Method and system for generating personalized home moxibustion guidance scheme based on internet of things

Abstract

The application discloses a personalized home moxibustion guidance scheme generation method and system based on the Internet of Things, and relates to the technical field of data processing. The method comprises the following steps: collecting real-time physiological data and moxibustion equipment operation data of a target user, and dynamically dividing a current moxibustion stage; extracting key biological characteristics in the current moxibustion stage, and combining user historical moxibustion reaction records to calculate a personalized heat tolerance dynamic threshold; inputting the key biological characteristics and the heat tolerance dynamic threshold into a pre-trained adaptive adjustment model to output an optimal temperature-time length regulation parameter group; generating a personalized moxibustion guidance scheme comprising a staged temperature control curve, a single hole recommended time length and real-time biological feedback monitoring points according to the optimal temperature-time length regulation parameter group, and driving the moxibustion equipment to perform preheating. The application effectively balances the moxibustion temperature and the moxibustion time length, and improves the personalized level of home moxibustion.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a method and system for generating personalized home moxibustion guidance plans based on the Internet of Things. Background Technology

[0002] With the development of IoT technology and smart health devices, home moxibustion health care is becoming increasingly popular. Existing technologies can collect users' physiological data and device operating status through smart moxibustion devices, and perform simple temperature and time control based on preset programs, providing a preliminary foundation for the quantitative management of the moxibustion process.

[0003] However, traditional home moxibustion mainly relies on personal subjective feelings and fixed experience duration for operation, lacking objective and quantitative control standards, resulting in huge differences in effects and risks among different users; moreover, traditional methods are difficult to balance moxibustion temperature and moxibustion time, and excessive pursuit of high temperature or excessively long single moxibustion time can easily cause skin burns, blisters, or even depletion of body fluids leading to adverse reactions such as internal heat. Summary of the Invention

[0004] This invention provides a method and system for generating personalized home moxibustion guidance plans based on the Internet of Things, aiming to solve the technical problem that existing home moxibustion technologies struggle to balance moxibustion temperature and duration.

[0005] In view of the above problems, the present invention provides a method and system for generating personalized home moxibustion guidance programs based on the Internet of Things.

[0006] In a first aspect, the present invention provides a method for generating personalized home moxibustion guidance plans based on the Internet of Things, including: Collect real-time physiological data of the target user and operating data of the moxibustion device to dynamically divide the current moxibustion stage; Key biomarkers within the current moxibustion stage are extracted and combined with the user's historical moxibustion response records to calculate a personalized dynamic threshold for heat tolerance. The key biofeatures and dynamic heat tolerance thresholds are input into a pre-trained adaptive adjustment model, and the optimal temperature-duration control parameter set is obtained from the output. Based on the optimal temperature-duration control parameter set, a personalized moxibustion guidance plan is generated, which includes a phased temperature control curve, a suggested duration for a single acupoint, and key points for real-time biofeedback monitoring, and the moxibustion device is driven to perform preheating.

[0007] Secondly, this invention provides a personalized home moxibustion guidance program generation system based on the Internet of Things, comprising: The dynamic stage division module is used to collect real-time physiological data of the target user and operating data of the moxibustion device to dynamically divide the current moxibustion stage; The heat tolerance threshold calculation module is used to extract key biological characteristics within the current moxibustion stage and, in conjunction with the user's historical moxibustion reaction records, calculate a personalized dynamic heat tolerance threshold. The regulation parameter output module is used to input the key biofeatures and the dynamic threshold of heat tolerance into the pre-trained adaptive regulation model and output the optimal temperature-duration regulation parameter set. The guidance scheme generation module is used to generate a personalized moxibustion guidance scheme based on the optimal temperature-duration control parameter group, which includes a phased temperature control curve, a suggested duration for a single acupoint, and key points for real-time biofeedback monitoring, and to drive the moxibustion device to perform preheating.

[0008] One or more technical solutions provided in this invention have at least the following technical effects or advantages: This invention provides a method and system for generating personalized home moxibustion guidance plans based on the Internet of Things (IoT). It collects real-time user physiological data and moxibustion device operation data, dynamically divides moxibustion stages, and accurately matches different conditioning states during the moxibustion process. It calculates personalized dynamic heat tolerance thresholds based on the user's historical moxibustion reactions, achieving objective quantification of the user's heat tolerance level. Furthermore, it outputs optimal temperature-duration control parameters based on an adaptive adjustment model, effectively balancing moxibustion temperature and duration. Finally, it generates personalized moxibustion guidance plans and drives device preheating, providing standardized and personalized operational guidelines for home moxibustion, avoiding burns and other problems caused by improper subjective operation, effectively balancing moxibustion temperature and duration, and improving the personalization level of home moxibustion. Attached Figure Description

[0009] Figure 1 A flowchart illustrating the method for generating personalized home moxibustion guidance schemes based on the Internet of Things, as provided in an embodiment of the present invention; Figure 2 A schematic diagram of the structure of the IoT-based personalized home moxibustion guidance scheme generation system provided in an embodiment of the present invention; The components represented by each number in the attached diagram are explained below: The module includes a dynamic stage division module 11, a heat tolerance threshold calculation module 12, a control parameter output module 13, and a guidance scheme generation module 14. Detailed Implementation

[0010] This invention provides a method and system for generating personalized home moxibustion guidance plans based on the Internet of Things, which is used to solve the technical problem that existing home moxibustion techniques are unable to balance moxibustion temperature and moxibustion duration.

[0011] Example 1, as Figure 1 As shown, this invention provides a method for generating personalized home moxibustion guidance plans based on the Internet of Things, the method comprising: S100: Collects real-time physiological data of the target user and operating data of the moxibustion device, and dynamically divides the current moxibustion stage.

[0012] In this embodiment of the invention, real-time physiological data of the target user and operational data of the moxibustion device are collected to dynamically divide the current moxibustion stage. Traditional home moxibustion does not combine the user's real-time physiological state and the operational status of the moxibustion device to divide the moxibustion stage, but relies solely on a fixed procedure or subjective feeling to control the moxibustion process. This results in problems such as vague stage division and poor temperature control and duration adaptability, which not only fails to ensure the effective exertion of the moxibustion thermal effect, but also easily leads to safety risks such as skin burns and internal heat due to improper stage control. Therefore, it is necessary to collect multi-dimensional real-time data of the user and combine it with personal historical moxibustion data to dynamically divide the moxibustion stage, providing an accurate stage basis for subsequent personalized heat tolerance calculation and adjustment parameter generation.

[0013] Step S100 in the method provided in this embodiment of the invention includes: First, real-time physiological data of the target user and operating data of the moxibustion device are collected.

[0014] The real-time physiological data includes the rate of change of skin surface temperature and heart rate variability index collected by wearable devices and flexible sensors, and the operating data of the moxibustion device includes the real-time temperature of the moxibustion head and the current cumulative working time.

[0015] Real-time physiological data are quantitative indicators reflecting the user's real-time bodily responses during moxibustion, including skin surface temperature change rate and heart rate variability. The skin surface temperature change rate is the magnitude of change in skin surface temperature per unit time; the heart rate variability is the difference in heart rate between successive heartbeats, reflecting the state of the body's autonomic nervous system regulation. Moxibustion device operation data reflects the real-time working status of the device, including the real-time temperature of the moxibustion head and the current cumulative working time. The real-time temperature of the moxibustion head is the current temperature of the heating element of the moxibustion head; the current cumulative working time is the cumulative running time of the moxibustion device since its initial startup.

[0016] For example, a wearable device collects the user's heart rate variability (HRV) index, and a flexible sensor is attached to the skin at the moxibustion site to collect the rate of change of skin surface temperature. Simultaneously, the moxibustion device is connected via the Internet of Things (IoT) to read the real-time temperature of the moxibustion head and the current cumulative working time. For instance, for target user A, the wearable device on the wrist collects an HRV index of 40ms, the flexible sensor attached to the abdominal moxibustion site collects a skin surface temperature change rate of 0.8℃ / min, and the IoT reads a real-time temperature of 42℃ for the moxibustion head and a current cumulative working time of 5 minutes.

[0017] Secondly, the current moxibustion stage is dynamically divided.

[0018] The dynamic division of the current moxibustion stage includes: The rate of change of skin surface temperature is compared with a preset initial stress threshold, and the duration during which the rate of change is consistently higher than the initial stress threshold is defined as the initial stage of heat stress. After the initial stage of heat stress ends, when the heart rate variability index falls into the steady-state regulation range and the real-time temperature of the moxibustion head remains stable, it is divided into the heat effect duration stage. When the current accumulated working time reaches the personalized stage duration benchmark generated based on the user's historical moxibustion response records, it is divided into the preparation stage for the end of the treatment.

[0019] First, the rate of change in skin surface temperature is compared with a preset initial stress threshold. The duration during which the rate of change remains above the initial stress threshold is defined as the initial stage of heat stress. The initial stress threshold is a universal critical value used to distinguish between imperceptible / mild heating and a clear heat stress response, determined through statistical analysis of initial response data on the rate of change in skin surface temperature under standard safe moxibustion stimulation collected from a large number of healthy subjects in previous clinical trials. The initial stage of heat stress is the initial time period during which the user's skin exhibits a clear heat stress response after moxibustion is initiated. The collected rate of change in skin surface temperature is compared with the preset initial stress threshold. If the rate of change remains above the threshold, the time period is defined as the initial stage of heat stress. For example, if the preset initial stress threshold is 0.5℃ / min, and user A's skin surface temperature change rate is 0.8℃ / min for 3 minutes, which is higher than 0.5℃ / min, then 3 minutes is defined as the initial stage of heat stress.

