Method and system for monitoring and early warning mental health of special occupational population based on combination of smart watch and AI

Abstract

The invention belongs to the technical field of medical treatment, discloses a mental health monitoring and early warning method for special occupational crowds based on a method of combining a smart watch and an AI, and relates to mental health monitoring and early warning and application for the special occupational crowds based on the method of combining the smart watch and the AI. The watch has the functions of collecting physiological indexes such as heart rate variability (HRV), electrodermal response (SCR), sleep efficiency, core body temperature fluctuation, blood pressure change trend and the like, and analyzing the collected data through an AI-carried big data platform, so that the real-time mental health condition of a user can be quickly and dynamically analyzed; the method is suitable for daily self-management and health monitoring of special occupational people.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of medical treatment, and particularly relates to a method for monitoring and early warning of the mental health of special occupational groups based on the combination of an intelligent watch and AI. BACKGROUND

[0002] I. The closest prior art

[0003] Patent name / document: "IoT-based wearable device, system and method for the management of a stress level and mental health of a human body" (WO2019012471A1). The system at least includes:

[0004] One or more body sensors (such as electrodermal activity sensors, skin temperature sensors, electrocardiograms, electromyograms, 9-axis motion sensors, etc.) for collecting body data such as physiological parameters, body movements or heat consumption;

[0005] Environmental sensors for collecting context data related to the environment;

[0006] A main processing unit for receiving and processing body data and environmental data, identifying body and mental health risks by comparing reference values and combining machine learning / artificial intelligence models, generating a "body score" or rating for assessing stress levels or mental health status; also supporting remote communication, database storage, treatment recommendations, etc.

[0007] II. Technical problems

[0008] Although this patent covers core technologies such as body + environment multi-source fusion, artificial intelligence risk assessment, IoT remote communication systems, etc., the proposed "special occupational group-oriented" mental health monitoring and early warning system still has the following key deficiencies:

[0009] 1. Lack of NLP and subjective semantic input fusion capability

[0010] WO2019012471A1 mainly focuses on the fusion and processing of physiological and environmental sensor data, and does not involve natural language processing (NLP) techniques such as questionnaire text, voice tone, etc. It also does not correct subjective scoring bias.

[0011] 2. Lack of grading early warning mechanism and group trend analysis function

[0012] Although the prior art can generate a physical / mental health "score" based on data, it does not establish a "three-level risk warning" and "group mental health trend visualization" (such as heat map, workplace risk hotspot analysis, etc.) module like the scheme, nor does it have the function of macro-level analysis to assist in formulating health promotion plans. SUMMARY

[0013] In view of the problems of the prior art, the present application provides a method for monitoring and warning the mental health of special professional groups based on the combination of smart watches and AI.

[0014] The present application is implemented as follows: a method for monitoring and warning the mental health of special professional groups based on the combination of smart watches and AI includes:

[0015] Step 1: Evaluate the mental health status of professional groups from the living environment and the working environment; realize real-time data collection through wearable smart devices and environmental perception terminals;

[0016] (wear a smart watch (with non-invasive, light, waterproof / anti-sweat / anti-damp / anti-fog, high temperature / corrosion resistant characteristics) to collect real-time physiological indicators such as heart rate variability (HRV), skin conductance response (SCR), sleep efficiency (including sleep staging data), core body temperature, and blood pressure trend; through Bluetooth or Wi-Fi, transmit data to the terminal; deploy embedded terminals (such as smart sensors) in professional places to monitor environmental factors such as light intensity, noise decibels, air quality (PM2.5, VOCs, CO2 concentration, total dust concentration (TSP), PM10), temperature and humidity, physical hazards (vibration level), mold / allergen level indicators, ultraviolet radiation intensity (UVI), and ionizing radiation level in real time; record data such as working hours and task intensity; collect subjective feedback using natural interaction methods such as voice input and text chat; analyze semantic sentiment using natural language processing (NLP) technology)

[0017] Step 2: Based on AI learning, construct a dynamic evaluation model in combination with reality, combine traditional statistical methods to realize real-time updating and risk warning of evaluation indicators; construct a dynamic evaluation system based on natural language processing (NLP), multi-factor decision tree model, and emotion recognition algorithm to automatically analyze questionnaire text semantics, voice tone, correct subjective score bias, and realize real-time calculation and three-level warning of mental health risk index;

[0018] Step 3: Present group mental health trends, environmental risk heat maps (such as the noise exceeding rate of each workplace), and other indicators in real time through wearable smart devices, analyze the main factors affecting the mental health status of professional groups, and assist managers in formulating macro health promotion plans.

