A wearable physiological cycle monitoring system based on multi-sensor fusion

Abstract

The application relates to the technical field of computers, and specifically discloses a wearable physiological cycle monitoring system based on multi-sensor fusion, aiming to solve the technical problems that existing menstrual period management applications take fertility prediction as the core, deviate from the needs of contemporary women to understand their bodies and obtain a sense of security, and traditional basal body temperature measurement methods are inconvenient to operate. The system comprises a multi-sensor fusion module, a local preprocessing module, a generative AI semantic reconstruction module and an Agent-driven closed-loop interaction module. The multi-sensor fusion module collects wrist skin temperature through an infrared thermocouple sensor, collects environmental temperature through an NTC thermistor sensor and collects body motion data through a six-axis inertial measurement unit. The local preprocessing module completes sleep state recognition, data cleaning, core body temperature estimation and physiological feature extraction on the wearable device end and generates a JSON format output. The generative AI semantic reconstruction module uses a large language model to convert physiological data into personalized health interpretation that focuses on users' body cognition and emotional needs. The Agent-driven closed-loop interaction module realizes complete closed-loop management of intelligent analysis and personalized reminders through memory storage and a decision engine. The application combines multi-sensor fusion with local preprocessing, improves the real-time performance and privacy security of data processing, realizes health interpretation with humanistic care through generative AI empowerment and accurately meets the core needs of contemporary women.

Description

Technical Field

[0001] This invention belongs to the field of computer technology, specifically relating to a wearable physiological cycle monitoring system based on multi-sensor fusion. Background Technology

[0002] Wearable physiological monitoring technology has made significant progress in recent years, with multi-sensor fusion becoming the mainstream technical approach for acquiring high-confidence physiological data. In the field of women's health monitoring, menstrual cycle management based on core body temperature (CBT) has always been a key focus for both clinical and consumer applications. However, existing technologies generally suffer from low data utilization and limited semantic expression, making it difficult for monitoring systems to meet the deeper needs of contemporary women for physical health awareness and psychological security.

[0003] At the implementation level, traditional wearable devices primarily rely on a single temperature sensor to collect wrist skin temperature data, failing to fully consider the dynamic impact of ambient temperature changes on the body's heat dissipation process. This results in limited accuracy in core body temperature estimation. More critically, even with multi-sensor data fusion, existing wearable systems remain at the "data display" stage, lacking the ability to deeply process massive amounts of physiological characteristic data. Consequently, users only receive cold numerical curves and probability predictions, failing to obtain personalized health interpretations tailored to their individual physical condition. This contradiction of abundant data but lack of insight poses a severe challenge to wearable physiological monitoring devices in terms of long-term user retention and stickiness.

[0004] Against this technological backdrop, how to efficiently reconstruct multi-dimensional physiological characteristic data semantically and transform it into health reminders with humanistic value through generative artificial intelligence technology has become a core technical challenge that urgently needs to be addressed in this field. Existing menstrual cycle management applications overemphasize fertility prediction, creating a significant deviation from the genuine needs of contemporary women to "understand their bodies and gain a sense of security." This not only limits the depth of user experience but also hinders the further development of wearable physiological monitoring technology in the fields of women's health management and body awareness. Summary of the Invention

[0005] This application provides a wearable menstrual cycle monitoring system based on multi-sensor fusion, which aims to solve the technical problem that existing menstrual management applications, which are centered on fertility prediction, have a serious deviation from the real needs of contemporary women to "understand their bodies and gain a sense of security". At the same time, it overcomes the technical defects of traditional basal body temperature measurement methods, such as inconvenience in operation, easy data interference, and difficulty in interpretation.

