Health monitoring system based on large model and wearable device

Abstract

The invention discloses a health monitoring system based on a large model and wearable equipment, which belongs to the technical field of health monitoring and comprises a data acquisition module, a data preprocessing module, a data feature extraction module, a data feature recognition module and a suggestion generation module. According to the invention, data collected by various health monitoring devices are integrated to monitor and evaluate the health state of the user, and comprehensive health description for the user can be formed.

Description

Technical Field

[0001] The present invention belongs to the technical field of health monitoring, and in particular relates to a health monitoring system based on a large model and a wearable device. Background Art

[0002] With the advancement of science and technology and the improvement of people's health awareness, health monitoring equipment has become an indispensable part of modern life, including smart watches, headphones, and brain-computer interface wearable devices. These health monitoring devices grasp people's physical health conditions through accurate and real-time data monitoring and analysis, and analyze people's physical health conditions based on their health conditions, and propose adaptive health management methods so that people can manage them in a timely manner.

[0003] Current health monitoring equipment has the following shortcomings during use:

[0004] 1. Insufficient data integration: Data collected by various health monitoring devices is often scattered and lacks a unified analysis framework, making it difficult to form a comprehensive description of the user's health;

[0005] 2. Lack of intelligent analysis and reasoning: Health monitoring devices lack deep analysis and reasoning capabilities and are unable to make comprehensive judgments based on complex physiological and psychological data, resulting in limited accuracy and personalization of recommendations;

[0006] 3. Limitations of mental health monitoring: Health monitoring devices mostly focus on physiological data and are unable to effectively assess users’ mental health, especially in terms of mood swings and stress management, and lack targeted advice.

[0007] In view of this, a health monitoring system based on large models and wearable devices is designed to solve the above problems. Summary of the Invention

[0008] To solve the problems raised in the above background technology, the present invention provides a health monitoring system based on a large model and wearable devices, which has the characteristics of being able to form a comprehensive description of the user's health and improve the accuracy and personalization of suggestions.

[0009] To achieve the above objectives, the present invention provides the following technical solutions: a health monitoring system based on a large model and a wearable device, comprising:

[0010] The data acquisition module collects the user's EEG data, motion data, conversation data, and physiological data through wearable devices;

[0011] A data preprocessing module, which preprocesses the collected user data;

[0012] Data feature extraction module, extracts key features from preprocessed user data;

[0013] The data feature recognition module identifies users' cognitive health problems and motor problems based on clustering and regression models, and identifies users' mental problems based on a fusion model of long short-term memory, convolutional neural networks, and random forests. Based on the identified cognitive health problems, motor problems, and mental problems, the user's health status is identified. The expression is:

[0014] y=αx1+βx2+χx3

[0015] Where: α, β and χ are pre-set weight factors, x1 represents the user's cognitive health problem coefficient, x2 represents the user's motor problem coefficient, and x3 represents the user's mental problem coefficient;

[0016] The suggestion generation module generates suggestions based on the identified user health status.

[0017] Furthermore, the exercise data of the data acquisition module includes the number of steps, exercise intensity, heart rate and calories consumed, the conversation data includes voice data, and other physiological data includes skin electrical response and heart rate variability.

[0018] Furthermore, the preprocessing of the data preprocessing module includes data cleaning, denoising, outlier detection and standardization.

[0019] Furthermore, the key features of the EEG data of the data feature extraction module include multi-band energy and spectrum diagrams, the key features of the motion data include exercise load, heart rate variability and calorie consumption rate, and the key features of the conversation data include emotional state and subjective feelings.

[0020] Furthermore, the cognitive health problems of the data feature recognition module include cognitive functions, motor problems include motor status exceeding an appropriate range and poor health recovery indicators, and mental problems include abnormal mental and emotional fluctuations.

[0021] Compared with the prior art, the present invention has the following beneficial effects:

[0022] 1. The present invention integrates data collected by various health monitoring devices to monitor and evaluate the user's health status, and can form a comprehensive description of the user's health.