[0020] Secondly, after the initial stage of heat stress ends, when the heart rate variability index falls into the steady-state regulation range and the real-time temperature of the moxibustion head remains stable, it is divided into the heat effect duration stage.

[0021] The steps for obtaining the steady-state adjustment range include: Obtain the initial heart rate variability index set continuously monitored by the target user in a non-moxibustion resting state, calculate the median value, and use it as the individual resting baseline value. From the user's historical moxibustion reaction records, all treatment periods marked as effective and without adverse reactions were screened, and the heart rate variability index during the stable period within the period was extracted to form a set of historical effective data. Calculate the upper and lower quartiles of the historical valid data set, use the individual resting baseline value as the central reference point, and determine the steady-state adjustment range based on the upper and lower quartiles.

[0022] First, the initial heart rate variability (HRV) index set continuously monitored by the target user in a non-moxibustion resting state is obtained, and the median value is calculated as the individual resting baseline value. The individual resting baseline value is the median of the initial HRV index set continuously monitored by the user in a non-moxibustion resting state, reflecting the user's basic physiological state without external stimulation. Continuous HRV indexes of the user in a non-moxibustion resting state are collected to form an initial HRV index set, and the median of this set is calculated as the individual resting baseline value. For example, if 10 sets of HRV indexes are collected from user A in a non-moxibustion resting state: 35ms, 38ms, 40ms, 42ms, 45ms, 46ms, 48ms, 50ms, 52ms, and 55ms, the median is 45.5ms, which is user A's individual resting baseline value.

[0023] Secondly, from the user's historical moxibustion response records, all treatment periods marked as effective and without adverse reactions are screened, and the heart rate variability index during the stable period within each period is extracted to form a historical effective data set. The user's historical moxibustion response records are complete records storing the user's physiological data, equipment operation data, moxibustion effects, and whether any adverse reactions occurred during each past moxibustion session. An effective treatment period without adverse reactions refers to a complete moxibustion period in the user's historical moxibustion history where the moxibustion effect was marked as effective and no adverse reactions occurred throughout. A stable period refers to a time within the effective treatment period where the user's physiological state tends to be stable without significant fluctuations. The historical effective data set is a data set formed by summarizing the heart rate variability index during the stable period of all effective and adverse reaction-free treatment periods.

[0024] For example, User A has 5 historical moxibustion records. Three effective treatment periods without adverse reactions are selected and denoted as Period 1, Period 2, and Period 3: Period 1 stable period heart rate variability index: 43ms, 44ms, 45ms; Period 2 stable period heart rate variability index: 46ms, 47ms, 45ms; Period 3 stable period heart rate variability index: 44ms, 46ms, 48ms; Summarizing all the above indicators, the historical effective data set for User A is: {43ms, 44ms, 45ms, 46ms, 47ms, 45ms, 44ms, 46ms, 48ms}.

[0025] Next, the upper and lower quartiles of the historical valid data set are calculated, and the personal resting baseline value is used as the central reference point to determine the steady-state adjustment range based on the upper and lower quartiles.

[0026] First, calculate the upper and lower quartiles of the historical valid data set. The upper quartile refers to the values ​​in the top 75% after sorting the data set from smallest to largest; that is, 75% of the data are less than or equal to this value, and 25% are greater than it. The lower quartile refers to the values ​​in the top 25% after sorting the data set from smallest to largest; that is, 25% of the data are less than or equal to this value, and 75% are greater than it. Sort all indicators in the historical valid data set in ascending order; calculate the quartile index using the formula: position = (n+1)×p, where n is the number of data points, and p is 0.25 or 0.75, thus determining the values ​​corresponding to the top 25% and top 75% positions in the sorted set.

[0027] For example, sorting user A's historical valid data set from smallest to largest, we get: 43ms, 44ms, 44ms, 45ms, 45ms, 46ms, 46ms, 47ms, 48ms; the total number of data is 9. The lower quartile position is (9+1)×25%=2.5, and we take the average of the 2nd and 3rd values, i.e., (44ms+44ms) / 2=44ms; the upper quartile position is (9+1)×75%=7.5, and we take the average of the 7th and 8th values, i.e., (46ms+47ms) / 2=46.5ms; therefore, the lower quartile of this historical valid data set is 44ms, and the upper quartile is 46.5ms.

[0028] Secondly, using the individual resting baseline value as the central reference point, the steady-state regulation interval is determined based on the upper and lower quartiles. The steady-state regulation interval refers to the range of values ​​within which the user's heart rate variability index tends to stabilize and autonomic nervous system regulation is normal during moxibustion, determined by using the individual resting baseline value as the central reference point and combining the upper and lower quartiles of historical effective data sets. This range is used to determine whether the user has entered the sustained thermal effect phase. Using the individual resting baseline value as the central reference point, the lower quartile is used as the lower limit of the interval, and the upper quartile as the upper limit. If the individual resting baseline value deviates from the range formed by the upper and lower quartiles, the range of the upper and lower quartiles is used to ensure that the interval closely matches the user's historical effective physiological state.

[0029] For example, user A's personal resting baseline is 45.5ms, the lower quartile of the historical valid data set is 44ms, and the upper quartile is 46.5ms; with 45.5ms as the central reference point, the steady-state adjustment range is determined to be 44ms~46.5ms.

[0030] Furthermore, after the initial heat stress phase, when the heart rate variability index falls within the steady-state regulation range and the real-time temperature of the moxibustion head remains stable, the period is designated as the sustained heat effect phase. The sustained heat effect phase refers to the core moxibustion stage after the initial heat stress phase, when the user's physiological state enters a stable regulation phase and the moxibustion heat effect continues to penetrate and exert its effects. At this time, the user experiences no significant discomfort, and the moxibustion effect is optimal. After the initial heat stress phase, the user's heart rate variability index and the real-time temperature of the moxibustion head are continuously monitored. When the heart rate variability index consistently falls within the steady-state regulation range and the real-time temperature of the moxibustion head remains stable—that is, when the real-time temperature of the moxibustion head does not fluctuate significantly (with a fluctuation range not exceeding ±1℃)—this period is designated as the sustained heat effect phase.

[0031] For example, after the initial 3-minute heat stress phase for user A, continuous monitoring revealed that the heart rate variability index stabilized between 45ms and 46ms, falling into the steady-state regulation range of 44ms to 46.5ms, and the real-time temperature of the moxibustion head stabilized at 45℃ with a fluctuation range of ±0.5℃. Therefore, starting from the end of the 3-minute phase, the heat effect duration phase was defined until the current cumulative working time reached the personalized phase duration benchmark.

[0032] Finally, when the current cumulative working time reaches the personalized stage duration benchmark generated based on the user's historical moxibustion response records, it is designated as the treatment completion preparation stage. The personalized stage duration benchmark refers to selecting all treatment records marked as safely completed and effective from the user's historical moxibustion response records, extracting the current cumulative working time corresponding to each record to form a historical safe working time set, and calculating the median of this set as the user's personalized treatment stage duration threshold. The treatment completion preparation stage refers to the stage where, when the current cumulative working time reaches the personalized stage duration benchmark, the moxibustion treatment is about to end, and is used to gradually reduce the moxibustion temperature and prepare for stopping moxibustion.

[0033] Specifically, from the user's historical moxibustion response records, all treatment records marked as safely completed and effective are screened, and the current cumulative working time corresponding to each record is extracted to form a set of historical safe working times; the median of this set is calculated, which is the personalization stage duration benchmark; the current cumulative working time of the moxibustion device is monitored in real time, and when the current cumulative working time reaches the personalization stage duration benchmark, the time period is divided into the treatment completion preparation stage.

[0034] For example, User A's historical safe working time set is: 22min, 24min, 25min, 26min, 28min; the median of this set is 25min, so User A's personalized stage duration benchmark is 25min; when User A's current cumulative working time for this moxibustion reaches 25min, the current stage is divided into the treatment end preparation stage, and preparations for stopping moxibustion begin.

[0035] In this embodiment of the invention, by collecting multi-dimensional physiological and equipment operation data, combined with the user's personal basic physiological data and historical moxibustion data, dynamic and personalized division of moxibustion stages is achieved. This abandons the traditional subjective division method, accurately matches the user's real-time physiological state and moxibustion process, and provides accurate stage basis for subsequent calculation of dynamic heat tolerance threshold and generation of optimal control parameters, thereby improving the objectivity and adaptability of moxibustion process control.

[0036] S200: Extract key biomarkers within the current moxibustion stage and combine them with the user's historical moxibustion response records to calculate a personalized dynamic threshold for heat tolerance.