[0019] Further, the sleep staging data, heart rate variability (HRV) and other indicators collected by the wearable device are added in the evaluation of the physiological health status of the professional population in step 1, and the HRV, SCR, sleep efficiency, and working environment noise value (collected by the built-in microphone of the watch, with a cutoff value > 85 dB) are input into the multi-factor decision tree to output a mental health risk index; the smart watch has a built-in data caching mechanism that can temporarily store key data when the connection is abnormal or lost, and automatically upload the data after the connection is restored; the device supports face recognition unlocking and has a positioning and finding function.

[0020] Further, in step 2, the evaluation system is used, combined with the environmental parameters (light, noise, air quality, etc.) and physiological data collected by the smart device, to monitor and warn the mental health status of the professional population in advance through a multi-factor decision tree model algorithm; when the calculated health risk index reaches the preset threshold or the key physiological indicators (such as abnormal reduction of HRV, continuous noise exceeding the limit) exceed the safety range, the watch will send real-time warnings to the user through screen flickering prompts, vibration alarms, and APP push messages.

[0021] Another object of the present application is to provide a mental health monitoring and warning system for special professional populations based on a smart watch combined with AI, which comprises:

[0022] A data collection module is used to evaluate the mental health status of professional populations from living and working environments; dynamic data is collected in real time through wearable smart devices and environmental perception terminals;

[0023] A warning module is used to combine reality and build a dynamic evaluation model based on AI learning, and to realize real-time updating and risk warning of evaluation indicators combined with traditional statistical methods; a dynamic evaluation system based on natural language processing (NLP), multi-factor decision tree model, and emotion recognition algorithm is constructed to automatically analyze the semantic and tone of the questionnaire text, correct the subjective score deviation, and realize real-time calculation and three-level warning of the mental health risk index;

[0024] An analysis module is used to present the group mental health trend, environmental risk heat map (such as the noise exceeding rate of each workplace), and other indicators in real time through wearable smart devices, analyze the main factors affecting the mental health status of professional populations, and assist managers in formulating macro health promotion plans.

[0025] Another object of the present application is to provide a computer device, which comprises a memory and a processor, the memory stores a computer program, and the computer program is executed by the processor to make the processor execute the steps of the mental health monitoring and warning method for special professional populations based on a smart watch combined with AI.

[0026] Another object of the present application is to provide a computer readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the method for monitoring and early warning of the mental health of special professional groups based on the combination of smart watches and AI.

[0027] Another object of the present application is to provide an information data processing terminal for implementing the mental health monitoring and early warning system for special professional groups based on the combination of smart watches and AI.

[0028] In combination with the above technical solutions and the technical problems solved, the technical solution to be protected by the present application has the following advantages and positive effects:

[0029] The technical solution to be protected by the present application has the following advantages and positive effects: It provides a reference for formulating a mental health evaluation system suitable for professional groups in China; and it can accurately clarify the relationship between the work environment and the mental health status of professional groups, propose feasible improvement measures for possible risk factors in the work environment, optimize the work environment of professional groups, and provide a reference value for relevant departments to formulate health promotion plans.

[0030] The present solution first realizes: multi-modal real-time fusion (physiological environment synchronous acquisition triggered by HRV+SCR+environmental noise>85dB threshold); three-level early warning mechanism (screen flicker (2Hz)+vibration (5 seconds)+APP push composite alarm); group-level analysis capability (generate noise over-standard rate heat map, visualize workplace risk distribution)

[0031] Traditional means lagging problem: solved the "delayed response of questionnaires", through NLP algorithm analysis of voice tone, the psychological evaluation cycle is shortened from week level to minute level;

[0032] Complex environment interference: adopt cache-continuation mechanism, in the truck cab and other signal unstable scenes, still guarantee the continuity of HRV / SCR data (key data loss rate <0.1%)

[0033] Multi-source data correction: use environmental sensor data (such as vibration level 0.8m / s 2 ) to explain the abnormal fluctuation of SCR, reduce the false positive rate.

[0034] Overturn the "psychological evaluation is not quantifiable" bias, traditional view believes that: mental health status is difficult to directly represent by physiological indicators. The present solution uses a decision tree quantification model (multi-factor decision tree core logic) to demonstrate the effect, and the risk index in the embodiment 7 is consistent with the clinical diagnosis rate of 92%.

[0035] Break the "wearable devices are not suitable for industrial environment" cognitive device characteristics breakthrough: waterproof / shockproof: IP68 certified (vibration tolerance > 1.0 m / s 2 ); extreme environment adaptation: -20℃ ~ 60℃ working temperature range (covering scenarios such as steelmaking / polar operations). BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 is a psychological health monitoring and early warning method flowchart for special professional groups based on the method of combining smart watches with AI provided by the embodiments of the present application.

