease characterized by elevated blood uric acid levels, leading to the deposition of urate crystals in the joints and triggering acute arthritis attacks, often manifesting as severe pain, redness, swelling, and fever. Traditionally, the diagnosis and management of gout have relied primarily on blood uric acid testing, symptom assessment, and imaging studies. However, these methods often fail to provide real-time, continuous physiological data and cannot effectively warn of gout attacks.

[0003] Existing gout monitoring methods have several limitations. First, blood uric acid testing requires a hospital visit, has a limited frequency, and cannot reflect the dynamic fluctuations in uric acid levels in daily life or its correlation with lifestyle habits. Second, patients often seek medical help only after symptoms appear, by which time a gout attack has already occurred, missing the optimal intervention window. Third, although some studies have attempted to use wearable devices for health monitoring, there are few solutions specifically designed for high-precision, personalized monitoring of local physiological changes that precede gout attacks. For example, some general-purpose temperature or activity monitoring devices may not be able to capture the microenvironmental changes specific to the first metatarsophalangeal joint region before a gout attack, such as subtle abnormalities in local temperature and pressure. Furthermore, existing warning systems are often based on simple threshold judgments, lacking consideration for individual differences and failing to deeply analyze the personalized triggers that lead to gout attacks, resulting in insufficient accuracy and personalized guidance in warnings.

[0004] Currently, the market lacks a wearable device that integrates high-precision sensors, intelligent algorithms, and a user-friendly interface to provide early, continuous, and personalized warnings of gout risk and offer targeted intervention suggestions. Existing technologies still have significant room for improvement in terms of comfort, data accuracy, algorithm intelligence, and trigger identification. In particular, the microenvironmental control of the sensor-skin interface, the deep fusion analysis of multimodal data, and personalized trigger identification based on causal reasoning remain challenges for current technology.

[0005] Therefore, it is necessary to propose a wearable gout warning sock and its warning system to solve the above problems. Summary of the Invention

[0006] The purpose of this invention is to solve the problems of poor real-time performance, lack of personalization, difficulty in identifying triggers, and the need to improve the accuracy of early warning in existing gout monitoring and early warning methods.

[0007] To achieve the above objectives, the present invention specifically adopts the following technical solution:

[0008] A wearable gout warning sock and its warning system, comprising a sock body made of elastic and breathable fabric material, wherein the sock body is locally reinforced in the first metatarsophalangeal joint area;

[0009] Multiple sensor components are integrated inside the body of the sock. The sensor components are strategically deployed in the medial region of the first metatarsophalangeal joint and arranged in an arc shape, distributed along the natural anatomical contour of the joint capsule. The sensor components include miniature temperature sensors and miniature pressure sensors. An active microenvironment control structure is provided around the sensor components.

[0010] A data processing unit is used to receive and process temperature and pressure data from the sensor assembly, and internally integrates a multimodal deep spatiotemporal fusion early warning algorithm, a personalized risk trajectory prediction and uncertainty quantification algorithm, and a personalized cause identification algorithm for causal reasoning.

[0011] A wireless communication module is used to send the data processed by the data processing unit to an external smart device.

[0012] Furthermore, the sensor assembly is packaged using a dual-layer structure, the dual-layer structure comprising:

[0013] A skin-friendly, thermally conductive silicone layer, as the inner layer, with a thickness of 0.3 mm to 0.5 mm, contains the micro temperature sensor probe and the micro pressure sensor probe, and is in direct contact with the user's skin;

[0014] An elastic fabric base layer serves as the outer layer, embedding a flexible circuit. The skin-friendly, thermally conductive silicone layer is connected to the elastic fabric base layer via detachable magnetic contacts, allowing the sensor assembly to be detached from the sock body.