[0006] The first aspect of this application provides a wearable physiological cycle monitoring system based on multi-sensor fusion, including a multi-sensor fusion module, a local preprocessing module, a generative AI semantic reconstruction module, and an agent-driven closed-loop interaction module. The multi-sensor fusion module is used to simultaneously collect the user's physiological signals and environmental data through multiple sensors, and transmit the collected raw data to the local preprocessing module. The local preprocessing module is used to perform sleep state recognition, data cleaning, core body temperature estimation and physiological feature extraction on the received raw data, generate JSON format output containing highly condensed physiological feature data, and send the JSON format output to the generative AI semantic reconstruction module. The generative AI semantic reconstruction module is used to receive the JSON format output and input the JSON format output into a large language model. By utilizing the semantic understanding and generation capabilities of the large language model, physiological characteristic data is transformed into personalized humanistic health interpretation information that is non-fertility-oriented and focuses on body cognition and sense of security. The Agent-driven closed-loop interaction module is used to construct an intelligent agent with memory, autonomous decision-making, and proactive reminder capabilities. Combined with a hardware haptic feedback unit, it realizes a complete closed-loop management of physiological state perception, intelligent analysis, personalized reminders, and user feedback.

[0007] In one optional implementation, the multi-sensor fusion module includes an infrared temperature measurement unit, an environmental compensation temperature measurement unit, and a motion state detection unit. The infrared temperature measurement unit uses an infrared thermopile sensor to collect the user's wrist skin temperature at a resolution of 0.01℃ via a UART interface. The environmental compensation temperature measurement unit uses an NTC thermistor temperature sensor to collect the raw value of the ambient temperature and transmit it to the local preprocessing module for Celsius conversion. The motion state detection unit uses a six-axis inertial measurement unit to collect three-axis acceleration data through an I2C interface, which is used to determine the user's motion state and duration of stillness.

[0008] In one optional implementation, the local preprocessing module includes a sleep state recognition unit, a data cleaning unit, a core body temperature estimation unit, and a physiological feature extraction unit. The sleep state recognition unit calculates the variance of motion volume within a preset time window based on the triaxial acceleration data collected by the motion state detection unit. When the variance of motion volume is lower than a preset threshold and the duration reaches the sleep window determination condition, the user is identified as entering a sleep state. The data cleaning unit is used to remove abnormal fluctuation data from the data identified as sleep periods, and retain stable and continuous sleep data segments. The core body temperature estimation unit is used to estimate the user's core body temperature by correcting the skin temperature data during the sleep period after washing and combining it with the ambient temperature data. Specifically, it adopts the core body temperature estimation algorithm in the paper by Sato et al., which estimates the core body temperature by the formula equal to the skin temperature plus the correction coefficient multiplied by the difference between the skin temperature and the ambient temperature. The physiological feature extraction unit is used to extract the daily representative core body temperature value and the nighttime core body temperature amplitude from the estimated core body temperature time series data. The daily representative core body temperature value is the 99th percentile value of the core body temperature data for that day, and the nighttime core body temperature amplitude is the difference between the maximum and minimum core body temperature values ​​within the day-night cycle.

[0009] In one optional implementation, the local preprocessing module further includes a JSON output unit for packaging the extracted physiological feature data into a standard JSON format, which includes a user identification field, a collection time field, a daily representative core body temperature field, a nighttime core body temperature amplitude field, and a sleep quality assessment field.

[0010] In one optional implementation, the generative AI semantic reconstruction module includes a semantic understanding unit and a personalized generation unit; The semantic understanding unit is used to parse the physiological feature data in the JSON format output and extract the values ​​and trends of key physiological indicators. The personalized generation unit is used to transmit the parsed physiological feature data as part of the input prompt to a large language model, instructing the large language model to generate health interpretation text that focuses on the user's physical perception, emotional state and sense of security, rather than rhetoric centered on fertility assessment or ovulation prediction.

[0011] In one optional implementation, the Agent-driven closed-loop interaction module includes a memory storage unit, a decision engine unit, an alert generation unit, and a feedback processing unit. The memory storage unit is used to store the user's long-term physiological characteristics, periodic behavioral patterns, and historical health interpretation feedback records to build a personalized user profile. The decision engine unit is used to compare and analyze the current physiological characteristic data with the historical data in the user's personalized profile, determine the stage category to which the current physiological state belongs, and generate a corresponding response strategy. The reminder generation unit is used to generate humanistic reminder content that conforms to user preferences and current status according to the response strategy output by the decision engine unit, and to convey it to the user through the visual feedback unit or the tactile feedback unit. The feedback processing unit is used to receive user confirmation, modification or ignore operations on the reminder content, and update the feedback results to the memory storage unit to optimize the accuracy of subsequent decision engine judgments.

