Intelligent behavior nursing system for behavior monitoring of old people

Abstract

The invention relates to the technical field of signal devices, and provides an intelligent behavior nursing system for old people behavior monitoring, which comprises a sensor module, an evaluation module, a behavior analysis module and an early warning module, the evaluation module evaluates the state of the old people according to the state data acquired by the sensor module to form an evaluation result, the behavior analysis module triggers the behavior analysis of the old people according to the evaluation result to form a behavior analysis result, and the early warning module sends an early warning prompt to a guardian according to the behavior analysis result. Through mutual cooperation of the sensor module, the evaluation module and the behavior analysis module, the system has a strong personalized monitoring capability, solves the problem of multi-dimensional data isolation, ensures that the system has a multi-dimensional data fusion analysis capability, comprehensively monitors the health state of the old people from different angles, and improves the safety of the old people. And accurate monitoring and evaluation of daily behaviors of each old person are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of signal devices, and in particular to an intelligent behavior care system for monitoring the behavior of the elderly. Background Art

[0002] As the problem of population aging becomes increasingly serious, the need for daily care for the elderly is gradually increasing. Many elderly people are prone to falling or getting lost due to mobility problems or cognitive impairment. Most current care systems rely on manual monitoring or simple monitoring equipment, which makes it difficult to achieve all-weather, efficient behavior monitoring and early warning.

[0003] For example, the video care system based on the behavior patterns of the elderly disclosed in Chinese patent CN206775650U monitors the status of the elderly through video, but ignores the changes in the elderly's own symptoms and cannot achieve specific behavioral care of the person, resulting in a lack of targeting and reliability, and easily leads to unreliable monitoring capabilities.

[0004] In addition, there are the following problems in the prior art:

[0005] 1. Traditional elderly monitoring systems usually use fixed thresholds to detect abnormal behavior. For example, the system may set a standard pace or activity frequency, and trigger an alarm if the elderly's behavior deviates from these standards. Although this method can provide monitoring to some extent, it is often not personalized and difficult to identify individual differences.

[0006] 2. Most existing elderly monitoring systems are only applicable to a single environment, such as indoors or in a specific area. When the elderly enter different environments (such as outdoors, parks), the monitoring system often has difficulty adapting to changes in the scene and cannot adjust the monitoring strategy according to environmental changes.

[0007] 3. Some existing elderly monitoring devices require the elderly to actively wear, operate or manually set them, and many elderly people may find it difficult to adapt to these complex device operations due to their physical condition or cognitive limitations. In addition, some monitoring systems rely too much on cameras and sound monitoring technology, which may infringe on the privacy of the elderly or cause them discomfort.

[0008] The present invention is made to solve the common problems in this field such as insufficient personalized monitoring, inability to comprehensively analyze multi-dimensional data, insufficient intelligent early warning, poor monitoring adaptability, inconvenience in use, low intelligence and insufficient interactive capabilities. Summary of the invention

[0009] The purpose of the present invention is to propose an intelligent behavior care system for monitoring the behavior of the elderly in response to the current deficiencies.

[0010] In order to overcome the shortcomings of the prior art, the present invention adopts the following technical solutions:

[0011] An intelligent behavior care system for monitoring elderly people's behavior, the intelligent behavior care system comprises a server, the intelligent behavior care system also comprises a sensor module, an evaluation module, a behavior analysis module and an early warning module, the server is connected to the sensor module, the evaluation module, the behavior analysis module and the early warning module respectively;

[0012] The sensor module is used to collect status data of the elderly, the evaluation module evaluates the status of the elderly according to the status data collected by the sensor module to form an evaluation result, the behavior analysis module triggers analysis of the elderly's behavior according to the evaluation result to form a behavior analysis result, and the early warning module issues an early warning prompt to the guardian according to the behavior analysis result;

[0013] The evaluation module includes a data fusion unit and an evaluation unit, wherein the data fusion unit performs fusion processing on the status data of the elderly, and the evaluation unit evaluates the status of the elderly according to the fused status data to form an evaluation result;

[0014] The sensor module is worn on the elderly and moves along with the movement of the elderly.