[0015] Furthermore, the temperature sensor and the pressure sensor in the sensor assembly are designed as an integrated temperature and pressure sensing unit, integrated in the same micro package. The micro pressure sensor adopts a piezoresistive thin film structure with a thermally conductive layer covering it. The micro temperature sensor is attached to the underside of the thermally conductive layer. The integrated temperature and pressure sensing unit shares a set of micro signal conditioning circuits.

[0016] Furthermore, the active microenvironment control structure includes:

[0017] A microporous moisture-absorbing layer, woven from highly absorbent fibers, with a thickness of 0.5 mm to 1.0 mm and a porosity of 60% to 80%, is used to absorb localized sweat and keep the interface between the sensor assembly and the skin dry;

[0018] A hydrophobic barrier membrane is disposed between the microporous moisture-absorbing layer and the sensor assembly to prevent sweat absorbed in the microporous moisture-absorbing layer from back-permeating to the surface of the sensor assembly.

[0019] Furthermore, the specific execution steps of the multimodal deep spatiotemporal fusion early warning algorithm include:

[0020] Continuously collected temperature sequences and pressure sequence Constructed as a multidimensional input matrix, where The length of the sampling sliding window;

[0021] The spatial features of the multidimensional input matrix are extracted using a one-dimensional convolutional kernel of a convolutional neural network to generate a high-dimensional feature vector mapping.

[0022] The high-dimensional feature vector is mapped and input into a long short-term memory network. Its forget gate, input gate, and output gate structures are used to extract temporal dependencies and generate a sequence of hidden states. An attention mechanism is introduced to perform weighted fusion of the hidden state sequences and calculate the context vector. :

[0023]

[0024]

[0025]

[0026] in, To score attention, For weight coefficients; use the context vector The system classifies gout attacks using a fully connected layer and outputs a warning score for the current attack.

[0027] Furthermore, the specific execution steps of the personalized risk trajectory prediction and uncertainty quantification algorithm include:

[0028] A risk prediction model based on Bayesian deep learning is constructed, and a Monte Carlo Dropout mechanism is introduced in the inference stage; for a given input sample, the Dropout layer is kept on, and... After several random forward propagation samplings, the set of prediction results is obtained. ;

[0029] Calculate the expected value of the prediction results As a risk prediction trajectory within a future time window:

[0030]

[0031] Calculate the variance of the prediction results As a quantitative indicator of model uncertainty:

[0032]

[0033] when Exceeding the preset risk threshold and When the value falls below the preset threshold, a warning command is output.

[0034] Furthermore, the specific execution steps of the personalized incentive identification algorithm for causal reasoning include:

[0035] Define users' external behavioral characteristics as intervention variables. Changes in foot physiological signs were used as the outcome variable. User basal metabolic rate is a confounding variable. ;

[0036] Structural causal models are used to describe the causal relationships between variables, and counterfactual inference frameworks are used to estimate the individual intervention effects. :

[0037]

[0038] in, Operators represent active intervention in behavior; when targeting a specific behavior... Calculated If the behavior exceeds the preset significance threshold three times consecutively, it will be identified as a personalized sensitive trigger for the user.

[0039] A warning system for wearable gout warning socks includes the aforementioned wearable gout warning socks and a smart application running on the external smart device, the smart application having:

[0040] A structured diet input interface and a built-in purine load calculator are used to record the user's diet and calculate the purine load values ​​for a single meal and a day;

[0041] The diet-physiology correlation analysis module uses a personalized trigger identification algorithm based on causal reasoning to automatically correlate dietary intake with subsequent changes in physiological parameters and identify personalized triggers that increase the risk of gout.

[0042] The personalized timing optimization function for medication reminders calculates the peak time of purine load based on individual metabolic characteristics and dietary records, and sets medication reminders 1 to 2 hours before the peak.

[0043] The multi-level early warning mechanism includes Level 1, Level 2, and Level 3 early warnings, each corresponding to different risk levels and intervention recommendations.