[0012] In one alternative implementation, the haptic feedback unit employs an air pump structure to generate haptic stimulation signals through minute changes in air pressure, thereby achieving a more immersive and natural interactive experience.

[0013] In one optional implementation, the local preprocessing module runs on the main control board of the wearable device, which is a microcontroller with low-power WiFi and Bluetooth dual-mode wireless communication capabilities; the generative AI semantic reconstruction module and the agent-driven closed-loop interaction module run on an external terminal or cloud server that is wirelessly connected to the wearable device.

[0014] In one optional implementation, the multi-sensor fusion module further includes a calibration unit for offset compensation of the infrared temperature measurement unit based on the ambient temperature before each acquisition begins, to ensure the accuracy of skin temperature measurement.

[0015] In the wearable menstrual cycle monitoring system based on multi-sensor fusion provided in this application, the multi-sensor fusion module collects wrist skin temperature through an infrared thermopile sensor, ambient temperature through an NTC thermistor sensor, and body motion data through a six-axis inertial measurement unit, achieving simultaneous acquisition of multi-dimensional physiological data. The local preprocessing module completes intelligent sleep state determination, data cleaning, scientific estimation of core body temperature, and high-dimensional extraction of physiological characteristics on the wearable device, outputting JSON format data that can be directly used by large language models. The generative AI semantic reconstruction module utilizes the semantic understanding capabilities of large language models to transform cold physiological data into personalized health interpretations that focus on users' physical cognition and emotional needs, completely overturning the existing menstrual cycle management model oriented towards fertility. The agent-driven closed-loop interaction module achieves complete closed-loop management from physiological state perception to intelligent analysis, personalized reminders, and finally user feedback through the synergistic effect of memory storage, decision engine, and tactile feedback.

[0016] Thus, this application, through a technical solution combining multi-sensor data fusion and local preprocessing, completes complex data processing and feature extraction on the wearable device, reducing reliance on cloud computing and improving the real-time performance and privacy security of data processing. Through generative AI-enabled semantic reconstruction, it upgrades traditional data display to a health interpretation with humanistic care, precisely meeting the core needs of contemporary women to "understand their bodies and gain a sense of security." Through an agent-driven closed-loop interaction mechanism, the system possesses the ability to continuously learn and adaptively adjust, providing users with a more accurate and personalized menstrual cycle management experience. Through a scientific core body temperature estimation algorithm, it effectively overcomes the shortcomings of traditional basal body temperature measurement methods, such as inconvenience and susceptibility to environmental interference, providing users with a reliable and convenient means of body temperature monitoring. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall technical architecture of the wearable physiological cycle monitoring system based on multi-sensor fusion proposed in this invention. Figure 2 This is a schematic diagram illustrating the core principle framework of the generative AI semantic reconstruction module in this invention, which transforms physiological feature data into personalized, humanistic health interpretation information that is non-fertility-oriented and focuses on body cognition and sense of security. Figure 3 This is a flowchart illustrating the main processing stages of the local preprocessing module in this invention, including sleep state recognition, data cleaning, core body temperature estimation, and physiological feature extraction. Figure 4 This is a schematic diagram of the multi-level interaction relationship and data flow of the Agent-driven closed-loop interaction module in this invention, including memory storage, decision engine, reminder generation, and feedback processing. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] In the accompanying drawings, the size of constituent elements, the thickness of layers, or areas may sometimes be exaggerated for clarity. Therefore, any implementation of this disclosure is not necessarily limited to the dimensions shown in the drawings, and the shapes and sizes of the components in the drawings do not reflect true proportions. Furthermore, the drawings schematically illustrate ideal examples, and any implementation of this disclosure is not limited to the shapes or values ​​shown in the drawings.

[0020] Reference Figure 1 , Figure 1 This is a schematic diagram of the structure of a wearable physiological cycle monitoring system based on multi-sensor fusion proposed in an embodiment of this application. As shown in the figure, the system includes: a multi-sensor fusion module, a local preprocessing module, a generative AI semantic reconstruction module, and an agent-driven closed-loop interaction module.