[0015] Optionally, the sensor module includes a vest, a gait monitoring unit, a motion recognition unit, a physiological monitoring unit, an environmental perception unit and a data storage device, wherein the gait monitoring unit, the motion recognition unit, the physiological monitoring unit and the environmental perception unit are arranged on the vest, the gait monitoring unit collects the body posture and gait information of the elderly, the motion recognition unit collects the stability data of the elderly, the physiological monitoring unit collects the physiological state data of the elderly, the environmental perception unit collects the environmental data of the environment in which the elderly are located, and the data storage device stores the collected body posture and gait information, stability data, physiological state data and environmental data;

[0016] The state data includes the body posture and gait information, stability data, physiological state data, and environmental data.

[0017] Optionally, the fusion unit includes a data preprocessing component, a feature extraction component, and a time synchronization component, wherein the data preprocessing component is used to preprocess the state data, the feature extraction component is used to extract gait, movement, physiology and environmental key features from the state data to form core feature data, and the time synchronization component is used to perform time alignment and synchronization processing on the extracted core feature data to form the elderly at t iThe core feature data vector x(t i ).

[0018] Optionally, the evaluation unit obtains the elderly person at t after being processed by the fusion unit. i The core data vector x(t i ), the state index S(x) of the elderly is calculated according to the following formula:

[0019]

[0020] Where x0 is the core feature data vector of the elderly in normal state, that is, the reference value of normal state, D(x(t i ), x0) is t i The Euclidean distance between the feature data at the moment and the normal state feature;

[0021] If the state index S(x) of the elderly is less than the fence warning threshold T set by the system, it means that the elderly's behavior is abnormal, and the behavior analysis module is triggered to analyze the elderly's behavior.

[0022] Optionally, the behavior analysis module includes a behavior feature extraction unit and a behavior analysis unit, wherein the behavior feature extraction unit extracts at least four preprocessed behavior features of the elderly from the fusion unit in real time to form a multidimensional behavior feature vector, and the behavior analysis unit performs analysis based on the multidimensional behavior feature vector to identify abnormalities in the elderly's behavior.

[0023] Optionally, the early warning module includes an early warning unit and an interactive unit, wherein the early warning unit triggers early warning information according to the analysis result, and the interactive unit interactively prompts the manager with the early warning information;

[0024] Among them, the early warning unit includes a comparator and an early warning trigger. The comparator obtains the analysis result and the monitoring threshold Worn set by the system, and compares the analysis result with the set monitoring threshold to form a comparison result. The early warning trigger triggers the early warning information according to the comparison result.

[0025] Optionally, the interaction unit includes a communicator and an identity verifier, wherein the communicator is used to send the warning information to the manager, and the identity verifier identifies the identity of the manager and allows the communicator to send the warning information to the manager after the verification is passed.

[0026] Optionally, the gait monitoring unit is arranged in the lower back or waist area of ​​the vest, and collects the body posture and gait data of the elderly;

[0027] Wherein, the gait monitoring unit includes an accelerometer and a gyroscope, the accelerometer collects acceleration data of the elderly, and the gyroscope collects body posture change data of the elderly.

[0028] Optionally, the motion recognition unit is arranged on the shoulder of the vest and captures the body stability data of the elderly.

[0029] Optionally, the identity verifier compares the identity data collected in real time with the verification data pre-stored in the database of the server. If the identity data and the verification data are the same, the verification is successful, otherwise the verification fails.

[0030] The beneficial effects achieved by the present invention are:

[0031] 1. Through the cooperation between the sensor module, evaluation module, and behavior analysis module, personalized analysis is performed according to the usual behavior patterns, and a personalized monitoring plan suitable for the elderly is established, so that the system has a strong personalized monitoring capability, solves the problem of multi-dimensional data isolation, and ensures that the system has multi-dimensional data fusion analysis capabilities, comprehensively monitors the health status of the elderly from different angles, and ensures accurate monitoring and evaluation of the daily behavior of each elderly person;

[0032] 2. Through the cooperation of the sensor module and the evaluation module, the behavior data of the elderly can be analyzed to ensure that the system can obtain a full range of elderly status data. Through the processing of the evaluation module, the health and behavior status of the elderly can be dynamically monitored, so that the system has a strong data fusion capability and real-time monitoring function;