[0044] Furthermore, the intelligent application also has a doctor-assisted diagnostic data package generation function, which automatically generates a weekly health report that includes a temperature change trend graph, a causal graph of dietary records and temperature changes, a timeline of early warning events, and curves of changes in multimodal physiological characteristics.

[0045] Furthermore, the smart application also has a shoe-changing reminder function. When the overall foot temperature rises by more than 1.5 times the baseline standard deviation, and the pressure sensor detects multiple local pressure points that persist for more than 30 minutes, and the system is not in a level 2 or 3 gout warning state, it will issue a reminder to change to loose and breathable shoes.

[0046] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0047] 1. The present invention, by deploying miniature temperature and pressure sensors in the first metatarsophalangeal joint region and arranging them in an arc, can accurately capture local microenvironmental changes before a gout attack and provide continuous, real-time physiological data.

[0048] 2. The data processing unit of this invention integrates a multimodal deep spatiotemporal fusion early warning algorithm, a personalized risk trajectory prediction algorithm, and an uncertainty quantification algorithm. It can perform in-depth analysis of multimodal data such as temperature and pressure, improve the accuracy and reliability of early warning, and quantify the uncertainty of prediction. Attached Figure Description

[0049] Figure 1 This is a schematic diagram showing a partial structure of the main body of the sock in the first metatarsophalangeal joint region of the present invention.

[0050] Figure 2 This is a side view of the unfolded partial structure of the main body of the sock in the first metatarsophalangeal joint region of the present invention.

[0051] Reference numerals: 1. Sensor assembly; 2. Skin-friendly thermally conductive silicone layer; 3. Elastic fabric base layer; 4. Microporous moisture-absorbing layer; 5. Hydrophobic barrier membrane. Detailed Implementation

[0052] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0053] This invention discloses a wearable gout warning sock and its warning system. The system integrates high-precision physiological sensors, data processing algorithms, and an intelligent user interface to achieve early, continuous, and personalized warning and management of gout attack risk.

[0054] Combination Figure 1 and Figure 2 As shown, the wearable gout warning socks of the present invention mainly include a sock body, a sensor assembly 1, a data processing unit, and a wireless communication module.

[0055] The main body of the sock is made of a high-performance elastic and breathable fabric, which can be blended with elastic fibers such as spandex and Lycra, and natural fibers such as cotton and modal, or synthetic fibers such as polyester and nylon. This ensures that the sock has good breathability, moisture-wicking properties, and a comfortable wearing experience, while maintaining sufficient elasticity to adapt to different foot shapes. The main body of the sock is locally reinforced in the first metatarsophalangeal joint area. This is achieved by increasing the knit density in this area, using more abrasion-resistant fiber materials, or by heat-pressing or sewing an additional elastic support layer on the inside. This local reinforcement aims to provide stable support for the sensor assembly 1, reduce sensor displacement during movement, and improve the durability of this critical area. The reinforced area is slightly thicker than other parts to provide moderate pressure support while still maintaining comfort.

[0056] Multiple sensor components 1 are integrated on the inner side of the sock body, strategically deployed in the medial region of the first metatarsophalangeal joint. Specifically, the sensor components 1 are arranged in an arc shape. Figure 1 and Figure 2 The unfolded planar view is distributed along the natural anatomical contours of the joint capsule. This ensures the sensors fit snugly against the joint surface, accurately capturing subtle changes in the joint microenvironment before a gout attack, such as pressure changes caused by increased local temperature and tissue swelling. Each sensor assembly 1 includes a miniature temperature sensor and a miniature pressure sensor.

[0057] Miniature temperature sensors: NTC thermistors or RTD sensors can be selected, characterized by small size, fast response, and high accuracy. They can monitor local skin temperature with an accuracy of 0.1 degrees Celsius or even higher.

[0058] Miniature pressure sensors: These can be either piezoresistive thin-film sensors or capacitive pressure sensors, offering excellent flexibility and high sensitivity. They can detect minute pressure changes caused by tissue edema or inflammation. The sensor size is only a few millimeters square, ensuring no impact on wearing comfort.