[0021] In this embodiment, the wearable device worn by the user synchronously collects the user's physiological signals and environmental data through a multi-sensor fusion module. The collected raw data is transmitted to a local preprocessing module for noise reduction, cleaning, and feature extraction. The extracted high-dimensional physiological feature data is packaged into JSON format and sent to a generative AI semantic reconstruction module. This module utilizes the semantic understanding capabilities of a large language model to transform the data into a caring health interpretation. Finally, the Agent-driven closed-loop interaction module generates personalized reminders and interaction strategies based on the interpretation content and the user's preference profile, and delivers them to the user through a haptic feedback unit, completing a full closed loop from perception to feedback.

[0022] In one embodiment, this application also provides a wearable physiological cycle monitoring system based on multi-sensor fusion, wherein the multi-sensor fusion module includes an infrared temperature measurement unit, an environmental compensation temperature measurement unit, and a motion state detection unit. The infrared temperature measurement unit is used to collect the user's wrist skin temperature; specifically, the infrared temperature measurement unit uses an infrared thermopile sensor, which continuously collects the user's wrist skin temperature data at a resolution of 0.01℃ via a UART interface. The raw voltage signal output by the sensor is quantized by an analog-to-digital converter and then transmitted to a local preprocessing module for temperature value calibration.

[0023] The environmental compensation temperature measurement unit is used to collect raw ambient temperature values. Specifically, the environmental compensation temperature measurement unit uses an NTC thermistor temperature sensor, which collects raw ambient temperature values ​​via an I2C interface at a preset sampling rate. These raw values ​​are then transmitted to a local preprocessing module, which performs Celsius conversion and serves as a basis for offset compensation of the infrared temperature measurement unit's measured values.

[0024] The motion state detection unit is used to determine the user's motion state and duration of stillness. Specifically, the motion state detection unit uses a six-axis inertial measurement unit to collect three-axis acceleration data via an I2C interface. The local preprocessing module calculates the composite acceleration based on the magnitude of the three-axis acceleration data and determines the user's motion state (e.g., walking, stillness, vigorous exercise) accordingly. Simultaneously, it records the user's duration of stillness, providing basic data for subsequent sleep state recognition.

[0025] In one embodiment, this application also provides a wearable physiological cycle monitoring system based on multi-sensor fusion. In this system, the aforementioned local preprocessing module includes a sleep state recognition unit, a data cleaning unit, a core body temperature estimation unit, and a physiological feature extraction unit. The sleep state recognition unit is used to identify when a user enters a sleep state. Specifically, the sleep state recognition unit calculates the variance of the amount of movement within a preset time window (e.g., 5 minutes) based on the triaxial acceleration data collected by the motion state detection unit. When the variance of the amount of movement is lower than a preset threshold (e.g., 0.1 g²) and the duration reaches the sleep window determination condition (e.g., 30 consecutive minutes), the system recognizes that the user has entered a sleep state and marks this period as a sleep data segment.

[0026] The data cleaning unit is used to remove abnormal fluctuation data from the data identified as sleep periods. Specifically, the data cleaning unit traverses all temperature sampling points in the sleep data segment and uses a preset amplitude limiting filtering algorithm to remove spike noise caused by poor sensor contact or instantaneous action, retaining stable and continuous sleep data segments for subsequent core body temperature estimation.

[0027] The core body temperature estimation unit is used to estimate the user's core body temperature. Specifically, the core body temperature estimation unit estimates the user's core body temperature based on skin temperature data during sleep after washing, combined with ambient temperature data for correction. This embodiment uses the core body temperature estimation algorithm from the paper by Sato et al., estimating core body temperature through a formula. The formula for calculating core body temperature is as follows:

[0028] in, This represents the estimated core body temperature. This indicates the skin temperature data after washing. Indicates ambient temperature. These are correction factors. The value is dynamically adjusted based on individual user differences and historical data, and is usually between 0.5 and 0.8 to ensure that the estimated core body temperature conforms to human physiological laws.

[0029] The physiological feature extraction unit is used to extract key physiological features from the estimated core body temperature time series data. Specifically, the daily representative core body temperature value is the 99th percentile value of the core body temperature data on that day (representing the peak body temperature in a resting state), and the nighttime core body temperature amplitude is the difference between the maximum and minimum core body temperature values ​​within the day-night cycle (reflecting the diurnal fluctuation of body temperature).