[0033] 3. Through the cooperation of the evaluation module and the behavior analysis module, the system can accurately analyze the behavior of the elderly, identify abnormal situations, detect changes in the elderly's body and environment, and provide personalized abnormal evaluation based on the elderly's daily habits, thus enhancing the system's intelligence and personalized monitoring capabilities;

[0034] 4. Through the cooperation between the behavior analysis module and the early warning module, managers or guardians can dynamically control the behavior of the elderly, ensuring that the entire system has the advantages of good personalized monitoring capabilities, high intelligence, strong multi-dimensional data fusion capabilities, reliable early warning analysis capabilities, high ease of use and strong interactive capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The present invention can be further understood from the following description in conjunction with the accompanying drawings. The components in the figures are not necessarily drawn to scale, but the emphasis is placed on illustrating the principles of the embodiments. In different views, the same reference numerals designate the same parts.

[0036] Figure 1It is an overall block diagram of the present invention.

[0037] Figure 2 It is a block diagram of the sensor module and the elderly of the present invention.

[0038] Figure 3 It is a block diagram of the state data collected by the sensor module of the present invention and the fusion unit.

[0039] Figure 4 It is a schematic structural diagram of the vest and the sensor module of the present invention.

[0040] Figure 5 It is a front view schematic diagram of the vest and the sensor module of the present invention.

[0041] Figure 6 Schematic side view of the vest and sensor module of the present invention.

[0042] Figure 7 for Figure 6 Schematic cross-sectional view at AA in the middle.

[0043] Explanation of the accompanying drawings: 1. vest; 2. setting head; 3. chest strap sensor; 4. temperature sensor; 5. air quality sensor; 6. humidity sensor; 7. heart rate sensor; 8. blood oxygen saturation sensor; 9. blood pressure sensor; 10. vibration motor; 11. electrode patch; 12. accelerometer; 13. light sensor; 14. body temperature sensor. DETAILED DESCRIPTION

[0044] The following is an explanation of the embodiments of the present invention through specific embodiments. Those skilled in the art can understand the advantages and effects of the present invention from the contents disclosed in this specification. The present invention can be implemented or applied through other different specific embodiments, and the details in this specification can also be modified and changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. In addition, the drawings of the present invention are only simple schematic illustrations and are not depicted according to actual sizes. It is stated in advance. The following embodiments will further explain the relevant technical contents of the present invention in detail, but the disclosed contents are not intended to limit the scope of protection of the present invention.

[0045] Embodiment 1: According to Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 , Figure 6 , Figure 7 As shown, this embodiment provides an intelligent behavior care system for elderly behavior monitoring, the intelligent behavior care system includes a server, the intelligent behavior care system also includes a sensor module, an evaluation module, a behavior analysis module and an early warning module;

[0046] The sensor module is used to collect status data of the elderly, the evaluation module evaluates the status of the elderly according to the status data collected by the sensor module to form an evaluation result, the behavior analysis module triggers analysis of the elderly's behavior according to the evaluation result to form a behavior analysis result, and the early warning module issues an early warning prompt to the guardian according to the behavior analysis result;

[0047] The server is connected to the sensor module, the evaluation module, the behavior analysis module and the early warning module respectively, and stores the intermediate data and process data of the sensor module, the evaluation module, the behavior analysis module and the early warning module in the database of the server for query and call;

[0048] The intelligent behavior care system also includes a central processor, which is respectively connected to the sensor module, the evaluation module, the behavior analysis module and the early warning module, and centrally controls the sensor module, the evaluation module, the behavior analysis module and the early warning module based on the central processor, so as to store the control data of the central processor in the database of the server, thereby improving the accuracy and reliability of the entire system in monitoring and caring for the elderly behavior;

[0049] The evaluation module includes a data fusion unit and an evaluation unit, wherein the data fusion unit performs fusion processing on the status data of the elderly, and the evaluation unit evaluates the status of the elderly according to the fused status data to form an evaluation result;

[0050] Wherein, the sensor module is worn on the elderly and moves along with the movement of the elderly;