[0059] Active microenvironment control structure: An active microenvironment control structure is provided around the sensor assembly 1. This structure aims to optimize the interface environment between the sensor and the skin, improving the accuracy and stability of data acquisition. The structure includes:

[0060] Microporous moisture-absorbing layer 4: Woven from specially modified polyester fibers, bamboo fibers, or microfibers, with a thickness of 0.5 mm to 1.0 mm and a porosity between 60% and 80%. This layer is located close to the skin surface and can quickly absorb local sweat, preventing sweat from accumulating and forming a liquid film, thereby avoiding interference with the measurements of temperature and pressure sensors.

[0061] Hydrophobic isolation membrane 5: disposed between the microporous moisture-absorbing layer 4 and the sensor assembly 1. This isolation membrane is made of hydrophobic materials such as polytetrafluoroethylene (PTFE) or polyethersulfone (PES) and has nanoscale pore size, which can effectively prevent sweat absorbed by the moisture-absorbing layer from back-permeating to the surface of the sensor assembly 1, ensuring that the sensor is always in a dry and stable working environment.

[0062] Sensor assembly 1 with dual-layer structure packaging: Sensor assembly 1 adopts a dual-layer structure packaging to improve its durability, comfort and maintainability.

[0063] Skin-friendly thermally conductive silicone layer 2: As the inner layer, its thickness is 0.3 mm to 0.5 mm. This layer is made of medical-grade skin-friendly thermally conductive silicone, which has good biocompatibility, thermal conductivity, and flexibility. Miniature temperature sensor probes and miniature pressure sensor probes are embedded within this silicone layer, ensuring direct contact between the layer and the user's skin for efficient and accurate physiological signal transmission.

[0064] Elastic fabric base layer 3: As the outer layer, it embeds flexible circuitry. This base layer is made of highly elastic, abrasion-resistant fabric material and integrates a flexible printed circuit board (FPCB) for connecting the sensor probe and transmitting signals. The skin-friendly, thermally conductive silicone layer 2 is connected to the elastic fabric base layer 3 via detachable magnetic contacts. The contacts consist of micro-magnets and conductive metal sheets, ensuring a stable electrical connection and easy disassembly. This allows the sensor assembly 1 to be easily separated from the sock body for convenient washing or sensor replacement.

[0065] Integrated Temperature and Pressure Sensing Unit: The temperature and pressure sensors in sensor assembly 1 are designed as an integrated temperature and pressure sensing unit, integrated within the same miniature package. This minimizes the sensor size and ensures that temperature and pressure data are acquired synchronously within the same microscopic region.

[0066] Miniature pressure sensors employ a piezoresistive thin-film structure, consisting of conductive ink or a thin metal film on a flexible polymer substrate. When the film is compressed and deformed, its resistance changes, thereby measuring the pressure.

[0067] The piezoresistive thin film structure is covered with a thermally conductive layer, which is a thin layer of thermally conductive silicone or thermally conductive polymer.

[0068] The miniature temperature sensor is closely attached to the underside of the heat-conducting layer to ensure accurate sensing of the local temperature transmitted through the heat-conducting layer.

[0069] The integrated temperature and pressure sensing unit shares a set of miniature signal conditioning circuits, which are responsible for converting the sensor's analog signals into digital signals and performing preliminary amplification, filtering, and analog-to-digital conversion to improve data quality and transmission efficiency.

[0070] The data processing unit includes a microcontroller (MCU) or a low-power system-on-a-chip (SoC). This unit is responsible for receiving and processing temperature and pressure data from sensor component 1. To achieve advanced early warning functions, the data processing unit integrates three types of algorithms:

[0071] Multimodal deep spatiotemporal fusion early warning algorithm: used for deep learning analysis of continuously acquired temperature and pressure data. Specific execution steps include:

[0072] Data construction: Continuously collected temperature sequences and pressure sequence Construct a multidimensional input matrix. Wherein, The length of the sampling sliding window can be set, for example, to collect temperature and pressure data within one hour, i.e., once per minute. =60. This matrix integrates temperature and pressure data in both time and spatial dimensions.