[0037] In this embodiment of the invention, key biometrics within the current moxibustion stage are extracted and combined with the user's historical moxibustion response records to calculate a personalized dynamic threshold for heat tolerance. Different users exhibit significant individual differences in skin heat tolerance and physiological regulation capabilities. Traditional moxibustion relies solely on a uniform tolerance standard to determine safety boundaries, which cannot adapt to the user's real-time physiological state and historical tolerance patterns, easily leading to deviations in tolerance threshold determination and thus risks such as burns and inflammation. Based on the dynamic division of moxibustion stages, key biometrics of the heat effect duration stage need to be extracted. Personalized warning lines and tolerance baselines are constructed by combining these with the user's historical moxibustion safety data. The thresholds are then dynamically reduced based on real-time physiological states to achieve adaptive tightening of the safety boundaries, providing a precise personalized safety benchmark for subsequent parameter input.

[0038] Step S200 in the method provided in this embodiment of the invention includes: During the duration of the thermal effect, the rate of change of skin surface temperature and heart rate variability were continuously monitored. The instantaneous physiological load index is generated by weighted fusion calculation based on the ratio of the skin surface temperature change rate to the historical mean at the lower limit of the steady-state regulation range, and the ratio of the heart rate variability index to the individual resting baseline value. From the user's historical moxibustion response records, query the historical valid records that are closest to the current real-time temperature of the moxibustion head, and extract the average skin surface temperature change rate and average heart rate variability index before reaching the preset warning line during the duration of the thermal effect, and calculate the individual's historical tolerance baseline. The steps for determining the preset warning line include: From the user's historical moxibustion response records, all treatment records marked as safely completed and effective are selected to form a safe record set; For each record in the safety record set, extract the skin surface temperature change rate and heart rate variability index of the last steady state during the thermal effect duration and before the end of the recording to form a safety critical feature set; Calculate the upper quantile of the skin surface temperature change rate data for all skin surface temperature change rate data in the safety critical feature set, and use it as the warning line for skin surface temperature change rate. Calculate the specified upper quantile for all heart rate variability index data, and use it as the warning line for heart rate variability index; The preset warning line is composed of the skin surface temperature change rate warning line and the heart rate variability index warning line. The instantaneous physiological load index is compared with the personal historical tolerance baseline. Taking into account the percentage of completion of the current cumulative working hours relative to the personalization phase duration benchmark, and the degree to which the instantaneous physiological load index approaches the personal historical tolerance baseline, the personal historical tolerance baseline is dynamically reduced. The reduced result is output as the dynamic threshold of heat tolerance.

[0039] First, during the duration of the thermal effect, the rate of change of skin surface temperature and heart rate variability are continuously monitored. Within the thermal effect duration defined by S100, flexible sensors and wearable devices continuously collect data on the rate of change of skin surface temperature and heart rate variability at the user's acupuncture site, ensuring real-time and continuous data collection. For example, after user A enters the thermal effect duration phase, continuous monitoring shows a skin surface temperature change rate of 0.6℃ / min and a heart rate variability of 45ms.

[0040] Secondly, a weighted fusion calculation is performed based on the ratio of the skin surface temperature change rate to the historical mean at the lower limit of the steady-state regulation range, and the ratio of the heart rate variability index to the individual resting baseline value, to generate the instantaneous physiological load index. The instantaneous physiological load index is a comprehensive indicator used to quantify the user's current real-time physiological load, obtained by weighted fusion of the ratios of the skin surface temperature change rate and the heart rate variability index. The historical mean at the lower limit of the steady-state regulation range refers to the average value of the heart rate variability index equal to the lower limit of the steady-state regulation range in the user's historical valid data set. The instantaneous physiological load index is obtained by calculating the ratio of the skin surface temperature change rate to the historical mean at the lower limit of the steady-state regulation range; calculating the ratio of the heart rate variability index to the individual resting baseline value; and performing a weighted fusion calculation on the two ratios.

[0041] For example, user A's steady-state regulation interval lower limit is 44ms, the historical average skin surface temperature change rate is 0.4℃ / min, and the personal resting baseline is 45.5ms; the current skin surface temperature change rate is 0.6℃ / min, 0.6 / 0.4=1.5; the current heart rate variability index is 45ms, 45 / 45.5≈0.99; the weights are set as follows: skin surface temperature change rate accounts for 0.6, heart rate variability index accounts for 0.4; the instantaneous physiological load index P=1.5×0.6+0.99×0.4=1.296.

[0042] Next, from the user's historical moxibustion response records, query the historical valid records that are closest to the current real-time temperature of the moxibustion head, and extract the average skin surface temperature change rate and average heart rate variability index before reaching the preset warning line during the duration of the thermal effect, and calculate the individual's historical tolerance baseline.

[0043] The steps for determining the preset warning line include: First, from the user's historical moxibustion reaction records, all records marked as safely completed and effective are selected to form a safe record set. The safe record set refers to the dataset composed of all records marked as safely completed and effective from the user's historical moxibustion reaction records. The user's historical moxibustion reaction records are traversed, and records with adverse reactions such as burns or fever, or those that are ineffective, are removed. Only records marked as safely completed and effective are retained, and these are summarized to form the safe record set. For example, if user A has 9 historical moxibustion records, after removing 2 ineffective records and 1 record showing fever, the remaining 5 safe and effective records constitute the safe record set.

[0044] Secondly, for each record in the safety record set, the skin surface temperature change rate and heart rate variability index of the last steady-state data during the duration of the thermal effect and before the end of the recording are extracted to form a safety critical feature set. The safety critical feature set refers to the set of features composed of the skin surface temperature change rate and heart rate variability index of the last steady-state data during the duration of the thermal effect and before the end of the recording in the safety record. For each record in the safety record set, the last steady-state data during the duration of the thermal effect and before the end of the recording are located, and the corresponding skin surface temperature change rate and heart rate variability index are extracted and summarized to form the safety critical feature set.

[0045] For example, from user A's 5 safety records, the skin surface temperature change rate and heart rate variability index of the last steady state before the end of the recording during the thermal effect phase are extracted to form a set of safety critical features, such as: skin surface temperature change rate: 0.68, 0.7, 0.72, 0.73, 0.75 (℃ / min) and heart rate variability index: 45, 46, 46, 47, 48 (ms).

[0046] In addition, the upper quartile of all skin surface temperature change rate data in the safety critical feature set is calculated as the warning line for skin surface temperature change rate. The upper quartile is the percentile used to determine the warning threshold value, such as the 95th percentile. The warning line for skin surface temperature change rate is the 95th percentile value of the skin surface temperature change rate data in the safety critical feature set, which is the safety warning threshold value for this indicator. All skin surface temperature change rate data in the safety critical feature set are extracted, sorted, and the 95th percentile value is calculated as the warning line for skin surface temperature change rate. The calculation method is the same as that for the upper and lower quartiles in S100. For example, user A's skin surface temperature change rate data: 0.68, 0.7, 0.72, 0.73, 0.75 (℃ / min), the 95th percentile value position = (5+1)×0.95=5.7, that is, the warning line for the user's skin surface temperature change rate = (0.73+0.75) / 2=0.745℃ / min.

[0047] Next, calculate the upper quantile of all heart rate variability (HRV) data as the HRV warning line. The HRV warning line is the 95th percentile value of the HRV data in the safety threshold feature set, which is the safety warning threshold for this indicator. Extract all HRV data from the safety threshold feature set, sort them, and calculate the 95th percentile value as the HRV warning line. For example, if user A's HRV data is 45, 46, 46, 47, 48 (ms), the warning line for this user's HRV is (47+48) / 2 = 47.5ms.

[0048] Furthermore, the preset warning line is jointly formed by the skin surface temperature change rate warning line and the heart rate variability index warning line. The preset warning line is a composite safety warning boundary composed of the skin surface temperature change rate warning line and the heart rate variability index warning line; exceeding either line indicates an excessive physiological load. The preset warning line is formed by combining the skin surface temperature change rate warning line and the heart rate variability index warning line. For example, user A's preset warning line is: skin surface temperature change rate ≥ 0.745℃ / min, or heart rate variability index ≥ 47.5ms.

[0049] Based on this, from the user's historical moxibustion response records, the historical valid records that are closest to the current real-time temperature of the moxibustion head are queried, and the average skin surface temperature change rate and average heart rate variability index before reaching the preset warning line during the duration of the thermal effect are extracted accordingly, and the personal historical tolerance baseline is calculated.

[0050] Specifically, the average skin surface temperature change rate and average heart rate variability index are extracted during the duration of the thermal effect before reaching the preset warning line, and an individual's historical tolerance baseline is calculated, including: Based on the retrieved historical records, the time point when the rate of change of skin surface temperature first continuously exceeds the warning line of the rate of change of skin surface temperature or the time point when the heart rate variability index first continuously exceeds the warning line of the heart rate variability index is located during the duration of the thermal effect is used as the trigger point of the preset warning line. Extract the skin surface temperature change rate data sequence and the heart rate variability index data sequence within a preset time period before the trigger point; Calculate the arithmetic mean of the skin surface temperature change rate data sequence as the average skin surface temperature change rate; Calculate the arithmetic mean of the heart rate variability index data sequence as the mean heart rate variability index; The average skin surface temperature change rate and the average heart rate variability index are normalized and weighted and summed, and the calculation results are output as the individual's historical tolerance baseline.