[0051] Optionally, the sensor module includes a vest 1, a gait monitoring unit, a motion recognition unit, a physiological monitoring unit, an environmental perception unit and a data storage device, wherein the gait monitoring unit, the motion recognition unit, the physiological monitoring unit and the environmental perception unit are arranged on the vest 1, the gait monitoring unit collects the body posture and gait information of the elderly, the motion recognition unit collects the stability data of the elderly, the physiological monitoring unit collects the physiological state data of the elderly, the environmental perception unit collects the environmental data of the environment in which the elderly are located, and the data storage device stores the collected body posture and gait information, stability data, physiological state data and environmental data;

[0052] Wherein, the state data includes the body posture and gait information, stability data, physiological state data, and environmental data;

[0053] like Figure 7As shown, the gait monitoring unit is arranged in the lower back or waist area of ​​the vest 1, and collects the body posture and gait data of the elderly;

[0054] Wherein, the gait monitoring unit includes a pedometer, an accelerometer 12 and a gyroscope, the pedometer is used to record the number of steps of the elderly, the accelerometer 12 collects the acceleration data of the elderly, and the gyroscope collects the body posture change data of the elderly;

[0055] Among them, the pedometer records the number of steps walked is a conventional means, so in this embodiment, it is not repeated one by one;

[0056] Optionally, the motion recognition unit is arranged at the shoulder of the vest 1 and captures the body stability data of the elderly;

[0057] The physiological monitoring unit includes a heart rate sensor 7, a blood oxygen saturation sensor 8, a blood pressure sensor 9, a body temperature sensor 14, and a data collector. The heart rate sensor 7 is used to collect heart rate data of the elderly, the blood oxygen saturation sensor 8 is used to collect blood oxygen saturation data of the elderly, the blood pressure sensor 9 is used to collect blood pressure data of the elderly, and the body temperature sensor 14 collects body temperature data of the elderly. The data collector collects the heart rate data collected by the heart rate sensor 7, the blood oxygen saturation data collected by the blood oxygen saturation sensor 8, the blood pressure data collected by the blood pressure sensor 9, and the body temperature data collected by the body temperature sensor 14;

[0058] Among them, the heart rate sensor 7 (ECG or PPG sensor) is installed on the front chest or collar of the vest 1 to monitor the heart rate of the elderly and detect the heart rate and rhythm; the blood oxygen saturation sensor 8 (SpO2 sensor) is installed on the shoulder or collar of the vest 1 to monitor the blood oxygen content by detecting skin blood flow fluctuations; the blood pressure sensor 9 (pulse wave sensor) is installed on the side waist or lower back area of ​​the vest 1 to estimate the blood pressure status of the elderly through pulse wave analysis; the body temperature sensor 14 is installed in the armpit, back or neck of the vest 1 to obtain real-time body temperature data by monitoring the skin surface temperature;

[0059] The environment sensing unit includes a temperature sensor 4, a humidity sensor 6, a light sensor 13, a light sensor 13, an air quality sensor 5 and an air pressure sensor. The temperature sensor 4 is used to monitor the temperature of the environment in which the elderly live, to ensure that the ambient temperature is in a comfortable range and to prevent overcooling or overheating. The humidity sensor 6 is used to detect the humidity in the ambient air to prevent excessively high or low humidity from affecting the health of the elderly. The light sensor 13 is used to monitor the ambient light intensity to ensure that the light in the living environment of the elderly is suitable and to avoid the influence of excessively strong or weak light on vision or activities. The air quality sensor 5 is used to monitor pollutants in the air, such as PM2.5, CO2 or VOC, to ensure that the air is clean and suitable for breathing. The air pressure sensor is used to monitor the ambient air pressure, especially when the air pressure fluctuates violently, to determine whether the elderly are in a suitable air pressure range;

[0060] The air quality sensor 5 includes a PM2.5 sensor and a VOC sensor. The PM2.5 sensor is used to collect the concentration of particulate matter in the environment, and the VOC sensor is used to monitor harmful gases in the environment.