[0073] Spatial Feature Extraction: Spatial features are extracted from the multidimensional input matrix using a one-dimensional convolutional kernel of a convolutional neural network (CNN). The one-dimensional convolutional kernel can effectively capture the spatial correlation of local temperature and pressure patterns in the sensor array, generating a high-dimensional feature vector mapping.

[0074] Temporal dependency extraction: The high-dimensional feature vector is mapped and input into a Long Short-Term Memory (LSTM) network. LSTM, through its unique forget gate, input gate, and output gate structure, can effectively process and learn temporal dependencies in long-term series data, generating hidden state sequences. .

[0075] Attention-based weighted fusion: An attention mechanism is introduced to perform weighted fusion of the hidden state sequences. The attention mechanism can identify the time points or features in the sequence that contribute most to the warning and assign them higher weights, thereby calculating the context vector. Its calculation process involves calculating an attention score. and weighting coefficients Finally, the weighted summation of the context vector is obtained. .

[0076] Warning score output: The context vector is used for output. The system uses one or more fully connected layers to classify gout attacks and outputs a warning score for the current attack. This score can be a probability value between 0 and 1, or a multi-class risk level.

[0077] Personalized Risk Trajectory Prediction and Uncertainty Quantification Algorithm: This algorithm aims to predict gout risk trends over a future period and assess the reliability of the prediction results. Specific execution steps include:

[0078] Model Construction: Construct a risk prediction model based on Bayesian deep learning. Bayesian deep learning models can simultaneously provide predicted values ​​and the uncertainty of those predictions.

[0079] Monte Carlo Dropout Sampling: A Monte Carlo Dropout mechanism is introduced during the model's inference phase. For a given input sample (such as current temperature and pressure data, historical risk data, etc.), the Dropout layer is kept on, and sampling is performed... Sub-random forward propagation sampling (e.g.) =100), each sampling yields a prediction result. This forms a set of prediction results. .

[0080] Risk prediction trajectory calculation: Calculate the expected value of the prediction result. This serves as a risk prediction trajectory within a future time window (e.g., the next 6 or 12 hours).

[0081] Uncertainty quantification: Calculating the variance of prediction results As a quantitative indicator of model uncertainty, a larger variance indicates higher prediction uncertainty.

[0082] Early warning instruction confirmation: When the calculated expected value Exceeding the preset risk threshold and variance When the confidence level falls below a preset threshold, a warning command is output. Ensure that warnings are only issued when the model has a high confidence level in the prediction results to avoid false alarms.

[0083] Personalized Trigger Identification Algorithm Based on Causal Reasoning: This algorithm utilizes causal reasoning methods to deeply analyze the causal relationship between user behavior and physiological changes, identifying personalized triggers that increase the risk of gout. Specific execution steps include:

[0084] Variable definition:

[0085] Intervention variables These represent a user's external behavioral characteristics, such as the intake of specific foods (e.g., high-purine foods), prolonged standing, strenuous exercise, and alcohol consumption. These behaviors are acquired through user input in smart applications or through other sensors (such as activity trackers).

[0086] Outcome variable This represents changes in the physiological signs of the foot, specifically manifested as localized temperature increases and / or abnormal pressure detected by the sensor.

[0087] Confounding variables These represent a user's basal metabolic rate, such as age, gender, body mass index (BMI), kidney function, baseline uric acid level, and medication use. These variables are entered during user registration or obtained from other medical records.