[0037] Figure 2 is a structural block diagram of a psychological health monitoring and early warning system for special professional groups based on the method of combining smart watches with AI provided by the embodiments of the present application.

[0038] Figure 3 is a psychological health assessment specification flowchart provided by the embodiments of the present application.

[0039] Figure 4 is a schematic diagram of the internal circuit of a smart watch provided by the embodiments of the present application.

[0040] Figure 5 is a data comparison chart before and after intervention provided by the embodiments of the present application.

[0041] Figure 6 is a data trend chart during the intervention process provided by the embodiments of the present application.

[0042] Figure 7 is a comparison chart of the missed detection rate of the traditional mode and the smart watch provided by the embodiments of the present application.

[0043] Figure 8 is a comparison chart of the time to eliminate 50% missed detection rate of the traditional mode and the smart watch provided by the embodiments of the present application.

[0044] Figure 9 is a comparison chart of the intervention completion rate of the traditional mode and the smart watch provided by the embodiments of the present application. DETAILED DESCRIPTION

[0045] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0046] In the existing field of occupational health management, the conventional monitoring means often rely on physical examination, questionnaire and interview, etc. static data acquisition methods, which have the problems of update lag and individual differences not reflecting in real time, and are difficult to meet the dynamic tracking needs of high-risk occupational groups on mental health status. The present application establishes a dynamic monitoring framework of continuous collection, physiological behavior and environmental factors parallel processing by combining smart watch and artificial intelligence algorithm, which fundamentally solves the industry technology short board of traditional monitoring unable to respond in real time and lack of interactive feedback. The core is to open up the multi-source data channel, integrate the living and working environment parameters, physiological sensing data and psychological scale results, and form a calculable and iterative risk assessment base.

[0047] In terms of system working principle, the smart watch as a front-end sensing terminal relies on its built-in heart rate sensor, accelerometer and environmental noise detection module to capture the individual's heart rate variability, skin electric response, sleep staging and external acoustic environment parameters in real time. These data ensure data integrity and time sequence continuity through local caching and breakpoint resume mechanism when the network is unstable or the device is dropped. At the same time, the watch end supports face recognition and positioning function, improves the security of data transmission and user identity correspondence, and provides stable and reliable original data source for the back-end analysis.

[0048] In the back-end evaluation link, a multi-level model architecture integrating traditional statistics and machine learning is constructed. The AI learning module based on multi-factor decision tree and Natural Language Processing (NLP) algorithm jointly models the subjective scale input and objective physiological parameters, automatically corrects the distortion caused by questionnaire filling bias or individual perception difference. The semantic analysis and emotion recognition algorithm deeply analyzes the user's voice tone and text, and establishes a psychological risk portrait complementary to physiological signals. This multi-modal fusion method makes the dynamic calculation of risk index more consistent with the psychological fluctuation characteristics in real scene.

[0049] In terms of risk early warning mechanism, the system sets dynamic threshold and upper and lower limits of key physiological indicators to realize multi-level triggering strategy. Once abnormal trend or dangerous signal is detected, the terminal immediately prompts the user in the form of vibration, screen flicker and mobile APP push, and synchronously uploads the event to the background server for managers to carry out hierarchical intervention and tracking. This real-time closed-loop early warning mechanism greatly shortens the time window from abnormal occurrence to intervention response, and has significant prevention and control value for sudden psychological risk events.

[0050] At the group level, the system can draw the psychological health trend curve and environmental risk heat map of the professional population through the statistical integration of multi-source data, and intuitively present the long-term effect of environmental factors on mental health. Through dynamic visualization and causal analysis, the system can identify the main environmental variables and work intensity factors that affect the psychological health status of the group, providing scientific support for the macro intervention policy of enterprises or organizations. This process realizes the three-dimensional dynamic closed-loop analysis of individuals, groups and environment.

[0051] This method breaks through the limitations of traditional single evaluation tools and changes the psychological health risk monitoring from static to dynamic, from individual to group, and from passive to active. The key problem it solves in industrial application is how to combine real-time data streams from wearable devices with the multi-modal processing capabilities of artificial intelligence to achieve continuous monitoring and early warning of the mental health of professional populations. The system built in this way not only has forward-looking value in mental health intervention, but also provides a replicable and scalable technical paradigm for future occupational health management.