[0061] Optionally, the fusion unit includes a data preprocessing component, a feature extraction component, and a time synchronization component, wherein the data preprocessing component is used to preprocess the state data, the feature extraction component is used to extract gait, movement, physiology and environmental key features from the state data to form core feature data, and the time synchronization component is used to perform time alignment and synchronization processing on the extracted core feature data to form the current core feature data vector x of the elderly;

[0062] The data preprocessing component preprocesses the state data according to the following formula (steps S1 to S3):

[0063] S1. Data collection:

[0064] Acquire the body posture and gait information collected by the gait monitoring unit, the stability data collected by the motion recognition unit, the physiological state data collected by the physiological monitoring unit, and the environmental data collected by the environmental perception unit;

[0065] The gait monitoring unit collects the body posture and gait information of the elderly, including acceleration data and gyroscope data; it determines the pace, speed and gait stability by collecting the body's acceleration and angular velocity;

[0066] The motion recognition unit collects stability data of the elderly and captures the elderly's walking, standing, sitting, and bending movements through a motion recognition sensor (such as an accelerometer 12);

[0067] The physiological monitoring unit collects physiological status data of the elderly, such as heart rate, blood oxygen saturation, and blood pressure physiological parameters;

[0068] The environmental sensing unit collects environmental data around the elderly, including temperature, humidity, light intensity, air quality (PM2.5, VOC), and air pressure;

[0069] S2. Data cleaning:

[0070] Identify and remove noisy data, erroneous data, or outliers;

[0071] The specific processing is as follows:

[0072] Identifying noisy data:

[0073] Perform noise detection on the data of the accelerometer 12 and the gyroscope sensor, and use filtering technology (such as low-pass filtering or Kalman filtering) to remove high-frequency interference signals;

[0074] For example, gait monitoring data may be disturbed by occasional jitter, and these high-frequency jitters can be eliminated by filtering;

[0075] Identifying erroneous data:

[0076] Monitoring possible data errors of sensors; for example, the heart rate sensor 7 may record abnormal heart rate data due to poor contact; identifying erroneous data by setting a reasonable threshold range (for example, heart rate data that is too high or too low may be invalid);

[0077] Remove outliers:

[0078] Use statistical methods (such as the 3-times standard deviation method or the box plot method) to detect and delete abnormal data that is beyond the normal range; for example, the ambient temperature data may occasionally jump to extremely high or low values, and these abnormal points need to be deleted;

[0079] S3. Missing data completion:

[0080] Handle missing data due to sensor failure or data transmission interruption to ensure data continuity;

[0081] Missing data completion methods help ensure data continuity; here are some common processing methods:

[0082] Linear interpolation: For continuous data that is missing in a short period of time (such as gait data and heart rate data), linear interpolation is used to complete it; that is, linear estimation is performed based on the points before and after the missing data; for example, assuming that the heart rate sensor 7 loses connection for a short period of time, the heart rate data before and after are used for interpolation;

[0083] Mean filling: For signals with relatively stable environmental data (such as temperature and humidity), the mean of the previous time period is used to fill missing values; this is suitable for environmental data that changes slowly;

[0084] In this embodiment, the current position of the elderly is set as the coordinate origin, and a spatial rectangular coordinate system xyz is established, wherein the X axis: the acceleration in the left-right direction, the Y axis: the acceleration in the front-back direction, and the Z axis: the acceleration in the vertical direction (up and down);

[0085] The feature extraction component extracts key features from gait, movement, physiology, and environmental state data to form core feature data. The specific extraction steps (S11 to S15) are as follows:

[0086] S11. Gait feature extraction:

[0087] The gait monitoring unit collects the gait data of the elderly through the accelerometer 12 and the gyroscope, including acceleration and angular velocity information; the steps for extracting gait features are as follows:

[0088] S111, extracting gait cycle:

[0089] By analyzing the data from the acceleration sensor, the periodic changes in the steps are detected and the gait cycle (i.e. the beginning and end of each step) is identified;

[0090] Specifically, the acceleration signal in the vertical direction (usually the Z-axis acceleration) is periodically detected, and the cycle of each step is identified by detecting the peaks and troughs;

[0091] S112, extracting pace:

[0092] Gait speed is obtained by calculating the time interval between each step (i.e., gait cycle);

[0093] Specifically, the average stride length of the elderly is calculated through the historical records of the number of steps and movement distance of the elderly over a period of time. This method directly extracts features from historical walking data and avoids the complexity of using an estimation model. The specific process is as follows:

[0094] 1) Collect historical data:

[0095] First, we need to collect the walking data of the elderly, including:

[0096] Step count: The total number of steps taken by the elderly over a period of time is recorded by the accelerometer 12 of the gait monitoring unit or the pedometer;

[0097] Walking distance: GPS is used to record the total distance that the elderly have walked during this period;

[0098] Select historical data for one day, one week or longer to ensure that the step size calculation is stable and accurate enough;

[0099] 2) Calculate the average step length:

[0100] After collecting enough historical data, the average stride length SL of the elderly is calculated using the following formula:

[0101]

[0102] In the formula, distance is the total moving distance (unit: meter) of the elderly in a period of time recorded by the GPS device, and STEPN is the total number of steps of the elderly in the same period of time recorded by the pedometer;

[0103] S113, gait stability characteristics:

[0104] By analyzing the volatility of acceleration data, the stability characteristics of gait are extracted; elderly people with unstable gait may show large acceleration changes and angular velocity fluctuations;

[0105] Calculate the standard deviation of acceleration and angular velocity. A larger standard deviation indicates a more unstable gait. For example, the standard deviation of vertical acceleration reflects the stability of gait.

[0106] Specific:

[0107] 1) StepSpeed ​​is calculated by step length and gait cycle:

[0108]

[0109] In the formula, SL is the average step length of each step of the elderly, obtained through historical data, in meters (m), and StepTime is the time interval between each step of the elderly, in seconds (s);

[0110] The gait cycle refers to the time it takes for an elderly person to take each step. The periodic changes of each step are detected by analyzing the peaks and troughs of the acceleration sensor data.

[0111] The gait period T is defined as the time interval between two consecutive foot strikes:

[0112] T=t step(n+1) -t step(n) ;

[0113] Where, t step(n) and t step(n+1) are the time points of step n and step n+1 respectively, in seconds (s);

[0114] The time point of each step is detected through the peak in the acceleration data (usually the maximum acceleration when the foot lands), so as to calculate the gait cycle of each step;

[0115] The gait stability index is measured by analyzing the volatility of acceleration and angular velocity data; generally speaking, the more stable the gait, the smaller the fluctuation of acceleration and angular velocity, so the stability of gait is quantified by calculating the standard deviation of acceleration and angular velocity;

[0116] 2) Formula for standard deviation of acceleration:

[0117] Assume that within a period of time, the acceleration data in the vertical direction (such as the Z axis) are a1, a2, …, a n , then the standard deviation of acceleration σ a The calculation formula is:

[0118]

[0119] In the formula, a i is the acceleration value at the i-th moment, is the average acceleration during this period:

[0120]

[0121] Where n is the total number of samples of acceleration data;

[0122] 3) Formula for standard deviation of angular velocity:

[0123] Similarly, suppose that in the same time period, the angular velocity data measured in a certain direction (such as rotation around the X axis) are ω1, ω2, …, ω m , then the standard deviation of the angular velocity σ ω The calculation formula is:

[0124]

[0125] In the formula, ω i is the angular velocity value at the i-th moment, is the average angular velocity:

[0126]

[0127] Where m is the total number of samples of angular velocity data;

[0128] 4) Calculate the gait stability index:

[0129] Acceleration standard deviation σ a and the standard deviation of angular velocity σ ω As an indicator of gait stability; the larger the standard deviation, the more unstable the gait;

[0130] The gait stability index S is obtained by combining the standard deviation of the vertical acceleration with the standard deviation of the angular velocity. 稳定性 ;

[0131] Specifically, the gait stability index S 稳定性 Calculate according to the following formula:

[0132]

[0133] S12, action feature extraction:

[0134] The action recognition unit mainly captures the activity status of the elderly, including standing, walking, sitting, and bending;

[0135] The steps of action feature extraction are as follows:

[0136] S121, Action Classification:

[0137] By classifying accelerometer data, the elderly’s action state (e.g., standing, walking, sitting, bending) can be identified;

[0138] Use thresholds to classify acceleration data and extract signature features of specific actions; for example, when sitting down, the acceleration in the Z-axis direction changes significantly, and when bending over, the acceleration in the Y-axis direction changes significantly;