[0030] In one embodiment, this application also provides a wearable menstrual cycle monitoring system based on multi-sensor fusion. In this system, the local preprocessing module further includes a JSON output unit for packaging the extracted physiological feature data into a standard JSON format. The data packet generated by the JSON output unit includes the following fields: UserID, Timestamp, DailyCoreTemp, NightAmplitude, and SleepScore. After being packaged, the JSON output is sent to the generative AI semantic reconstruction module via the wearable device's wireless communication module.

[0031] In one embodiment, this application also provides a wearable physiological cycle monitoring system based on multi-sensor fusion, in which the generative AI semantic reconstruction module includes a semantic understanding unit and a personalized generation unit. The semantic understanding unit is used to parse physiological feature data in JSON format output; specifically, the semantic understanding unit reads JSON data packets, extracts the values ​​(such as 36.7℃) and trends (such as "up 0.2℃ from yesterday") of key physiological indicators, and transforms this structured data into text descriptions that can be processed by large language models.

[0032] The personalized generation unit is used to generate non-fertility-oriented health interpretation text. Specifically, the personalized generation unit uses parsed physiological characteristic data as part of the input prompt and transmits it to a large language model. The input prompt explicitly instructs the large language model to ignore fertility assessments or ovulation predictions, and instead focus on the user's physical sensations, emotional state, and need for security. For example, the prompt might include the instruction, "This user's core body temperature today is XX, and their sleep quality last night was YY. Please provide health advice from the perspective of body maintenance and self-care." The large language model then generates personalized health interpretation text based on this instruction.

[0033] In one embodiment, this application also provides a wearable physiological cycle monitoring system based on multi-sensor fusion. In this system, the Agent-driven closed-loop interaction module includes a memory storage unit, a decision engine unit, an reminder generation unit, and a feedback processing unit. The memory storage unit is used to build a personalized user profile; specifically, the memory storage unit stores the user's long-term physiological characteristics (daily core body temperature and amplitude over the past 30 days or even longer), periodic behavioral patterns (such as fixed exercise time and wake-up time), and historical health interpretation feedback records (user's preference markings for specific types of interpretations) based on a time-series database.

[0034] The decision engine unit is used to determine the current physiological state and generate response strategies. Specifically, the decision engine unit compares and analyzes the current physiological characteristic data with historical data in the user's personalized profile. Through a preset stage classification algorithm (such as morphological analysis based on body temperature curves), it determines the stage category of the current physiological state (such as menstruation, follicular phase, luteal phase, or, for those who are more concerned about health, the "adjustment period" or "stable period"), and generates corresponding response strategies.

[0035] The reminder generation unit is used to generate human-centered reminder content. Specifically, based on the response strategy output by the decision engine unit and combined with the user's preference settings (e.g., whether they prefer morning or evening reminders, text or haptic feedback), the reminder generation unit generates specific reminder content. The reminder content is delivered to the user through a visual feedback unit (displayed on the screen) or a haptic feedback unit.

[0036] The feedback processing unit is used to optimize subsequent decisions. Specifically, it receives user confirmation, modification, or ignoring of reminder content and updates the feedback results to the memory storage unit. If a user repeatedly ignores a certain type of reminder, the system will automatically reduce the weight of that type of reminder. If a user provides positive feedback on a certain interpretation (such as "very helpful"), the system will enhance the interpretation depth of the relevant features, thereby optimizing the accuracy of the subsequent decision engine's judgment.

[0037] In one embodiment, this application also provides a wearable physiological cycle monitoring system based on multi-sensor fusion, in which the haptic feedback unit adopts an air pump structure. The haptic feedback unit incorporates a miniature air pump and a miniature airbag, generating tactile stimulation signals through minute changes in air pressure. Compared to traditional vibration motors, this air pressure-based haptic feedback offers higher discernibility and a more natural interactive experience, enabling the transmission of discreet physiological status reminders to the user without disturbing others.

[0038] In one embodiment, this application also provides a wearable menstrual cycle monitoring system based on multi-sensor fusion. In this system, a local preprocessing module runs on the main control board of the wearable device. Specifically, the main control board of the wearable device uses a microcontroller (such as the ESP32 series) with low-power WiFi and Bluetooth dual-mode wireless communication capabilities. This microcontroller is responsible for controlling sensor data acquisition, executing local preprocessing algorithms, and sending the processed JSON data to the user's mobile phone via Bluetooth or to a cloud server via WiFi. The generative AI semantic reconstruction module and the agent-driven closed-loop interaction module run on an external terminal (such as a smartphone APP) or a cloud server connected to the wearable device via wireless communication, utilizing powerful computing capabilities for complex semantic understanding and personalized decision-making.