[0051] First, based on the retrieved historical records, the time point when the rate of change of skin surface temperature first consistently exceeds the warning line for the rate of change of skin surface temperature, or the time point when the heart rate variability index first consistently exceeds the warning line for the heart rate variability index, is identified as the trigger point for the preset warning line. The preset warning line trigger point is the time point when either the rate of change of skin surface temperature first consistently exceeds the warning line, or the time point when the heart rate variability index first consistently exceeds the warning line, occurs within the duration of the thermal effect. From the user's historical moxibustion response records, the historical records whose real-time temperature of the moxibustion head is closest to the current real-time temperature of the moxibustion head are retrieved. Within the matched historical records, the time point when any indicator first consistently exceeds the preset warning line within the duration of the thermal effect is identified as the trigger point.

[0052] For example, user A's current real-time temperature of the moxibustion head is 45℃. A search of historical records reveals a valid historical record with a temperature of 45.2℃, which is a match. In user A's matched historical records, the rate of change of skin surface temperature first continuously exceeds 0.745℃ / min at the 18th minute. This 18-minute period is the preset warning line trigger point.

[0053] Secondly, extract the skin surface temperature change rate data sequence and heart rate variability index data sequence within a preset time period before the trigger point. The preset time period is a statistical value based on the user's historical valid records, representing the typical stable time before reaching the warning line, such as 2 minutes. Starting from the trigger point and counting backwards for the preset time period, extract the skin surface temperature change rate data sequence and heart rate variability index data sequence within that time period. For example, if the trigger point is 18 minutes, extract the data sequence from 16 minutes to 18 minutes, measuring every 1 minute: Skin surface temperature change rate sequence: 0.7℃ / min, 0.71℃ / min, 0.72℃ / min; Heart rate variability index sequence: 46ms, 46.5ms, 47ms.

[0054] Next, the arithmetic mean of the skin surface temperature change rate data sequence is calculated as the average skin surface temperature change rate. The arithmetic mean of the skin surface temperature change rate data sequence is used to obtain the average skin surface temperature change rate. For example, the average skin surface temperature change rate = (0.7 + 0.71 + 0.72) / 3 = 0.71℃ / min.

[0055] Then, the arithmetic mean of the heart rate variability index data sequence is calculated as the mean heart rate variability index. The arithmetic mean of the heart rate variability index data sequence is calculated to obtain the mean heart rate variability index. For example, the mean heart rate variability index = (46 + 46.5 + 47) / 3 = 46.5 ms.

[0056] Subsequently, the average skin surface temperature change rate and the average heart rate variability index are normalized and weighted, and the calculation results are output as the personal historical tolerance baseline. The personal historical tolerance baseline is a basic heat tolerance reference value obtained by normalizing and weighting the user's historical safety data. Normalization refers to mapping the two types of indicators to a unified numerical range to eliminate dimensional differences. Weights are assigned according to the correlation between the two types of indicators and heat tolerance. After normalizing the average skin surface temperature change rate and the average heart rate variability index, the results are weighted and summed to output the personal historical tolerance baseline B.

[0057] The normalization process uses Min-Max normalization, with the formula: Normalized value = (Current indicator value - Historical minimum value of the indicator) / (Historical maximum value of the indicator - Historical minimum value of the indicator), where the historical minimum and maximum values ​​are taken from the extreme values ​​of the corresponding indicators in the user's safety critical feature set. The closer the normalization result is to 1, the closer it is to the user's historical safety tolerance upper limit. For example, the normalized average skin surface temperature change rate = (0.71-0.68) / (0.75-0.68)≈0.43, the normalized average heart rate variability index = (46.5-45) / (48-45)=0.5; the weight of the skin surface temperature change rate = 0.6, the weight of the heart rate variability index = 0.4; the personal historical tolerance baseline B = 0.43×0.6+0.5×0.4=0.458.

[0058] Finally, the immediate physiological load index is compared with the personal historical tolerance baseline. Combining the current cumulative working time as a percentage of completion relative to the personalized stage duration baseline, and the degree to which the immediate physiological load index approaches the personal historical tolerance baseline, the personal historical tolerance baseline is dynamically reduced. The reduced result is output as the dynamic heat tolerance threshold. The completion percentage is the proportion of the current cumulative working time to the personalized stage duration baseline. The degree to which the immediate physiological load index approaches the personal historical tolerance baseline is calculated by the ratio R of the immediate physiological load index P to the personal historical tolerance baseline B, and quantified according to the following rules: when R≤1, the degree of proximity is 1; when R>1, the degree of proximity is 1 / R. The dynamic heat tolerance threshold is a personalized safety threshold dynamically reduced in real time, gradually tightening as the moxibustion process progresses. Dynamic reduction formula: Dynamic heat tolerance threshold = Personal historical tolerance baseline × [1 - Completion percentage × (1 - Degree of proximity)].

[0059] For example, user A's current cumulative working time is 20 minutes, and the baseline time for the personalized phase is 25 minutes; completion percentage = 20 / 25 = 0.8; R = 1.296 / 0.458 ≈ 2.82 > 1; proximity = 1 / 2.82 ≈ 0.35; dynamic threshold for heat tolerance = 0.458 × [1 - 0.8 × (1 - 0.35)] ≈ 0.22.

[0060] In this embodiment of the invention, key biometric features of the sustained thermal effect stage are extracted, and personalized preset warning lines and personal historical tolerance baselines are constructed by combining the user's historical moxibustion safety data. Then, the tolerance baseline is dynamically reduced by real-time physiological load and the completion of the moxibustion process to generate a gradually tightening dynamic threshold for thermal tolerance. This achieves personalized, dynamic, and quantitative thermal tolerance standards, abandons the drawbacks of traditional uniform tolerance thresholds, adapts to the user's real-time physiological state and historical tolerance patterns, effectively reduces the risk of burns and inflammation, and provides accurate personalized input parameters for subsequent adaptive adjustment models.

[0061] S300: Input the key biofeedback and the dynamic threshold of heat tolerance into the pre-trained adaptive adjustment model, and output the optimal temperature-duration control parameter set.

[0062] In this embodiment of the invention, the key biometrics and dynamic heat tolerance threshold are input into a pre-trained adaptive adjustment model, which outputs the optimal temperature-duration control parameter set. Traditional moxibustion temperature and duration control relies on fixed empirical values, failing to dynamically adjust based on the user's real-time key biometrics and dynamic heat tolerance threshold. This easily leads to safety risks due to excessively high temperatures or durations, or ineffective treatment due to excessively low temperatures or insufficient durations. Therefore, it is necessary to train the adaptive adjustment model using historical safe and effective data, allowing the model to learn the optimal control rules for achieving therapeutic effects without exceeding the tolerance threshold, thereby outputting a temperature-duration control parameter set adapted to the user's real-time condition.

[0063] The steps for building the pre-trained adaptive adjustment model are as follows: First, historical moxibustion records were collected to form a sample dataset. Multiple sets of historical moxibustion records from the target user were collected, and all records marked as safe and effective were selected. For each set of records, the instantaneous physiological load index, dynamic heat tolerance threshold, real-time temperature of the moxibustion head, and current cumulative working time were extracted according to time series during the duration of the thermal effect. Simultaneously, the temperature adjustment value and duration adjustment value ultimately marked as safe and effective for that set of records were extracted. All of the above data were then aggregated to form the sample dataset.

[0064] For example, taking user A's historical records as an example, a single sample data point is as follows: Time-series input data: Instantaneous physiological load index sequence: 1.05, 1.12, 1.296, 1.35...; Dynamic heat tolerance threshold sequence: 0.48, 0.42, 0.22, 0.18...; Real-time temperature of moxibustion head: 44℃, 44.5℃, 45℃, 45.2℃...; Current cumulative working time: 15min, 16min, 20min, 21min.... Safe and effective adjustment values: Temperature adjustment value: -0.5℃; Duration adjustment value: -3min. Summarizing 1000 such safe and effective records from user A constitutes a sample data set.

[0065] Secondly, iterate through each record in the sample dataset; for each record's time-series input data, combine its final safe and effective temperature adjustment value and duration adjustment value into a tuple: (temperature adjustment value, duration adjustment value); use this tuple as the target control parameter group label for that record to complete the labeling of the entire sample.

[0066] For example, for a single sample of user A: input time-series data: instantaneous physiological load index of 1.296, dynamic threshold of heat tolerance of 0.22, and moxibustion head temperature of 45℃ at 20 min; label the target regulation parameter group: (-0.5℃, -3min), indicating that the temperature needs to be reduced by 0.5℃ and the duration shortened by 3 min; after completing the labeling of 1000 samples of user A, each sample corresponds to a unique target regulation parameter group label.

[0067] Next, an adaptive regulation model network architecture based on deep reinforcement learning is constructed. The feature extraction layer uses an LSTM network to process the input time-series data, such as the immediate physiological load index and the dynamic threshold of heat tolerance, to extract dynamic features in the time dimension. The policy decision layer uses a fully connected network to map the high-dimensional features output by the feature extraction layer into candidate policies for temperature adjustment and duration adjustment. The safety constraint layer introduces the dynamic threshold of heat tolerance as a constraint condition to filter out candidate policies whose immediate physiological load index exceeds the dynamic threshold of heat tolerance after prediction and regulation. The three-layer structure is integrated to form a complete adaptive regulation model network architecture, with the goal of outputting the optimal regulation parameters that achieve therapeutic effects without exceeding the safety threshold.