[0088] Structural Causal Models and Counterfactual Inference: Structural causal models (SCM) are used to describe the causal relationships between variables (e.g., a high-purine diet may lead to elevated blood uric acid, which in turn affects foot physiological signs). The counterfactual inference framework is used to estimate the individual intervention effect. Counterfactual inference aims to answer the question, "How would a user's physiological characteristics change if they did not perform a certain action?" Its calculation method is as described in claim 7, wherein... Operators represent active interventions in behavior; for example, assuming a user has ingested high-purine foods. , compared to the assumption that the user did not ingest The physiological changes over time need to be compared. A model that can simulate the distribution of outcome variables under different intervention conditions needs to be established.

[0089] Trigger determination: When targeting a specific behavior For example, calculated from the intake of a specific food. If a behavior exceeds a preset significance threshold three times consecutively on different dates or at different events, it is considered a personalized sensitive trigger for that user. The threshold can be set based on clinical experience and statistical data; for example, when… The value indicates that the behavior increases the risk of gout by more than 20%.

[0090] The wireless communication module is used to transmit data processed by the data processing unit to external smart devices, such as smartphones or tablets. This module employs Bluetooth Low Energy technology to achieve efficient data transmission and long battery life. The module may also include a small antenna and necessary radio frequency circuitry.

[0091] The present invention also provides a warning system for wearable gout warning socks. The system includes the aforementioned wearable gout warning socks and a smart application running on an external smart device. This smart application serves as the user interface for interacting with the warning socks and provides a range of advanced functions.

[0092] With a structured dietary input interface and a built-in purine load calculator, this smart application provides a user-friendly, structured dietary input interface, allowing users to easily record their daily food intake, portion sizes, and meal times. The application includes a comprehensive purine load database and calculator that can calculate the purine load values ​​for single meals and the entire day in real time based on the types and quantities of food entered by the user. This database is updated regularly and covers the purine content of various common foods.

[0093] The diet-physiology association analysis module automatically associates the user's dietary intake records (intervention variables) with the aforementioned causal relationship inference personalized inducement identification algorithm. ) and subsequent physiological parameter changes collected from the warning socks (outcome variable) Through analysis, this module can identify personalized triggers that increase a user's risk of gout, such as the intake of specific seafood, beer, or animal offal, and present these associations to the user in an intuitive way.

[0094] The personalized timing optimization feature for medication reminders calculates the peak time of purine load in the body based on the user's individual metabolic characteristics and dietary records. The application then sets a medication reminder 1 to 2 hours before this peak. For example, if a high-purine meal is expected to cause a uric acid peak 4 hours after the meal, the system may issue a medication reminder 2-3 hours after the meal to ensure the medication has the best effect before the uric acid peak.

[0095] A multi-level early warning mechanism is implemented through intelligent applications to differentiate between different risk levels and provide corresponding intervention suggestions.

[0096] Level 1 Warning (Low Risk): Triggered when there are slight abnormalities in local temperature or pressure, but not yet reaching the threshold for a gout attack. Users are advised to rest, avoid fatigue, drink plenty of water, and review their recent diet and activities.

[0097] Level 2 Warning (Medium Risk): Triggered when physiological parameters worsen further, but no obvious symptoms have yet appeared. Users are advised to immediately adjust their diet, avoid high-purine foods, consider taking preventative medications such as colchicine under a doctor's guidance, and closely monitor symptoms.

[0098] Level 3 Warning (High Risk): Triggered when local physiological parameters such as temperature and pressure reach typical precursor values ​​for an acute gout attack, and the personalized risk trajectory prediction algorithm provides a high-risk, high-confidence prediction. Users are strongly advised to seek immediate medical attention or take emergency intervention measures as advised by their doctor, such as taking painkillers or anti-inflammatory drugs, and complete bed rest.

[0099] The smart application features a doctor-assisted diagnostic data package generation function. This function automatically organizes all relevant user data within a certain period and generates a structured health report. The report content includes:

[0100] Temperature change trend graph: shows the continuous change curve of local temperature of the first metatarsophalangeal joint, highlighting abnormal fluctuations.