[0052] As shown in Figure 1 The method for monitoring and early warning the mental health of special professional populations based on the combination of smart watches and AI provided by the embodiment of the application comprises the following steps:

[0053] S101: Evaluate the mental health status of professional populations from the living environment and the working environment; realize real-time data collection through wearable smart devices and environmental perception terminals;

[0054] (wear a smart watch (with non-invasive, light, waterproof / anti-sweat / anti-moisture / anti-fog, high temperature / corrosion resistant characteristics) to collect real-time physiological indicators such as heart rate variability (Heart rate variability, HRV), skin conductance response (Skin Conductance Response, SCR), sleep efficiency (including sleep staging data), core body temperature, and blood pressure trend; through Bluetooth or Wi-Fi, transmit the data to the terminal; deploy embedded terminals (such as smart sensors) in professional places to monitor environmental factors such as light intensity, noise decibel, air quality (PM2.5, VOCs, CO2 concentration, total suspended particulate (TSP), PM10), temperature and humidity, physical hazards (vibration level), mold / allergen level, ultraviolet radiation intensity (UVI), and ionizing radiation level in real time; record data such as working hours and task intensity; collect subjective feedback through natural interaction methods such as voice input and text chat; analyze semantic sentiment using NLP technology)

[0055] S102: Based on AI learning, a dynamic evaluation model is constructed in combination with the actual situation, traditional statistical methods are combined, real-time updating of evaluation indexes and risk early warning are realized; a dynamic evaluation system based on NLP, a multi-factor decision tree model and an emotion recognition algorithm is constructed, questionnaire text semantics and voice intonation are automatically analyzed, subjective score deviation is corrected, real-time calculation of a mental health risk index and three-level early warning are realized;

[0056] S103: Through a wearable smart device, a group mental health trend and an environmental risk heat map index are presented in real time, main factors affecting the mental health status of a professional group are analyzed, and a macro health promotion plan is formulated for a manager.

[0057] The sleep staging data, HRV and other indexes collected by the wearable device are added in the evaluation of the physiological health status of the professional group in S101, a multi-factor decision tree is used, HRV, SCR, sleep efficiency and working environment noise value (collected by the built-in microphone of the watch, the cutoff value is > 85dB) are input, and a mental health risk index is output; the smart watch has a built-in data caching mechanism, which can temporarily store key data when the connection is abnormal or lost, and automatically upload after the connection is restored; the device supports face recognition unlocking and has a positioning and finding function.

[0058] In S102, the evaluation system is used in combination with the environmental parameters (light, noise, air quality, etc.) and physiological data collected by the smart device, a multi-factor decision tree model algorithm is used, and the mental health status of the professional group is monitored and warned in advance; when the calculated health risk index reaches a preset threshold or the key physiological indexes (such as abnormal reduction of HRV and continuous over-limit of noise) exceed the safety range, the watch will send real-time early warning to the user through screen flickering prompt, vibration alarm and APP push message.

[0059] As shown in Figure 2 The mental health monitoring and early warning system for special professional groups based on the combination of a smart watch and AI provided by the embodiment of the application includes:

[0060] A data collection module is used to collect data from living environment and working environment, and evaluate the mental health status of the professional group; through a wearable smart device and an environment perception terminal, dynamic data is collected in real time;

[0061] An early warning module is used to construct a dynamic evaluation model based on AI learning in combination with the actual situation, traditional statistical methods are combined, real-time updating of evaluation indexes and risk early warning are realized; a dynamic evaluation system based on NLP, a multi-factor decision tree model and an emotion recognition algorithm is constructed, questionnaire text semantics and voice intonation are automatically analyzed, subjective score deviation is corrected, real-time calculation of a mental health risk index and three-level early warning are realized;

[0062] An analysis module is configured to present real-time group mental health trends, environmental risk heat maps (e.g., noise exceeding standard rates in various workplaces), and other indicators through the smart wearable device, analyze main factors affecting the mental health status of the professional group, and assist managers in formulating macro health promotion plans.

[0063] The working principle of the embodiment of the present application can be described as follows:

[0064] The system acquires multidimensional data of special professional groups in real time through a data acquisition module composed of a smart watch and an environmental perception terminal. The smart watch is responsible for collecting heart rate variability, skin electrical response, sleep quality, activity amount, and other physiological and behavioral signals; the environmental perception terminal captures environmental parameters such as noise, light, and air quality in the living and working places. Through wireless transmission, these data are uploaded to the central processing platform to form a real-time dynamic raw data stream.

[0065] The early warning module is based on AI algorithm for intelligent modeling. The system uses natural language processing technology to analyze questionnaire texts and voice inputs, combines a multi-factor decision tree model and an emotion recognition algorithm, cross- validates user subjective feelings and objective physiological signals, and corrects the bias in psychological evaluation. The AI model will continuously learn historical data and real-time data, update weight parameters, and realize dynamic calculation of the mental health risk index. The index is divided into three threshold intervals, and when the index exceeds the threshold, the system automatically triggers the corresponding level of early warning prompt.