[0068] For example, an adaptive adjustment model network architecture is constructed for user A: Feature extraction layer: 2-layer LSTM with an input dimension of 4, corresponding to the immediate physiological load index, dynamic heat tolerance threshold, moxibustion head temperature, and cumulative duration, with a hidden layer dimension of 64; Policy decision layer: 2-layer fully connected layers with an output dimension of 2, corresponding to the temperature adjustment value and duration adjustment value; Safety constraint layer: embedding the S200 dynamic heat tolerance threshold formula to verify the safety of candidate strategies in real time; The overall network aims to maximize therapeutic efficacy while minimizing safety risks.

[0069] Subsequently, supervised training was performed on the adaptive adjustment model network architecture. The labeled sample dataset was divided into training, validation, and test sets in a 7:2:1 ratio. The time-series input data from the training set was fed into the adaptive adjustment model, and the model output the predicted set of control parameters. The error between the predicted control parameter set and the target control parameter set labels was calculated using the mean squared error (MSE). The weight parameters of the model network were updated using the backpropagation algorithm to minimize the prediction error. The above process was iterated until the model's error on the training set tended to stabilize.

[0070] For example, a dataset of 1000 samples is divided into 700 training sets, 200 validation sets, and 100 test sets. The time-series input data of the training set is input into an adaptive adjustment model, and the model outputs the predicted set of control parameters. The error between the predicted control parameter set and the target control parameter set is calculated using the mean squared error (MSE). The weight parameters of the model network are updated using the backpropagation algorithm to minimize the prediction error. For example, after 1000 iterations, the prediction result gradually approaches the correct value, and the training set error drops below 0.01.

[0071] Finally, training is stopped and the model pre-training is completed when the combined error between the control parameter set generated by the model on the independent validation set and the target control parameter set label is lower than the preset convergence threshold, and the probability that the predicted immediate physiological load index exceeds the dynamic threshold of heat tolerance after simulation application is lower than the safety risk threshold. The input data from the independent validation set is then input into the trained model to obtain the predicted control parameter set; the combined error between the predicted parameter set and the target label is calculated to determine if it is lower than the preset convergence threshold, such as 0.05; the predicted control parameters are simulated and applied to the moxibustion process, and the probability that the immediate physiological load index exceeds the dynamic threshold of heat tolerance is calculated to determine if it is lower than the safety risk threshold, such as 5%; if both conditions are met, training is stopped and the model pre-training is completed; if not, it returns to continue iterative training.

[0072] For example, the preset convergence threshold is 0.05, the safety risk threshold is 5%, the comprehensive MSE error of the 200 samples in the validation set is 0.03 < 0.05, after simulating the application of the prediction parameters, only 1 sample has a physiological load index close to the threshold, and the probability of exceeding the threshold is 0% < 5%; if both conditions are met, training is stopped and the pre-training of the user A's exclusive adaptive adjustment model is completed.

[0073] Then, the key biometrics and dynamic heat tolerance thresholds are input into a pre-trained adaptive adjustment model, which outputs the optimal temperature-duration control parameter set. For example, given the key biometrics and dynamic heat tolerance thresholds of target user A from S200, the following data is used as input vectors and fed into the adaptive adjustment model customized for user A: [Instantaneous physiological load index = 1.296, dynamic heat tolerance threshold = 0.22, skin surface temperature change rate = 0.71℃ / min, heart rate variability index = 46.5ms, real-time temperature of moxibustion head = 45℃, current cumulative duration = 20min]. The adaptive adjustment model outputs the optimal control parameter set: (temperature adjustment value = -0.5℃, duration adjustment value = -3min).

[0074] This step involves constructing a sample set by collecting historical safe and effective moxibustion data from users, labeling it with target regulation tags, building and training an adaptive regulation model based on deep reinforcement learning, and completing model pre-training by combining convergence thresholds and safety risk thresholds. This model can learn the temperature-duration regulation pattern of users achieving "therapeutic effect target + not exceeding heat tolerance threshold," laying the foundation for outputting the optimal regulation parameter set after inputting real-time key biometrics and dynamic heat tolerance thresholds. This ensures the effectiveness of moxibustion while mitigating safety risks such as burns and internal heat at the model level, achieving personalized and adaptive moxibustion regulation.

[0075] S400: Based on the optimal temperature-duration control parameter set, generate a personalized moxibustion guidance plan that includes a phased temperature control curve, a suggested duration for a single acupoint, and key points for real-time biofeedback monitoring, and drive the moxibustion device to perform preheating.

[0076] In this embodiment of the invention, a personalized moxibustion guidance plan is generated based on the optimal temperature-duration control parameter set. This plan includes a phased temperature control curve, a suggested duration for each acupoint, and key points for real-time biofeedback monitoring. The moxibustion device is then driven to perform preheating. Simply obtaining the optimal temperature-duration control parameter set is insufficient to directly guide home moxibustion practice. Traditional moxibustion lacks phased temperature control, duration allocation according to acupoint characteristics, and a real-time risk monitoring mechanism, easily leading to problems such as abrupt temperature control, uneven acupoint stimulation, and failure to provide timely risk warnings. Therefore, it is necessary to analyze the optimal parameters to form a phased temperature control curve, allocate duration for each acupoint based on its physiological characteristics, set key points for biofeedback monitoring, and ultimately generate a personalized moxibustion guidance plan that can be directly implemented.

[0077] Step S400 in the method provided in this embodiment of the invention includes: The optimal temperature-duration control parameter set is analyzed to obtain the recommended target temperature and remaining suggested duration for the duration of the thermal effect. Based on the recommended target temperature, the duration of the initial stage of thermal stress, and the steady-state adjustment range, a staged temperature control curve is generated. The remaining suggested duration is allocated proportionally according to the muscle thickness and sensitivity of the acupoints to be treated, and the suggested duration for each acupoint is calculated. The biofeedback monitoring points for triggering real-time alarms and pausing moxibustion are set as follows: the immediate physiological load index continuously exceeds the dynamic threshold of heat tolerance, or the skin surface temperature change rate suddenly increases beyond the preset change rate threshold. By summarizing the phased temperature control curves, the recommended duration for each acupoint, and the key points of real-time biofeedback monitoring, and combining the recommended target temperature with the remaining recommended duration, a personalized moxibustion guidance plan is generated.

[0078] First, the optimal temperature-duration control parameter set is analyzed to obtain the recommended target temperature and remaining suggested duration for the duration of the thermal effect. The recommended target temperature refers to the safe and effective steady-state temperature that needs to be maintained during the duration of the thermal effect after adjusting the optimal parameters. The remaining suggested duration refers to the actual moxibustion time remaining during the duration of the thermal effect after deducting the adjustment amount. The recommended target temperature is obtained by calculating the temperature adjustment value in the optimal temperature-duration control parameter set with the current real-time temperature; the remaining suggested duration is obtained by subtracting the duration adjustment value from the original remaining duration.

[0079] For example, User A's current real-time temperature of the moxibustion head is 45℃, the optimal temperature adjustment value is -0.5℃, and the recommended target temperature is 45-0.5=44.5℃; User A's current cumulative time is 20 minutes, the personalization stage time base is 25 minutes, the original remaining time is 5 minutes, the optimal time adjustment value is -3 minutes, and the remaining suggested time is 5-3=2 minutes.

[0080] Secondly, based on the recommended target temperature, the duration of the initial stage of thermal stress, and the steady-state adjustment range, a staged temperature control curve is generated.

[0081] Specifically, based on the recommended target temperature, the duration of the initial stage of thermal stress, and the steady-state adjustment range, a staged temperature control curve is generated, including: Starting from the initial temperature of moxibustion and taking the duration of the initial stage of heat stress as the total duration, calculate the average heating rate required to linearly increase the temperature from the initial temperature to the recommended target temperature. Using the recommended target temperature as the steady-state temperature and the remaining suggested duration as the duration of the isothermal phase, an isothermal control segment with a constant temperature is generated. Set a fixed cooling time and a cooling endpoint temperature, and calculate the average cooling rate required to linearly decrease from the recommended target temperature to the cooling endpoint temperature. The linear heating section, the constant temperature control section, and the linear cooling section are connected in chronological order to form a staged temperature control curve.

[0082] First, starting from the initial temperature of moxibustion and taking the duration of the initial heat stress phase as the total duration, calculate the average heating rate required for the temperature to rise linearly from the initial temperature to the recommended target temperature. The linear heating phase refers to the temperature control phase where the temperature rises at a constant rate from the initial moxibustion temperature to the recommended target temperature over the initial heat stress phase. The average heating rate refers to the temperature increase per minute within the linear heating phase. Average heating rate = (Recommended target temperature - Initial moxibustion temperature) / Duration of the initial heat stress phase. For example, if User A's initial moxibustion temperature is 37°C (human skin temperature), the initial heat stress phase lasts 3 minutes, and the recommended target temperature is 44.5°C, the average heating rate = (44.5 - 37) / 3 = 2.5°C / min.