[0101] Causal mapping of dietary records and temperature changes: visually demonstrates which dietary habits are causally related to physiological abnormalities, helping doctors quickly identify the triggers.

[0102] Warning event timeline: Records the occurrence time, risk level, and user intervention measures for all warning events.

[0103] Multimodal physiological characteristic change curves: In addition to temperature and pressure, these may also include derived composite indicators (such as inflammation indices), comprehensively reflecting changes in the joint microenvironment. This data package can be easily exported to PDF or other common formats, allowing users to provide it to doctors during medical visits to assist in more accurate diagnosis and treatment planning.

[0104] The shoe-changing reminder feature aims to prevent foot discomfort caused by excessive local pressure or poor breathability due to ill-fitting footwear, thereby indirectly reducing the risk of gout. The app will issue a reminder to change into loose-fitting, breathable shoes when the following conditions are met:

[0105] The overall foot temperature increased by more than 1.5 times the baseline standard deviation (the baseline standard deviation was learned from long-term data).

[0106] The pressure sensor detected multiple localized pressure points that persisted for more than 30 minutes.

[0107] Meanwhile, the system is not in a level 2 or 3 gout warning state (to avoid confusion with a gout attack). This reminder aims to guide users to improve the foot microenvironment in a timely manner and reduce potential mechanical stimulation or local overheating, which can be triggering factors for gout attacks.

[0108] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. The scope of patent protection of the present invention shall be determined by the claims. Similarly, any equivalent structural changes made based on the content of the present invention's specification shall also be included within the scope of protection of the present invention.

Claims

1. A wearable gout warning sock, characterized in that, include: A sock body made of an elastic, breathable fabric material, the sock body being locally reinforced in the first metatarsophalangeal joint region; Multiple sensor components (1) are integrated inside the body of the sock. The sensor components (1) are strategically deployed in the inner region of the first metatarsophalangeal joint and arranged in an arc shape, distributed along the natural anatomical contour of the joint capsule. The sensor components (1) include a miniature temperature sensor and a miniature pressure sensor. An active microenvironment control structure is provided around the sensor components (1). A data processing unit is used to receive and process temperature and pressure data from the sensor assembly (1), and internally integrates a multimodal deep spatiotemporal fusion early warning algorithm, a personalized risk trajectory prediction and uncertainty quantification algorithm, and a causal relationship reasoning personalized cause identification algorithm; A wireless communication module is used to send the data processed by the data processing unit to an external smart device.

2. The wearable gout warning socks according to claim 1, characterized in that, The sensor assembly (1) is packaged with a double-layer structure, the double-layer structure comprising: A skin-friendly thermally conductive silicone layer (2), as an inner layer, with a thickness of 0.3 mm to 0.5 mm, contains the micro temperature sensor probe and the micro pressure sensor probe, and is in direct contact with the user's skin; An elastic fabric base layer (3) serves as the outer layer, incorporating a flexible circuit. The skin-friendly thermally conductive silicone layer (2) is connected to the elastic fabric base layer (3) via detachable magnetic contacts, allowing the sensor assembly (1) to be detached from the sock body.

3. The wearable gout warning socks according to claim 1, characterized in that, The temperature sensor and the pressure sensor in the sensor assembly (1) are designed as an integrated temperature and pressure sensing unit, integrated in the same micro package. The micro pressure sensor adopts a piezoresistive thin film structure with a thermally conductive layer covering it. The micro temperature sensor is attached to the bottom of the thermally conductive layer. The integrated temperature and pressure sensing unit shares a set of micro signal conditioning circuits.