[0066] The analysis module processes and presents macro trends. It aggregates individual health risks into group trends and presents them through visualization, such as group mental health change curves, risk level distribution maps, and environmental risk heat maps. For example, if a certain type of workplace has long-term noise exceeding standard conditions, the system will automatically mark the risk hotspots and infer their potential impact on mental health. Such analysis results provide data support for managers to develop targeted health intervention measures.

[0067] Based on early warning and analysis, the system also has an adaptive feedback mechanism. After the user receives intervention (such as relaxation training, psychological counseling, and work environment optimization), the system will continue to track changes in related indicators and dynamically correct the risk assessment model. This closed-loop feedback mechanism ensures that the system can continuously optimize the evaluation accuracy and form a health profile library that takes both individualization and groupization into account in long-term application.

[0068] The method described in the application can be realized by a computer device and a computer readable storage medium. A processor in the computer device runs a program stored in a memory to complete the whole process of data acquisition, modeling, early warning and analysis according to the above steps; and an information data processing terminal provides a man-machine interface, so that individual users and managers can directly obtain health status reports and early warning information. Thus, the system realizes full-link automation from data acquisition, intelligent analysis to result output, and guarantees the real-time performance and reliability of psychological health monitoring of special occupational groups.

[0069] Embodiment one

[0070] Please refer to Figure 3 The application provides a technical solution: a special occupational group psychological health monitoring and early warning model construction and application, including the following steps:

[0071] Step one: the internal ability is evaluated from two dimensions (physiology, psychology) and seven aspects (past medical history, sleep status, cognitive function, anxiety, depression, life satisfaction, social support and social participation) (through a watch built-in APP and real-time collected HRV, SCR, sleep efficiency and staging, voice content analysis); the environment is evaluated from living environment and working environment (through intelligent sensor monitoring of light, noise, air quality, temperature and humidity, physical hazards (vibration level), mold / allergen level indicators, UVI, ionizing radiation level, etc.), to evaluate the psychological health status of the occupational group.

[0072] Step two: combine the actual situation and apply the scientific and practical evaluation system (based on a multi-factor decision tree model and time series analysis) to evaluate the psychological health status of the occupational group.

[0073] Step three: collect the baseline data of the occupational group through the user registration and real-time monitoring function of the intelligent watch, including basic information, physical activity and sleep condition, etc., and add a daily work efficiency punch-in after work every day (“today's task completion degree” “whether satisfied with today's work” “whether the team support is sufficient”), to analyze the main factors affecting the psychological health status of the occupational group.

[0074] Compared with existing indicators and analysis methods, the method overcomes the problems of lack of professional evaluation and guidance, and the incompleteness of the psychological health indicators of the professional population, and provides a unified psychological health evaluation system for the professional population in different environments and states by developing and applying a special professional population psychological health monitoring and early warning model, improves the applicability, universality and compliance of psychological evaluation, realizes home rehabilitation, and reduces the overall medical cost; and provides a complete solution for screening and evaluation, and completes the prevention and screening of psychological diseases of the professional population comprehensively, and carries out related psychological counseling and positive guidance in advance, which can improve the life enthusiasm and quality of life of the professional population, solves the problems of low screening efficiency of the psychological health status of the professional population, lack of prevention strategies and unified evaluation system, lack of psychological evaluation resources, and can efficiently and low-costly screen the individuals with psychological health problems of the professional population, and reduce the risk of deterioration of psychological problems.

[0075] Embodiment two

[0076] The internal ability in the step one is evaluated from seven aspects (previous medical history, sleep condition, anxiety emotion, depression emotion, life satisfaction, social support and social participation) of two dimensions (physiology and psychology).

[0077] The score index of the physiological health includes:

[0078] (1) Disease: the disease of the adult is determined according to the clinical diagnosis standard. (2) Sleep condition: the sleep condition is evaluated from seven aspects of sleep quality, sleep time, sleep duration, sleep efficiency, sleep disorder, hypnotic drug and daytime dysfunction. The higher the score is, the worse the sleep quality is, and the heart rate variability (HRV), skin conductance response (SCR) and sleep efficiency are collected in real time by wearing a smart watch, a sleep detector and other devices. (3) Cognitive function: the score index of the psychological health includes:

[0079] ① The physiological data such as HRV, SCR and sleep efficiency collected, and the environmental data such as the noise value of the working environment are input into a multi-factor decision tree model, and a psychological health risk index is output. ② A time series model is constructed, and the long-term collected data are analyzed to predict the change trend of the cognitive function related indexes, such as the downward or upward trend of the cognitive function over time.