[0039] In one embodiment, this application also provides a wearable menstrual cycle monitoring system based on multi-sensor fusion. In this system, the multi-sensor fusion module further includes a calibration unit. The calibration unit compensates for the offset of the infrared thermometer's measurements based on the ambient temperature before each data acquisition begins. Specifically, the calibration unit reads the ambient temperature collected by the NTC thermistor, calculates and compensates for the nonlinear effect of ambient temperature on the infrared thermometer results based on a pre-determined sensor temperature-voltage characteristic curve, ensuring the accuracy of skin temperature measurements under various ambient temperatures.

[0040] Thus, the wearable menstrual cycle monitoring system based on multi-sensor fusion provided in this application completes complex data processing and feature extraction on the wearable device through a technical solution combining multi-sensor data fusion and local preprocessing, reducing reliance on cloud computing and improving the real-time performance and privacy security of data processing. Through generative AI-enabled semantic reconstruction, it upgrades traditional data display to a health interpretation with humanistic care, accurately meeting the core needs of contemporary women to "understand their bodies and gain a sense of security." Through an agent-driven closed-loop interaction mechanism, the system has the ability to continuously learn and adaptively adjust, providing users with a more accurate and personalized menstrual cycle management experience. Through a scientific estimation algorithm for core body temperature, it effectively overcomes the shortcomings of traditional basal body temperature measurement methods, such as inconvenience and susceptibility to environmental interference, providing users with a reliable and convenient means of body temperature monitoring.

[0041] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0042] Those skilled in the art will understand that embodiments of this application can be provided as methods, apparatus, or computer program products. Therefore, embodiments of this application can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of this application can take the form of one or more computer-usable storage media (including but not limited to disk storage, etc.) containing computer-usable program code. It takes the form of a computer program product implemented on (such as optical memory, etc.).

[0043] This application describes embodiments with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0044] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0045] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0046] Although preferred embodiments of the present application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present application.

[0047] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0048] The above provides a detailed description of a wearable physiological cycle monitoring system based on multi-sensor fusion provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A wearable menstrual cycle monitoring system based on multi-sensor fusion, characterized in that, It includes a multi-sensor fusion module, a local preprocessing module, a generative AI semantic reconstruction module, and an agent-driven closed-loop interaction module; the multi-sensor fusion module is used to simultaneously collect the user's physiological signals and environmental data through multiple sensors, and transmit the collected raw data to the local preprocessing module; The local preprocessing module is used to perform sleep state recognition, data cleaning, core body temperature estimation, and physiological feature extraction on the received raw data, generating a JSON format output containing highly condensed physiological feature data, and sending the JSON format output to the generative AI semantic reconstruction module. The generative AI semantic reconstruction module is used to receive the JSON format output and input it into a large language model. Utilizing the semantic understanding and generation capabilities of the large language model, the physiological feature data is transformed into personalized, humanistic health interpretation information that focuses on body cognition and sense of security, without being fertility-oriented. The agent-driven closed-loop interaction module is used to construct an intelligent agent with memory, autonomous decision-making, and proactive reminder capabilities. Combined with a hardware haptic feedback unit, it realizes a complete closed-loop management of physiological state perception, intelligent analysis, personalized reminders, and user feedback.

2. The wearable menstrual cycle monitoring system based on multi-sensor fusion according to claim 1, characterized in that, The multi-sensor fusion module includes an infrared temperature measurement unit, an environmental compensation temperature measurement unit, and a motion state detection unit. The infrared temperature measurement unit uses an infrared thermopile sensor to collect the user's wrist skin temperature at a resolution of 0.01℃ via a UART interface. The environmental compensation temperature measurement unit uses an NTC thermistor temperature sensor to collect the raw ambient temperature value and transmit it to the local preprocessing module for Celsius conversion. The motion state detection unit uses a six-axis inertial measurement unit to collect three-axis acceleration data via an I2C interface to determine the user's motion state and duration of stillness.