4. The wearable gout warning socks according to claim 1, characterized in that, The active microenvironment control structure includes: A microporous moisture-absorbing layer, woven from highly absorbent fibers, with a thickness of 0.5 mm to 1.0 mm and a porosity of 60% to 80%, is used to absorb localized sweat and keep the interface between the sensor assembly and the skin dry; A hydrophobic barrier membrane is disposed between the microporous moisture-absorbing layer and the sensor assembly to prevent sweat absorbed in the microporous moisture-absorbing layer from back-permeating to the surface of the sensor assembly.

5. The wearable gout warning socks according to claim 1, characterized in that, The specific execution steps of the multimodal deep spatiotemporal fusion early warning algorithm include: Continuously collected temperature sequences and pressure sequence Constructed as a multidimensional input matrix, where The length of the sampling sliding window; The spatial features of the multidimensional input matrix are extracted using a one-dimensional convolutional kernel of a convolutional neural network to generate a high-dimensional feature vector mapping. The high-dimensional feature vector is mapped and input into a long short-term memory network. Its forget gate, input gate, and output gate structures are used to extract temporal dependencies and generate a sequence of hidden states. An attention mechanism is introduced to perform weighted fusion of the hidden state sequences and calculate the context vector. : in, To score attention, For weight coefficients; use the context vector The system classifies gout attacks using a fully connected layer and outputs a warning score for the current attack.

6. The wearable gout warning socks according to claim 1, characterized in that, The specific execution steps of the personalized risk trajectory prediction and uncertainty quantification algorithm include: A risk prediction model based on Bayesian deep learning is constructed, and a Monte Carlo Dropout mechanism is introduced in the inference stage; for a given input sample, the Dropout layer is kept on, and... After several random forward propagation samplings, the set of prediction results is obtained. ; Calculate the expected value of the prediction results As a risk prediction trajectory within a future time window: Calculate the variance of the prediction results As a quantitative indicator of model uncertainty: when Exceeding the preset risk threshold and When the value falls below the preset threshold, a warning command is output.

7. The wearable gout warning socks according to claim 1, characterized in that, The specific execution steps of the personalized incentive identification algorithm for causal relationship reasoning include: Define users' external behavioral characteristics as intervention variables. Changes in foot physiological signs were used as the outcome variable. User basal metabolic rate is a confounding variable. ; Structural causal models are used to describe the causal relationships between variables, and counterfactual inference frameworks are used to estimate the individual intervention effects. : in, Operators represent active intervention in behavior; when targeting a specific behavior... Calculated If the behavior exceeds the preset significance threshold three times consecutively, it will be identified as a personalized sensitive trigger for the user.

8. A warning system for wearable gout warning socks, characterized in that, The wearable gout warning socks include any one of claims 1 to 7, and a smart application running on the external smart device, the smart application having: A structured diet input interface and a built-in purine load calculator are used to record the user's diet and calculate the purine load values ​​for a single meal and a day; The diet-physiology correlation analysis module uses a personalized trigger identification algorithm based on causal reasoning to automatically correlate dietary intake with subsequent changes in physiological parameters and identify personalized triggers that increase the risk of gout. The personalized timing optimization function for medication reminders calculates the peak time of purine load based on individual metabolic characteristics and dietary records, and sets medication reminders 1 to 2 hours before the peak. The multi-level early warning mechanism includes Level 1, Level 2, and Level 3 early warnings, each corresponding to different risk levels and intervention recommendations.

9. The warning system of the wearable gout warning socks according to claim 8, characterized in that, The intelligent application also has a doctor-assisted diagnostic data package generation function, which automatically generates weekly health reports that include temperature change trend graphs, causal graphs of dietary records and temperature changes, timelines of early warning events, and curves of changes in multimodal physiological characteristics.

10. The warning system of the wearable gout warning socks according to claim 8, characterized in that, The smart application also has a shoe-changing reminder function. When the overall foot temperature rises by more than 1.5 times the baseline standard deviation, and the pressure sensor detects multiple local pressure points that persist for more than 30 minutes, and the system is not in a level 2 or 3 gout warning state, it will issue a reminder to change to loose and breathable shoes.

Images