[0080] The environment in the step one is evaluated from the living environment and the working environment. The environment condition evaluation system model is as follows:

[0081] The total score is 22, and 18-22, 12-17, 11 and below correspond to good, medium, poor and the like, respectively. The specific score details are shown in the following table.

[0082] Environment condition evaluation system

[0083]

[0084] According to the scale score, the influence of the environment of the professional population on the mental health status can be evaluated.

[0085] In the method, the special environmental factors of the professional population are considered in the mental evaluation index, which is beneficial to the unified evaluation of the evaluation system by different professional populations, and is very helpful for collecting, counting and managing the mental evaluation data of the professional population.

[0086] Embodiment three

[0087] In the step two, the mental health status score system is combined with the environment data (residential environment and working environment) evaluation system, and the mental health status of the professional population is comprehensively evaluated.

[0088] In the method, the evaluation system is analyzed and evaluated through practical application, and has good reliability and validity, as well as sensitivity and specificity.

[0089] In the method, the evaluation system is analyzed and evaluated through practical application, and has good reliability and validity, as well as sensitivity and specificity.

[0090] Embodiment four

[0091] In the step three, the baseline data of the professional population are collected, including basic information, sleep condition and the like, and the main factors affecting the mental health status of the professional population are analyzed.

[0092] The collection of the environment data is as follows:

[0093] The embedded terminal is deployed, and the working environment such as hygiene condition, noise condition, CO2 concentration, vibration level, air flow rate and mold level, working intensity and the like and the residential environment such as comfort degree, hygiene condition information are collected, and the environment status of the professional population is evaluated according to the evaluation system.

[0094] The collection of the mental health status is as follows:

[0095] The physiological indexes such as heart rate variability (HRV), skin conductance response (SCR) and sleep efficiency are collected through the smart watch, and are transmitted to the data center through Bluetooth or Wi-Fi, and a dynamic evaluation model is constructed based on AI learning, and the real-time update and risk early warning of the evaluation index are realized by combining the traditional statistical method; a dynamic evaluation system based on natural language processing (NLP), multi-factor decision tree model and emotion recognition algorithm is constructed, the subjective score deviation is corrected by automatically analyzing the questionnaire text semantics, voice tone and facial expression, and the real-time calculation and three-level early warning of the mental health risk index are realized.

[0096] The step four puts forward corresponding solving measures for the factors most related to the mental health outcome, so as to provide a theoretical basis and reference suggestion for the relevant departments to carry out health education and promotion plan.

[0097] Figure 4 The intelligent watch internal circuit schematic diagram provided by the embodiment of the application.

[0098] Embodiment six

[0099] (1) Based on the dynamic data acquisition of the smart watch:

[0100]

[0101] Dynamic evaluation model:

[0102] ① Multi-factor decision tree: input HRV, SCR, sleep efficiency, working environment condition, living environment condition, and output mental health risk index (0-100 points, ≥60 points trigger early warning).

[0103] ② Trend analysis: build a time series model to generate an environmental risk heat map.

[0104] ③ Conduct a self-evaluation of mental health status once a month, including PSQI, MMSE, HAMA, HAMD, and SWLS, and timely feedback the results to the user and store the data.

[0105] (2) Based on the smart watch technology route: collect HRV / SCR / sleep data through the smart watch, transmit to the mobile terminal APP through Bluetooth, and the APP preprocesses the data (denoising, standardization), then calculates the risk index through the multi-factor decision tree to determine whether to push the early warning. When the risk index is ≥60 points or the key parameters (such as continuous noise >85dB for more than 10 minutes, HRV continuously abnormally lower than the individual baseline value by 20%) exceed the threshold value, the watch end immediately triggers the screen red flashing warning (frequency 2Hz) and strong vibration (for 5 seconds), and the APP pushes the detailed early warning information to the user and the management platform.

[0106] Embodiment seven

[0107] Application example and effect verification

[0108] Step 1: Select 30 long-distance truck drivers (25-40 years old, average length of service 6.5 years) in a certain city as the research object. This group of people are long-term exposed to high-intensity work pressure, irregular work and rest, and high noise driving environment. The baseline evaluation shows that the detection rate of anxiety and depression symptoms (anxiety: 25.7%, depression: 22.3%) is significantly higher than the average level of other industries and the data of the Fifth National Physical Fitness Monitoring Report. Through this model, the mental health status of the drivers is evaluated, and the key influencing factors are analyzed.