[0083] Secondly, using the recommended target temperature as the steady-state temperature and the remaining suggested duration as the duration of the constant-temperature phase, a constant-temperature control segment is generated where the temperature remains constant. A constant-temperature control segment refers to a temperature control phase where the recommended target temperature is the steady-state temperature, the remaining suggested duration is the duration, and the temperature remains constant. The recommended target temperature is locked, and the duration is set as the remaining suggested duration, thus forming the constant-temperature control segment. For example, User A's constant-temperature control segment: steady-state temperature 44.5℃, duration 2 minutes.

[0084] Next, set a fixed cooling time and a fixed endpoint temperature, and calculate the average cooling rate required for a linear decrease from the recommended target temperature to the endpoint temperature. The linear cooling phase refers to the temperature control stage where the temperature drops uniformly from the recommended target temperature to the endpoint temperature. The average cooling rate refers to the temperature decrease per minute within the linear cooling phase. Average cooling rate = (Recommended target temperature - Ending temperature) / Fixed cooling time. For example, if user A sets a fixed cooling time of 1 minute and an endpoint temperature of 37°C, the average cooling rate = (44.5 - 37) / 1 = 7.5°C / min.

[0085] Finally, the linear heating segment, the isothermal control segment, and the linear cooling segment are connected in chronological order to form a staged temperature control curve. A staged temperature control curve is a continuous temperature control curve formed by splicing the linear heating segment, the isothermal control segment, and the linear cooling segment in chronological order. The parameters of each stage are integrated according to the chronological order of the linear heating segment, the isothermal control segment, and the linear cooling segment to form a complete temperature control curve. For example, User A's staged temperature control curve is as follows: Linear heating segment: 0~3 min, temperature rises from 37℃ to 44.5℃ at a rate of 2.5℃ / min; Isothermal control segment: 3~5 min, temperature remains constant at 44.5℃ for 2 min; Linear cooling segment: 5~6 min, temperature drops from 44.5℃ to 37℃ at a rate of 7.5℃ / min.

[0086] Then, the remaining suggested duration is proportionally allocated according to the muscle thickness and sensitivity of the acupoints to be treated, and the suggested duration for each acupoint is calculated.

[0087] The remaining suggested duration is allocated proportionally based on the muscle thickness and sensitivity of the acupoints to be treated, and the suggested duration for each acupoint is calculated, including: For each acupoint to be treated, the data sequence of skin surface temperature change rate during the heat effect sustained stage of the corresponding acupoint is extracted from the user's historical moxibustion response records. The variance of the sequence is calculated, and the reciprocal of the variance is normalized to obtain a stabilizing factor characterizing the degree of muscle thickness. From the user's historical moxibustion response records, the skin surface temperature change rate data of the corresponding acupoints in the initial stage of heat stress during historical treatment are extracted, the arithmetic mean is calculated and normalized to obtain the sensitivity factor characterizing personalized heat sensitivity. The stabilizing factor and the sensitive factor are weighted and summed to calculate the comprehensive allocation coefficient of the corresponding acupoint. The comprehensive allocation coefficients of all acupoints are then combined to form a coefficient set. Based on the proportion of the comprehensive allocation coefficient of each acupoint in the coefficient set to the total, the remaining suggested duration is allocated to obtain the single-acupoint suggested duration for each acupoint.

[0088] First, for each acupoint scheduled for moxibustion, the data sequence of skin surface temperature change rate during the duration of the thermal effect at the corresponding acupoint is extracted from the user's historical moxibustion response records. The variance of the sequence is calculated, and the reciprocal of the variance is normalized to obtain a stabilizing factor characterizing the degree of muscle richness. The stabilizing factor is an indicator of the degree of muscle richness at the acupoint; the higher the value, the richer the muscle. The process involves extracting the skin surface temperature change rate sequence during the historical thermal effect duration at the corresponding acupoint, calculating the variance of the sequence, taking the reciprocal of the variance, and performing Min-Max normalization to obtain the stabilizing factor.

[0089] For example, User A plans to apply moxibustion to the acupoints Guanyuan and Zusanli. Data sequences of skin surface temperature change rates during the historical treatment phase at the corresponding acupoints are extracted: Guanyuan acupoint sequence: 0.70, 0.71, 0.72, variance ≈ 0.0002, reciprocal = 5000, normalized stability factor = 0.6; Zusanli acupoint sequence: 0.68, 0.69, 0.70, variance ≈ 0.0001, reciprocal = 10000, normalized stability factor = 0.4.

[0090] Secondly, from the user's historical moxibustion response records, data on the rate of change of skin surface temperature at corresponding acupoints during the initial stage of heat stress during historical treatments are extracted. The arithmetic mean is calculated and normalized to obtain a sensitivity factor characterizing personalized heat sensitivity. The sensitivity factor is an indicator characterizing the personalized heat sensitivity of acupoints, obtained by normalizing the mean value of the indicator during the initial stage of heat stress; the higher the value, the more sensitive the acupoint. The rate of change of skin surface temperature at the initial stage of historical heat stress for corresponding acupoints is extracted, the arithmetic mean is calculated, and Min-Max normalization is performed to obtain the sensitivity factor. For example, the mean rate of change of skin surface temperature at Guanyuan acupoint is 0.65℃ / min, and the normalized sensitivity factor is 0.4; the mean rate of change of skin surface temperature at Zusanli acupoint is 0.75℃ / min, and the normalized sensitivity factor is 0.6.

[0091] Then, the stabilizing factor and the sensitive factor are weighted and summed to calculate the comprehensive allocation coefficient for the corresponding acupoint. The comprehensive allocation coefficients of all acupoints are then summed to form a coefficient set. The comprehensive allocation coefficient, obtained by weighting and summing the stabilizing factor and the sensitive factor, is an indicator representing the weight of the acupoint moxibustion duration allocation. The coefficient set refers to the sum of the comprehensive allocation coefficients of all planned moxibustion acupoints. Comprehensive allocation coefficient = stabilizing factor × 0.5 + sensitive factor × 0.5. For example, the comprehensive allocation coefficient for Guanyuan acupoint = 0.6 × 0.5 + 0.4 × 0.5 = 0.5; the comprehensive allocation coefficient for Zusanli acupoint = 0.4 × 0.5 + 0.6 × 0.5 = 0.5; the coefficient set is {0.5, 0.5}, and the total coefficient sum is 1.

[0092] Next, based on the proportion of the comprehensive allocation coefficient of each acupoint in the coefficient set to the total, the remaining suggested time is allocated to obtain the single-acupoint suggested time for each acupoint. The single-acupoint suggested time refers to the moxibustion time for a single acupoint obtained after allocating the remaining suggested time according to the proportion of the comprehensive allocation coefficient of the acupoint. Single-acupoint suggested time = remaining suggested time × (comprehensive allocation coefficient of acupoints / total coefficients). For example, user A has 2 minutes of remaining suggested time and a total coefficient of 1; the single-acupoint suggested time for Guanyuan acupoint = 2 × (0.5 / 1) = 1 minute; the single-acupoint suggested time for Zusanli acupoint = 2 × (0.5 / 1) = 1 minute.

[0093] Furthermore, the biofeedback monitoring criteria for triggering real-time alarms and pausing moxibustion are set as follows: the immediate physiological load index continuously exceeds the dynamic threshold for heat tolerance, or the skin surface temperature change rate suddenly increases beyond a preset change rate threshold. These biofeedback monitoring criteria refer to the risk assessment conditions that trigger moxibustion alarms and pause moxibustion. Two types of triggering conditions are set: the immediate physiological load index continuously exceeds the dynamic threshold for heat tolerance; and the skin surface temperature change rate suddenly increases beyond a preset change rate threshold. For example, the biofeedback monitoring criteria for User A are: if the immediate physiological load index continuously exceeds the dynamic threshold for heat tolerance by 0.22, an alarm will be triggered immediately; if the skin surface temperature change rate suddenly increases beyond the preset change rate threshold of 0.2℃ / min, an alarm will be triggered immediately.

[0094] Finally, the phased temperature control curves, the recommended duration for each acupoint, and the key points of real-time biofeedback monitoring are summarized. Combined with the recommended target temperature and remaining recommended duration, a personalized moxibustion guidance plan is generated. The complete plan is generated by integrating the phased temperature control curves, the recommended duration for each acupoint, the key points of real-time biofeedback monitoring, the recommended target temperature, and the remaining recommended duration. For example, User A's personalized moxibustion guidance plan: Temperature control curve: 0-3 minutes to 44.5℃, 3-5 minutes to maintain a constant temperature of 44.5℃, 5-6 minutes to cool down to 37℃; Moxibustion duration: 1 minute each for Guanyuan and Zusanli; Monitoring points: Real-time monitoring of physiological load index and skin temperature change rate, pausing immediately upon reaching the target; Recommended target temperature: 44.5℃, remaining total moxibustion time: 2 minutes. Users can perform moxibustion according to this personalized guidance plan.

[0095] In this embodiment of the invention, a phased temperature control curve of linear heating, constant temperature, and linear cooling is generated by analyzing the optimal control parameters. The duration of moxibustion at acupoints is allocated based on the thickness of the acupoint muscles and the heat sensitivity. Dual biofeedback monitoring points are set to ultimately form a complete and feasible personalized moxibustion guidance plan. This not only achieves smooth temperature control transition and precise acupoint stimulation, but also provides real-time risk warnings, completely solving the problems of rigid temperature control, uneven duration allocation, and lack of risk warnings in traditional moxibustion, further improving the safety and therapeutic effect of home moxibustion.