[0023] 2. Based on the fusion output of multiple models, the present invention can deeply analyze the data collected by various health monitoring devices, and infer adaptive suggestions after dual analysis of physiological and psychological data, thereby improving the accuracy and personalization of the suggestions.

[0024] 3. The present invention monitors the user's health based on physiological and psychological data, which can not only monitor the physical health but also the mental state, thereby improving the comprehensiveness of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a system framework diagram of the present invention;

[0026] In the figure: 1. Data acquisition module; 2. Data preprocessing module; 3. Data feature extraction module; 4. Data feature recognition module; 5. Suggestion generation module. DETAILED DESCRIPTION

[0027] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0028] See attached Figure 1 The present invention provides the following technical solutions: a health monitoring system based on a large model and a wearable device, comprising:

[0029] Data collection module 1, collects the user's EEG data, movement data, conversation data and physiological data through wearable devices;

[0030] Data preprocessing module 2, preprocessing the collected user data;

[0031] Data feature extraction module 3, extracting key features from pre-processed user data;

[0032] Data feature recognition module 4 identifies the user's cognitive health problems and motor problems based on clustering and regression models, identifies the user's mental problems based on the long short-term memory, convolutional neural network and random forest fusion model, and identifies the user's health status based on the identified user's cognitive health problems, motor problems and mental problems. The expression is:

[0033] y=αx1+βx2+χx3

[0034] Where: α, β and χ are pre-set weight factors, x1 represents the user's cognitive health problem coefficient, x2 represents the user's motor problem coefficient, and x3 represents the user's mental problem coefficient;

[0035] The suggestion generating module 5 generates suggestions based on the identified health status of the user.

[0036] Specifically, the exercise data of the data acquisition module 1 includes the number of steps, exercise intensity, heart rate and calories consumed, the conversation data includes voice data, and other physiological data includes skin electrical response and heart rate variability.

[0037] Specifically, the preprocessing of the data preprocessing module 2 includes data cleaning, denoising, outlier detection and standardization.

[0038] Specifically, the key features of the EEG data of the data feature extraction module 3 include multi-band energy and spectrum diagrams, the key features of the motion data include exercise load, heart rate variability and calorie consumption rate, and the key features of the conversation data include emotional state and subjective feelings.

[0039] Specifically, the cognitive health problems of the data feature recognition module 4 include cognitive functions, motor problems include motor status exceeding the appropriate range and poor health recovery indicators, and mental problems include abnormal mental and emotional fluctuations. Specific embodiments

[0041] Data acquisition module 1:

[0042] Applied to monitoring and providing recommendations for users' cognitive health, collecting users' EEG data, heart rate, and respiratory rate through smart watches and brain-computer devices;

[0043] Applied to monitoring and providing recommendations for users' sports health, the system collects users' sports data through smart watches and smart headphones, including steps, exercise intensity, heart rate, and calories burned;

[0044] Applied to user mental health monitoring and advice, wearable natural language processors and smart watches collect user conversation data and physiological data, including voice data, galvanic skin response, and heart rate variability;

[0045] The wearable natural language processor uses voice data to conduct a conversation with the user, asking about the user's daily emotions, stress levels, and daily conversations. This information is recorded in real time and analyzed based on the voice and converted into an evaluation of the user's emotional state and subjective feelings.

[0046] Data preprocessing module 2:

[0047] Applied to user cognitive health monitoring and recommendations, it receives user EEG, heart rate, and respiratory data collected by smart watches and brain-computer devices and performs pre-processing, including data cleaning, denoising, outlier detection, and standardization;

[0048] Applied to user sports health monitoring and recommendations, it receives user sports data collected by smart watches and smart headphones and performs pre-processing, including data cleaning, denoising, outlier detection and standardization;

[0049] Applied to user mental health monitoring and advice, it receives conversation chat data collected by wearable natural language processors and smart watches, as well as user emotional and physical data, and performs pre-processing, including data cleaning, denoising, outlier detection, and normalization;

[0050] Data feature extraction module 3:

[0051] Applied to user cognitive health monitoring and recommendations, extracting features strongly associated with stress and cognitive load from pre-processed user EEG and heart rate data, including EEG spectrum features and ERP;