[0109] Step 2: Continuous monitoring for 8 weeks by wearing smart watches and deploying intelligent sensors in the cockpit, collecting:

[0110] Physiological indicators: average HRV (baseline: 65 ms, after intervention: 72 ms), average sleep efficiency (baseline: 78%, after intervention: 83%), SCR trend.

[0111] Environmental indicators: work environment noise value (average peak noise: 92 dB, over-standard ratio: 35%), work duration (weekly average: 65 hours), light intensity, CO2 (average 1200 ppm, over-standard ratio 40%), vibration level (average 0.8 m / s2, over-standard ratio 25%), TSP (average 8.2 mg / m3, over-standard ratio 12%), PM10 (average 0.32 mg / m3, over-standard ratio 18%), UVI (average ultraviolet index 7, high exposure period ratio 30%), electromagnetic radiation level in the cockpit (complies with safety limits).

[0112] Comparison of smart watch data before and after intervention

[0113]

[0114]

[0115] Figure 5 Comparison of data before and after intervention.

[0116] Figure 6 Trend of data during intervention.

[0117] Subjective feedback: Regularly collect voice diaries and short questionnaires through the APP.

[0118] Step 3: Input HRV, SCR, sleep efficiency, work duration, and noise value into the multi-factor decision tree model in the form of a multi-factor decision tree, dynamically calculate and output the mental health risk index of each driver.

[0119] Model results show that more than 60% of drivers triggered a level 3 warning (index ≥ 60) at least once during the monitoring period, with the main contributing factors being sustained high-intensity driving (> 10 hours / day), excessive noise exposure (> 85 dB for more than 1 hour), and low sleep efficiency (< 80%). Compared to the control group evaluated by traditional questionnaires, the advantages of this model are:

[0120] (1) Improved detection sensitivity: 4 drivers who were later clinically diagnosed with mild anxiety / depression were identified 1-2 weeks earlier than the traditional questionnaire (which did not reach the diagnostic threshold during the same period).

[0121] (2) Improved timeliness of warnings: 85% of acute psychological stress reactions triggered by environmental stressors (such as sudden high noise or continuous overtime) were warned in real time or near real time (delay < 5 minutes), significantly better than traditional methods (which usually take several days to several weeks).

[0122] (3) Enhanced intervention targeting: Based on the risk heat map (such as noise exceeding areas, high-risk periods) and individual warning information output by the model, managers implemented targeted interventions: such as forced rest reminders, distribution of noise-reducing earplugs, and optimization of scheduling to reduce continuous driving time. After 8 weeks of intervention, the average mental health risk index of the driver group decreased by 15%, and the self-reported rates of anxiety / depression symptoms decreased by 18% and 12%, respectively. Through "dynamic monitoring - intelligent evaluation - accurate prediction", this model effectively identifies the risk factors of mental health in professional groups (such as excessive working hours, sleep deprivation, and environmental stressors). The research results provide quantitative basis for developing health promotion plans for professional groups, and can be extended to other high-pressure professional groups (such as medical staff, IT industry, customer service personnel, etc.).

[0123] I. Specific application fields or related products of the invention.

[0124] The present invention focuses on the field of occupational health management, providing real-time mental health monitoring and early warning services for high-pressure and high-risk professional groups through the deep integration of intelligent wearable devices and AI algorithms.

[0125] (1) Core application fields

[0126] Monitoring of mental health of special professional groups: for professional groups working in high-pressure and high-intensity environments or exposed to extreme conditions, such as transportation industry, medical industry, industrial field, technology and service industry, outdoor workers.

[0127] (2) Related products

[0128] ①Smartwatch terminal: physiological sensing (real-time collection of HRV, SCR, sleep staging); environmental perception (built-in microphone to monitor noise, compatible with external sensors); safety design (IP68 waterproof and dustproof, high temperature resistant, vibration resistant, unlocked by facial recognition to ensure data security)

[0129] ②Back-end analysis platform: dynamic assessment system (based on multi-factor decision tree model and NLP algorithm, real-time calculation of mental health risk index); early warning and intervention module (three-level early warning mechanism, group management);

[0130] II. Related evidence of the technical effects obtained by the embodiments of the present application.

[0131] Figure 7 Comparison of traditional mode and smartwatch missed detection rate.

[0132] Figure 8 Comparison of traditional mode and smartwatch elimination of 50% missed detection rate time.

[0133] Figure 9 Comparison of traditional mode and smartwatch intervention completion rate.

[0134] Through strict community trials (Examples 6 and 7) and quantitative data analysis, the present application is significantly superior to the traditional method in timeliness, accuracy and intervention effect.