[0096] Through the specific implementation methods described above, the embodiments of the present invention achieve the following technical effects: This invention provides a method and system for generating personalized home moxibustion guidance schemes based on the Internet of Things. First, it dynamically divides the initial stage of heat stress and the sustained stage of heat effect based on the user's real-time physiological characteristics, laying a phased foundation for subsequent precise control. Second, it extracts key user biometrics and combines them with historical safety data, using a unified percentile calculation logic to construct a personalized dynamic threshold for heat tolerance, establishing a safety boundary adapted to the user's physiological state. Third, it trains an adaptive adjustment model based on deep reinforcement learning, learning the optimal temperature-duration control pattern that achieves therapeutic effects without exceeding the safety threshold, outputting a set of control parameters that balances safety and efficacy. Finally, it analyzes the optimal parameters to generate a phased temperature control curve, allocating the duration of each acupoint based on the thickness of the acupoint muscles and the individual's heat sensitivity, and setting dual biofeedback monitoring points, ultimately forming a directly implementable personalized moxibustion guidance scheme that effectively balances moxibustion temperature and duration, improving the personalization level of home moxibustion.

[0097] Example 2, as Figure 2 As shown, this invention provides a personalized home moxibustion guidance program generation system based on the Internet of Things, the system comprising: The dynamic stage division module 11 is used to collect real-time physiological data of the target user and operating data of the moxibustion device, and dynamically divide the current moxibustion stage. The heat tolerance threshold calculation module 12 is used to extract key biological characteristics within the current moxibustion stage and combine them with the user's historical moxibustion reaction records to calculate a personalized dynamic heat tolerance threshold. The regulation parameter output module 13 is used to input the key biological characteristics and the dynamic threshold of heat tolerance into the pre-trained adaptive regulation model and output the optimal temperature-duration regulation parameter set. The guidance scheme generation module 14 is used to generate a personalized moxibustion guidance scheme based on the optimal temperature-duration control parameter group, which includes a phased temperature control curve, a suggested duration for a single acupoint, and key points for real-time biofeedback monitoring, and to drive the moxibustion device to perform preheating.

[0098] In one embodiment, the dynamic stage division module 11 is further configured to: The real-time physiological data includes the rate of change of skin surface temperature and heart rate variability index collected by wearable devices and flexible sensors, and the operating data of the moxibustion device includes the real-time temperature of the moxibustion head and the current cumulative working time.

[0099] The dynamic division of the current moxibustion stage includes: The rate of change of skin surface temperature is compared with a preset initial stress threshold, and the duration during which the rate of change is consistently higher than the initial stress threshold is defined as the initial stage of heat stress. After the initial stage of heat stress ends, when the heart rate variability index falls into the steady-state regulation range and the real-time temperature of the moxibustion head remains stable, it is divided into the heat effect duration stage. When the current accumulated working time reaches the personalized stage duration benchmark generated based on the user's historical moxibustion response records, it is divided into the preparation stage for the end of the treatment.

[0100] The steps for obtaining the steady-state adjustment range include: Obtain the initial heart rate variability index set continuously monitored by the target user in a non-moxibustion resting state, calculate the median value, and use it as the individual resting baseline value. From the user's historical moxibustion reaction records, all treatment periods marked as effective and without adverse reactions were screened, and the heart rate variability index during the stable period within the period was extracted to form a set of historical effective data. Calculate the upper and lower quartiles of the historical valid data set, use the individual resting baseline value as the central reference point, and determine the steady-state adjustment range based on the upper and lower quartiles.

[0101] In one embodiment, the heat tolerance threshold calculation module 12 is further configured to: During the duration of the thermal effect, the rate of change of skin surface temperature and heart rate variability were continuously monitored. The instantaneous physiological load index is generated by weighted fusion calculation based on the ratio of the skin surface temperature change rate to the historical mean at the lower limit of the steady-state regulation range, and the ratio of the heart rate variability index to the individual resting baseline value. From the user's historical moxibustion response records, query the historical valid records that are closest to the current real-time temperature of the moxibustion head, and extract the average skin surface temperature change rate and average heart rate variability index before reaching the preset warning line during the duration of the thermal effect, and calculate the individual's historical tolerance baseline. The steps for determining the preset warning line include: From the user's historical moxibustion response records, all treatment records marked as safely completed and effective are selected to form a safe record set; For each record in the safety record set, extract the skin surface temperature change rate and heart rate variability index of the last steady state during the thermal effect duration and before the end of the recording to form a safety critical feature set; Calculate the upper quantile of the skin surface temperature change rate data for all skin surface temperature change rate data in the safety critical feature set, and use it as the warning line for skin surface temperature change rate. Calculate the specified upper quantile for all heart rate variability index data, and use it as the warning line for heart rate variability index; The preset warning line is composed of the skin surface temperature change rate warning line and the heart rate variability index warning line. The instantaneous physiological load index is compared with the personal historical tolerance baseline. Taking into account the percentage of completion of the current cumulative working hours relative to the personalization phase duration benchmark, and the degree to which the instantaneous physiological load index approaches the personal historical tolerance baseline, the personal historical tolerance baseline is dynamically reduced. The reduced result is output as the dynamic threshold of heat tolerance.

[0102] Specifically, the average skin surface temperature change rate and average heart rate variability index are extracted during the duration of the thermal effect before reaching the preset warning line, and an individual's historical tolerance baseline is calculated, including: Based on the retrieved historical records, the time point when the rate of change of skin surface temperature first continuously exceeds the warning line of the rate of change of skin surface temperature or the time point when the heart rate variability index first continuously exceeds the warning line of the heart rate variability index is located during the duration of the thermal effect is used as the trigger point of the preset warning line. Extract the skin surface temperature change rate data sequence and the heart rate variability index data sequence within a preset time period before the trigger point; Calculate the arithmetic mean of the skin surface temperature change rate data sequence as the average skin surface temperature change rate; Calculate the arithmetic mean of the heart rate variability index data sequence as the mean heart rate variability index; The average skin surface temperature change rate and the average heart rate variability index are normalized and weighted and summed, and the calculation results are output as the individual's historical tolerance baseline.

[0103] In one embodiment, the guidance scheme generation module 14 is further configured to: The optimal temperature-duration control parameter set is analyzed to obtain the recommended target temperature and remaining suggested duration for the duration of the thermal effect. Based on the recommended target temperature, the duration of the initial stage of thermal stress, and the steady-state adjustment range, a staged temperature control curve is generated. The remaining suggested duration is allocated proportionally according to the muscle thickness and sensitivity of the acupoints to be treated, and the suggested duration for each acupoint is calculated. The biofeedback monitoring points for triggering real-time alarms and pausing moxibustion are set as follows: the immediate physiological load index continuously exceeds the dynamic threshold of heat tolerance, or the skin surface temperature change rate suddenly increases beyond the preset change rate threshold. By summarizing the phased temperature control curves, the recommended duration for each acupoint, and the key points of real-time biofeedback monitoring, and combining the recommended target temperature with the remaining recommended duration, a personalized moxibustion guidance plan is generated.

[0104] Specifically, based on the recommended target temperature, the duration of the initial stage of thermal stress, and the steady-state adjustment range, a staged temperature control curve is generated, including: Starting from the initial temperature of moxibustion and taking the duration of the initial stage of heat stress as the total duration, calculate the average heating rate required to linearly increase the temperature from the initial temperature to the recommended target temperature. Using the recommended target temperature as the steady-state temperature and the remaining suggested duration as the duration of the isothermal phase, an isothermal control segment with a constant temperature is generated. Set a fixed cooling time and a cooling endpoint temperature, and calculate the average cooling rate required to linearly decrease from the recommended target temperature to the cooling endpoint temperature. The linear heating section, the constant temperature control section, and the linear cooling section are connected in chronological order to form a staged temperature control curve.

[0105] The remaining suggested duration is allocated proportionally based on the muscle thickness and sensitivity of the acupoints to be treated, and the suggested duration for each acupoint is calculated, including: For each acupoint to be treated, the data sequence of skin surface temperature change rate during the heat effect sustained stage of the corresponding acupoint is extracted from the user's historical moxibustion response records. The variance of the sequence is calculated, and the reciprocal of the variance is normalized to obtain a stabilizing factor characterizing the degree of muscle thickness. From the user's historical moxibustion response records, the skin surface temperature change rate data of the corresponding acupoints in the initial stage of heat stress during historical treatment are extracted, the arithmetic mean is calculated and normalized to obtain the sensitivity factor characterizing personalized heat sensitivity. The stabilizing factor and the sensitive factor are weighted and summed to calculate the comprehensive allocation coefficient of the corresponding acupoint. The comprehensive allocation coefficients of all acupoints are then combined to form a coefficient set. Based on the proportion of the comprehensive allocation coefficient of each acupoint in the coefficient set to the total, the remaining suggested duration is allocated to obtain the single-acupoint suggested duration for each acupoint.