[0052] Applied to user sports health monitoring and recommendations, extracting features that are highly correlated with pre-processed user sports data and sports characteristics, including exercise load, heart rate variability, and calorie consumption rate;

[0053] Applied to user mental health monitoring and recommendations, it extracts features that are strongly correlated with mental characteristics from pre-processed user emotions and physical data, including emotional state, subjective feelings, galvanic skin response, and heart rate variability;

[0054] Data feature recognition module 4:

[0055] Applied to user cognitive health monitoring and recommendations, identifying user cognitive problems based on features extracted from user EEG data;

[0056] Applied to user sports health monitoring and recommendations, identifying user sports problems based on features extracted from user sports data;

[0057] The specific steps include:

[0058] Randomly select K sample points from several features extracted from the user's EEG data and motion data as the initial cluster centers and record them as u1, u2, ..., u k , where K represents the specified number of clusters;

[0059] The distance from each sample point in the data set to each cluster center is calculated using the Euclidean distance. The expression is:

[0060]

[0061] Where: (x2, y2) represents the coordinates of the sample point, (x1, y1) represents the coordinates of the cluster center;

[0062] Assign each sample point to the cluster with the nearest cluster center;

[0063] Recalculate the cluster center of each cluster, that is, the mean of all sample points in the cluster and update it. The expression is:

[0064]

[0065] Where: C j is represented as cluster j, and x is represented as cluster C j The sample points of |C j|denoted as cluster C j The number of sample points;

[0066] Repeat the above allocation and update steps until the set number of iterations is reached and clustering is completed;

[0067] The clustered features are used as independent variables, and the user's EEG data and movement data are used as dependent variables. The degree of influence of the features on the user's EEG and movement is analyzed through regression equations. The expression is:

[0068]

[0069] Where: α0, α1, ..., α n It is expressed as regression coefficient, x is the independent variable, and y is the dependent variable;

[0070] Based on the impact of the analyzed features on the user's EEG and movement, the key features that affect the user's cognition and movement are determined, and compared with the preset reference data of the user's cognition and movement key features to identify the user's cognitive and movement problems;

[0071] Cognitive problems include excessive cognitive load, decreased reaction speed, and increased stress;

[0072] Motion problems include user motion patterns, health recovery indicators, and anomaly detection, including:

[0073] The exercise mode assesses whether the user's exercise status is within the appropriate range, including whether the exercise intensity is too high or insufficient, and whether the heart rate meets the healthy standard;

[0074] The health recovery indicator evaluates the user's recovery after exercise and determines whether the user has accumulated fatigue or insufficient recovery;

[0075] Anomaly detection avoids the risk of overtraining or insufficient exercise intensity;

[0076] Applied to user mental health monitoring and recommendations, identifying user mental state issues based on features extracted from user emotions and physical data;

[0077] The specific steps include:

[0078] Input the user's emotional data into the long short-term memory network (LSTM) to identify the changing trend of emotions over time;

[0079] The user's emotional data is input into the convolutional neural network (CNN) to identify the changing trend of emotions over space;

[0080] The identified emotion temporal variation features and emotion spatial variation features are fused and input into the random forest to identify the user's current mental state problem;

[0081] Mental problems include abnormal mental and mood swings;

[0082] At the same time, the mental state recognition model is combined with the above-mentioned cognitive and movement recognition models to form an identification of the user's overall health status;

[0083] Suggestion generation module 5:

[0084] Applied to user cognitive health monitoring and recommendations, generating recommendations based on big data for identified user cognitive issues, including:

[0085] Conduct moderate cognitive function training to improve cognitive functions, such as memory, attention, comprehension, judgment, and language skills;

[0086] Add stress-relief exercises to relieve stress, such as meditation, deep breathing, or progressive muscle relaxation;

[0087] Users are advised to engage in moderate physical exercise;

[0088] Use aromatherapy to help with relaxation and stress reduction;

[0089] Users are advised to avoid excessive eating, especially foods high in sugar and fat;

[0090] Applied to user sports health monitoring and recommendations, it generates recommendations based on big data for identified user sports issues, including:

[0091] It is recommended that users reduce exercise intensity appropriately to prevent overtraining and injury;

[0092] Increase the frequency of aerobic exercise, such as brisk walking, running or swimming, to improve cardiopulmonary function;

[0093] Design alternating training plans for users, combining high-intensity training with low-intensity recovery days to promote adequate recovery of the body;

[0094] It is recommended that users incorporate strength training to improve muscle strength and bone density, and enhance overall physical fitness;

[0095] Adjust training time and intensity based on the user's health recovery indicators to ensure maximum training effect without affecting health;

[0096] Provide recommendations for flexibility training, such as dynamic stretching and yoga, to improve muscle recovery after exercise and prevent injuries;

[0097] Encourage users to maintain adequate hydration during exercise to avoid dehydration and maintain peak performance;

[0098] Based on exercise data, users are advised to set gradual improvement goals, such as increasing the number of steps per week or extending the training time, to maintain continuous progress;

[0099] If a user experiences an abnormal heart rate during exercise, it is recommended to stop training immediately and seek medical advice;

[0100] Based on the user's sports interests, it is recommended to join group sports or outdoor activities to increase the fun and persistence of exercise;

[0101] Applied to user mental health monitoring and recommendations, generating recommendations based on big data for identified user mental health issues, including:

[0102] Users are advised to practice deep breathing or meditation to help relieve anxiety;

[0103] Communicate with friends or family members to seek emotional support and relieve psychological stress;

[0104] Users are advised to engage in light aerobic exercise, such as walking or yoga, to improve their emotional state;

[0105] Use mindfulness exercises to help users focus on the present moment and reduce negative thoughts;

[0106] Listen to relaxing music or nature sounds to help soothe your mood and reduce stress;

[0107] Suggest users to set small goals and increase their sense of accomplishment and improve their mood by completing small tasks;

[0108] When you feel stressed, you can try writing a diary to record your emotions and feelings to help release your emotions;

[0109] If negative emotions persist, it is recommended to seek help from a professional psychologist;

[0110] Improve your mood by adjusting your work and rest schedule appropriately and ensuring adequate sleep;

[0111] Join social activities or interest groups to increase positive social interactions and alleviate loneliness.

[0112] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A health monitoring system based on a large model and wearable devices, characterized by: include: Data acquisition module (1), collects the user's EEG data, motion data, conversation data and physiological data through wearable devices; A data preprocessing module (2) performs preprocessing on the user data collected above; Data feature extraction module (3), extracting key features from pre-processed user data; The data feature recognition module (4) identifies the user's cognitive health problems and motor problems based on clustering and regression models, identifies the user's mental problems based on the long short-term memory, convolutional neural network and random forest fusion model, and identifies the user's health status based on the identified user's cognitive health problems, motor problems and mental problems. The expression is: y=αx1+βx2+χx3 Where: α, β and χ are pre-set weight factors, x1 represents the user's cognitive health problem coefficient, x2 represents the user's motor problem coefficient, and x3 represents the user's mental problem coefficient; The suggestion generation module (5) generates suggestions based on the identified user health status.

2. The health monitoring system based on a large model and a wearable device according to claim 1, characterized in that: The exercise data of the data acquisition module (1) includes the number of steps, exercise intensity, heart rate and calories consumed, the conversation data includes voice data, and other physiological data includes skin electrical response and heart rate variability.

3. The health monitoring system based on a large model and a wearable device according to claim 1, characterized in that: The preprocessing of the data preprocessing module (2) includes data cleaning, denoising, outlier detection and standardization.

4. The health monitoring system based on a large model and a wearable device according to claim 1, characterized in that: The key features of the EEG data of the data feature extraction module (3) include multi-band energy and spectrum, the key features of the motion data include motion load, heart rate variability and calorie consumption rate, and the key features of the conversation data include emotional state and subjective feelings.

5. The health monitoring system based on a large model and a wearable device according to claim 1, characterized in that: The cognitive health problems of the data feature recognition module (4) include cognitive function, motor problems include motor status exceeding the appropriate range and poor health recovery indicators, and mental problems include abnormal mental and emotional fluctuations.

Images