[0135] 1. Summary of empirical data (based on the 8-week driver test of Example 7)

[0136]

[0137]

[0138] Key conclusions:

[0139] Increased sensitivity: 1-2 weeks earlier identification of 4 mild anxiety patients who were subsequently diagnosed (traditional questionnaire missed detection rate 38%).

[0140] Timeliness change: 85% of environmental stress events (such as sudden noise) achieve a delay of <5 minutes of early warning (traditional method requires 3-5 days).

[0141] 2. Comparison and verification of technical effects

[0142] Evaluation dimensions Traditional questionnaire The present invention Risk identification delay 3-5 days (dependent on manual recovery) <5 minutes (sensor automatically triggers) Missed detection rate 38% (mainly subjective concealment) 4% (multi-modal fusion correction) Intervention coverage 35% (only active reporters) 92% (automatically captured hidden signals)

[0143] Mechanism advantage analysis:

[0144] Dynamic modeling: multi-factor decision tree integrates HRV drop (<50ms), environmental noise (>85dB) and voice emotion (NLP analysis), objective improvement of 40%.

[0145] Closed-loop intervention: tertiary warning triggers immediate action (e.g. mandatory rest), shortens the "assessment-intervention" chain (from weekly to minutes).

[0146] It should be noted that the embodiments of the present application can be realized by hardware, software, or a combination of software and hardware. The hardware portion can be realized by special logic; the software portion can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or a specially designed hardware. Those skilled in the art can understand that the above-mentioned devices and methods can be realized by computer executable instructions and / or included in processor control codes, for example, such codes are provided on a carrier medium, such as a magnetic disk, CD or DVD-ROM, programmable memory, such as read-only memory (firmware), or data carrier, such as optical or electronic signal carrier. The devices of the present application and their modules can be realized by hardware circuit, such as very large scale integrated circuit or gate array, semiconductor, such as logic chip, transistor, etc., or programmable hardware device, such as field programmable gate array, programmable logic device, etc., or by software executed by various types of processors, or by a combination of the above-mentioned hardware circuit and software, such as firmware.

[0147] The above description is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, any modification, equivalent replacement and improvement within the technical range disclosed by the present application, and within the spirit and principle of the present application, should be covered within the protection scope of the present application.

Claims

1. A mental health monitoring and early warning method based on a smart watch combined with artificial intelligence, characterized in that, The method comprises the following steps: Collecting physiological and environmental data of an individual through a smart watch, including heart rate variability, skin conductance response, sleep staging, and noise level; Temporarily storing the collected data locally through a data caching mechanism, and uploading the data to a backend analysis system after network recovery; Building a dynamic evaluation model based on a multi-factor decision tree and natural language processing on the backend, and performing fusion analysis on physiological data and psychological questionnaire text data; Triggering a real-time warning prompt on the smart watch when the calculated mental health risk index exceeds a preset threshold.

2. The method of claim 1, wherein, The warning prompt includes screen flickering, vibration alarm, and application message push.

3. A mental health monitoring and alert system, characterized by, The method comprises: A data collection module for collecting physiological and environmental data through a smart watch and caching the data; An analysis module for performing fusion modeling on physiological data and psychological scale data based on artificial intelligence algorithms, and generating a mental health risk index; An early warning module for issuing a warning signal to the user when the mental health risk index reaches a preset threshold.

4. The system of claim 3, wherein, The analysis module uses a decision tree model combined with an emotion recognition algorithm to dynamically evaluate multi-source data.

5. A method of multi-modal assessment of mental health status, characterized in that, The method comprises: Obtaining physiological parameters and environmental parameters of a user; Analyzing the user's voice tone and text semantics to obtain an emotion recognition result; Inputting the physiological parameters, environmental parameters, and emotion recognition result into a multi-factor model to calculate a mental health risk index.

6. The method of claim 5, wherein, The emotion recognition result is obtained based on joint extraction of acoustic features and text semantics.

7. A smartwatch for mental health risk monitoring, characterized in that, The method comprises: A physiological data collection unit for obtaining heart rate variability, skin conductance response, and sleep staging parameters; An environmental data collection unit for detecting noise intensity in the wearing environment; A storage unit for caching collected data when the network is disconnected; An output unit for providing visual and tactile alarms when the risk index exceeds a threshold.

8. The smart watch of claim 7, wherein, The smart watch further comprises a positioning unit for uploading the location when the user experiences a mental health risk.

9. A group situation analysis system for mental health of a professional group, characterized in that, The method comprises: A data aggregation module for receiving individual data from multiple smart watches; A statistical analysis module for generating a group mental health trend curve; A visualization module for generating an environmental risk heat map and displaying key influencing factors.

10. The system of claim 9, wherein, The statistical analysis module dynamically predicts the mental health trend based on a time series model.

Images