[0106] It should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for generating personalized home moxibustion guidance plans based on the Internet of Things, characterized in that, The method includes: Collect real-time physiological data of the target user and operating data of the moxibustion device to dynamically divide the current moxibustion stage; Key biomarkers within the current moxibustion stage are extracted and combined with the user's historical moxibustion response records to calculate a personalized dynamic threshold for heat tolerance. The key biofeatures and dynamic heat tolerance thresholds are input into a pre-trained adaptive adjustment model, and the optimal temperature-duration control parameter set is obtained from the output. Based on the optimal temperature-duration control parameter set, a personalized moxibustion guidance plan is generated, which includes a phased temperature control curve, a suggested duration for a single acupoint, and key points for real-time biofeedback monitoring, and the moxibustion device is driven to perform preheating.

2. The method for generating personalized home moxibustion guidance plans based on the Internet of Things according to claim 1, characterized in that, The real-time physiological data includes the rate of change of skin surface temperature and heart rate variability index collected by wearable devices and flexible sensors. The operating data of the moxibustion device includes the real-time temperature of the moxibustion head and the current cumulative working time.

3. The method for generating personalized home moxibustion guidance plans based on the Internet of Things according to claim 1, characterized in that, The current moxibustion stage is dynamically divided, including: The rate of change of skin surface temperature is compared with a preset initial stress threshold, and the duration during which the rate of change is consistently higher than the initial stress threshold is defined as the initial stage of heat stress. After the initial stage of heat stress ends, when the heart rate variability index falls into the steady-state regulation range and the real-time temperature of the moxibustion head remains stable, it is divided into the heat effect duration stage. When the current accumulated working time reaches the personalized stage duration benchmark generated based on the user's historical moxibustion response records, it is divided into the preparation stage for the end of the treatment.

4. The method for generating personalized home moxibustion guidance plans based on the Internet of Things according to claim 3, characterized in that, The steps for obtaining the steady-state adjustment range include: Obtain the initial heart rate variability index set continuously monitored by the target user in a non-moxibustion resting state, calculate the median value, and use it as the individual resting baseline value. From the user's historical moxibustion reaction records, all treatment periods marked as effective and without adverse reactions were screened, and the heart rate variability index during the stable period within the period was extracted to form a historical effective data set; Calculate the upper and lower quartiles of the historical valid data set, use the individual resting baseline value as the central reference point, and determine the steady-state adjustment range based on the upper and lower quartiles.

5. The method for generating personalized home moxibustion guidance plans based on the Internet of Things according to claim 1, characterized in that, Key biomarkers within the current moxibustion stage are extracted and combined with the user's historical moxibustion response records to calculate a personalized dynamic threshold for heat tolerance, including: During the duration of the thermal effect, the rate of change of skin surface temperature and heart rate variability were continuously monitored. The instantaneous physiological load index is generated by weighted fusion calculation based on the ratio of the skin surface temperature change rate to the historical mean at the lower limit of the steady-state regulation range, and the ratio of the heart rate variability index to the individual resting baseline value. From the user's historical moxibustion response records, query the historical valid records that are closest to the current real-time temperature of the moxibustion head, and extract the average skin surface temperature change rate and average heart rate variability index before reaching the preset warning line during the heat effect duration phase, and calculate the individual's historical tolerance baseline. The steps for determining the preset warning line include: From the user's historical moxibustion response records, all treatment records marked as safely completed and effective are selected to form a safe record set; For each record in the safety record set, extract the skin surface temperature change rate and heart rate variability index of the last steady state during the thermal effect duration and before the end of the recording to form a safety critical feature set; Calculate the upper quantile of the skin surface temperature change rate data for all skin surface temperature change rate data in the safety critical feature set, and use it as the warning line for skin surface temperature change rate. Calculate the specified upper quantile for all heart rate variability index data, and use it as the warning line for heart rate variability index; The preset warning line is composed of the skin surface temperature change rate warning line and the heart rate variability index warning line. The instantaneous physiological load index is compared with the personal historical tolerance baseline. Taking into account the percentage of completion of the current cumulative working hours relative to the personalization phase duration benchmark, and the degree to which the instantaneous physiological load index approaches the personal historical tolerance baseline, the personal historical tolerance baseline is dynamically reduced. The reduced result is output as the dynamic threshold of heat tolerance.

6. The method for generating personalized home moxibustion guidance plans based on the Internet of Things according to claim 5, characterized in that, The average skin surface temperature change rate and average heart rate variability were extracted before reaching the preset warning line during the duration of the thermal effect, and the individual's historical tolerance baseline was calculated, including: Based on the retrieved historical records, the time point when the rate of change of skin surface temperature first continuously exceeds the warning line of the rate of change of skin surface temperature or the time point when the heart rate variability index first continuously exceeds the warning line of the heart rate variability index is located during the duration of the thermal effect is used as the trigger point of the preset warning line. Extract the skin surface temperature change rate data sequence and the heart rate variability index data sequence within a preset time period before the trigger point; Calculate the arithmetic mean of the skin surface temperature change rate data sequence as the average skin surface temperature change rate; Calculate the arithmetic mean of the heart rate variability index data sequence as the mean heart rate variability index; The average skin surface temperature change rate and the average heart rate variability index are normalized and weighted and summed, and the calculation results are output as the individual's historical tolerance baseline.

7. The method for generating personalized home moxibustion guidance plans based on the Internet of Things according to claim 1, characterized in that, Based on the optimal temperature-duration control parameter set, a personalized moxibustion guidance plan is generated, including a phased temperature control curve, suggested duration for each acupoint, and key points for real-time biofeedback monitoring. The optimal temperature-duration control parameter set is analyzed to obtain the recommended target temperature and remaining suggested duration for the duration of the thermal effect. Based on the recommended target temperature, the duration of the initial stage of thermal stress, and the steady-state adjustment range, a staged temperature control curve is generated. The remaining suggested duration is allocated proportionally according to the muscle thickness and sensitivity of the acupoints to be treated, and the suggested duration for each acupoint is calculated. The biofeedback monitoring points for triggering real-time alarms and pausing moxibustion are set as follows: the immediate physiological load index continuously exceeds the dynamic threshold of heat tolerance, or the skin surface temperature change rate suddenly increases beyond the preset change rate threshold. By summarizing the phased temperature control curves, the recommended duration for each acupoint, and the key points of real-time biofeedback monitoring, and combining the recommended target temperature with the remaining recommended duration, a personalized moxibustion guidance plan is generated.

8. The method for generating personalized home moxibustion guidance plans based on the Internet of Things according to claim 7, characterized in that, Based on the recommended target temperature, the duration of the initial stage of heat stress, and the steady-state regulation range, a staged temperature control curve is generated, including: Starting from the initial temperature of moxibustion and taking the duration of the initial stage of heat stress as the total duration, calculate the average heating rate required to linearly increase the temperature from the initial temperature to the recommended target temperature. Using the recommended target temperature as the steady-state temperature and the remaining suggested duration as the duration of the isothermal phase, an isothermal control segment with a constant temperature is generated. Set a fixed cooling time and a cooling endpoint temperature, and calculate the average cooling rate required to linearly decrease from the recommended target temperature to the cooling endpoint temperature. The linear heating section, the constant temperature control section, and the linear cooling section are connected in chronological order to form a staged temperature control curve.

9. The method for generating personalized home moxibustion guidance plans based on the Internet of Things according to claim 7, characterized in that, The remaining suggested time is allocated proportionally based on the muscle thickness and sensitivity of the acupoints to be treated, and the suggested time for each acupoint is calculated, including: For each acupoint to be treated, the data sequence of skin surface temperature change rate during the heat effect sustained stage of the corresponding acupoint is extracted from the user's historical moxibustion response records. The variance of the sequence is calculated, and the reciprocal of the variance is normalized to obtain a stabilizing factor characterizing the degree of muscle thickness. From the user's historical moxibustion response records, the skin surface temperature change rate data of the corresponding acupoints in the initial stage of heat stress during historical treatment are extracted, the arithmetic mean is calculated and normalized to obtain the sensitivity factor characterizing personalized heat sensitivity. The stabilizing factor and the sensitive factor are weighted and summed to calculate the comprehensive allocation coefficient of the corresponding acupoint. The comprehensive allocation coefficients of all acupoints are then combined to form a coefficient set. Based on the proportion of the comprehensive allocation coefficient of each acupoint in the coefficient set to the total, the remaining suggested duration is allocated to obtain the single-acupoint suggested duration for each acupoint.

10. A personalized home moxibustion guidance program generation system based on the Internet of Things, characterized in that, The system for implementing the IoT-based personalized home moxibustion guidance program generation method according to any one of claims 1-9, the system comprising: The dynamic stage division module is used to collect real-time physiological data of the target user and operating data of the moxibustion device, and dynamically divide the current moxibustion stage. The heat tolerance threshold calculation module is used to extract key biological characteristics within the current moxibustion stage and, in conjunction with the user's historical moxibustion reaction records, calculate a personalized dynamic heat tolerance threshold. The regulation parameter output module is used to input the key biofeatures and the dynamic threshold of heat tolerance into the pre-trained adaptive regulation model and output the optimal temperature-duration regulation parameter set. The guidance scheme generation module is used to generate a personalized moxibustion guidance scheme based on the optimal temperature-duration control parameter group, which includes a phased temperature control curve, a suggested duration for a single acupoint, and key points for real-time biofeedback monitoring, and to drive the moxibustion device to perform preheating.

Images