3. The wearable menstrual cycle monitoring system based on multi-sensor fusion according to claim 2, characterized in that, The local preprocessing module includes a sleep state recognition unit, a data cleaning unit, a core body temperature estimation unit, and a physiological feature extraction unit. The sleep state recognition unit calculates the variance of motion volume within a preset time window based on the triaxial acceleration data collected by the motion state detection unit. When the variance of motion volume is lower than a preset threshold and the duration reaches the sleep window determination condition, the user is identified as entering a sleep state. The data cleaning unit is used to remove abnormal fluctuation data from the data identified as sleep state periods and retain stable and continuous sleep data segments. The core body temperature estimation unit is used to estimate the user's core body temperature based on the cleaned sleep period skin temperature data and combined with the ambient temperature data. The physiological feature extraction unit is used to extract daily representative core body temperature values ​​and nighttime core body temperature amplitudes from the estimated core body temperature time-series data.

4. The wearable menstrual cycle monitoring system based on multi-sensor fusion according to claim 3, characterized in that, The local preprocessing module also includes a JSON output unit, which packages the extracted physiological feature data into a standard JSON format, including a user identification field, a collection time field, a daily representative core body temperature field, a nighttime core body temperature amplitude field, and a sleep quality assessment field.

5. The wearable menstrual cycle monitoring system based on multi-sensor fusion according to claim 1, characterized in that, The generative AI semantic reconstruction module includes a semantic understanding unit and a personalized generation unit. The semantic understanding unit is used to parse the physiological feature data in the JSON format output and extract the values ​​and trends of key physiological indicators. The personalized generation unit is used to transmit the parsed physiological feature data as a component of the input prompt to a large language model, instructing the large language model to generate health interpretation text that focuses on the user's physical perception, emotional state, and sense of security.

6. The wearable menstrual cycle monitoring system based on multi-sensor fusion according to claim 1, characterized in that, The Agent-driven closed-loop interaction module includes a memory storage unit, a decision engine unit, an alert generation unit, and a feedback processing unit. The memory storage unit is used to store the user's long-term physiological characteristic historical data, periodic behavioral patterns, and historical health interpretation feedback records to build a personalized user profile. The decision engine unit is used to compare and analyze the current physiological characteristic data with the historical data in the user's personalized profile, determine the stage category to which the current physiological state belongs, and generate a corresponding response strategy; the reminder generation unit is used to generate humanistic reminder content that conforms to the user's preferences and current state according to the response strategy output by the decision engine unit, and convey it to the user through the visual feedback unit or the tactile feedback unit. The feedback processing unit is used to receive user confirmation, modification or ignore operations on the reminder content, and update the feedback results to the memory storage unit to optimize the accuracy of subsequent decision engine judgments.

7. The wearable menstrual cycle monitoring system based on multi-sensor fusion according to claim 6, characterized in that, The haptic feedback unit uses an air pump structure to generate haptic stimulation signals through minute changes in air pressure, achieving a more immersive and natural interactive experience.

8. The wearable menstrual cycle monitoring system based on multi-sensor fusion according to claim 1, characterized in that, The local preprocessing module runs on the main control board of the wearable device, which is a microcontroller with low-power WiFi and Bluetooth dual-mode wireless communication capabilities. The generative AI semantic reconstruction module and the agent-driven closed-loop interaction module run on an external terminal or cloud server that is wirelessly connected to the wearable device.

9. The wearable menstrual cycle monitoring system based on multi-sensor fusion according to claim 2, characterized in that, The multi-sensor fusion module also includes a calibration unit, which compensates for the offset of the infrared temperature measurement unit based on the ambient temperature before each acquisition begins, to ensure the accuracy of skin temperature measurement.

10. The wearable menstrual cycle monitoring system based on multi-sensor fusion according to claim 3, characterized in that, The core body temperature estimation unit uses a core body temperature estimation algorithm to estimate the core body temperature as equal to the skin temperature plus a correction factor multiplied by the difference between the skin temperature and the ambient temperature. The daily representative core body temperature value is the 99th percentile of the core body temperature data for that day. The nighttime core body temperature amplitude is the difference between the maximum and minimum core body temperature values ​​within the day-night cycle.

Images