[0139] Each action will cause specific changes in the acceleration on different axes. Based on these changes, different action states are identified by setting different thresholds;

[0140] In this embodiment, a process for identifying an action type and an action frequency is provided, specifically:

[0141] 1) Identification of standing state

[0142] In the standing state, the body is still and the acceleration changes little; especially on the Z axis, the acceleration caused by gravity is about 9.8m / s 2 ;

[0143] Identification method: When the acceleration of the X and Y axes is small (close to 0m / s 2 ), and the Z-axis acceleration is close to 9.8m / s 2 When the elderly are in a standing position,

[0144] The threshold conditions are:

[0145] |a x |<0.5m / s 2 And|a y |<0.5m / s 2 And|a z -9.8|<1m / s2 ;

[0146] In the formula, a x is the acceleration along the x-axis, and its value is measured by the accelerometer, that is, the component of the accelerometer along the x-axis, a y is the acceleration along the y-axis direction, and its value is measured by the accelerometer, that is, the component along the y-axis direction, a z is the acceleration along the z-axis, and its value is measured by the accelerometer, that is, the component along the z-axis;

[0147] 2) Identification of walking status

[0148] When walking, the acceleration on the x, y, and z axes will show periodic changes, especially the acceleration on the z axis has large fluctuations, indicating the up and down movement of the footsteps;

[0149] Identification method: When the periodic fluctuation of z-axis acceleration is detected, it is judged that the elderly are walking;

[0150] A threshold is set, and when the z-axis acceleration is detected to fluctuate periodically within a certain range, it is determined to be in walking state;

[0151] Threshold condition: 1m / s 2 <a z <15m / s 2 And the acceleration fluctuates periodically;

[0152] In the formula, a z is the acceleration along the z-axis, and its value is measured by the accelerometer, that is, the component along the z-axis;

[0153] That is, the periodic changes are determined by detecting the peaks and troughs of the z-axis;

[0154] 3) Identification of sitting status:

[0155] When an elderly person sits down, the acceleration on the z-axis suddenly decreases, indicating that the body moves downward, and the acceleration on the x-axis or y-axis also changes significantly;

[0156] Identification method: If the z-axis acceleration decreases rapidly and reaches a certain lower value, it is determined that the elderly person is sitting down;

[0157] Threshold Condition: And a z <6m / s 2 ;

[0158] In the formula, da z / dt represents the rate of change of z-axis acceleration. If the z-axis acceleration decreases rapidly and reaches a lower value, it is judged as sitting down;

[0159] 4) Identification of the bent state (indicating that the body is leaning forward without significant acceleration changes in the vertical direction):

[0160] When bending over, the acceleration on the y-axis (front and back direction of the body) will change significantly, usually with a positive value, indicating that the body is leaning forward;

[0161] Identification method: When the acceleration of the y-axis increases and the acceleration of the z-axis remains relatively stable, it is judged as a bending state;

[0162] Threshold condition: a y >2m / s 2 And|a z -9.8|<1m / s 2 ;

[0163] In the formula, a y is the acceleration along the y-axis direction, and its value is measured by the accelerometer, that is, the component along the y-axis direction, a z is the acceleration along the z-axis, and its value is measured by the accelerometer, that is, the component along the z-axis;

[0164] S122, action frequency A f :

[0165] Extract the frequency of different actions, such as the number of times you sit down / stand up, the number of times you stand up;

[0166] By performing temporal analysis on the classified action data, the number of occurrences of a specific action can be calculated; for example, the number of times an elderly person stands up or sits down each day can be recorded;

[0167] After identifying the elderly in a certain action state, the action frequency is calculated by counting the number of times the same action occurs; the specific process is as follows:

[0168] Action duration T:

[0169] The start and end of each action are determined by the obvious characteristics of the acceleration change; for example, the duration of the walking state can be determined by detecting the periodic changes in the Z-axis acceleration;

[0170] Action count times:

[0171] For the action frequency A f Calculation, counting the number of times an action occurs within a fixed time window (such as 1 minute or 1 hour);

[0172]

[0173] In the formula, times is the number of times an action occurs, and T is the duration of the action;

[0174] For example, if an elderly person sits down 5 times in 1 hour, the frequency of sitting down is 5 / 1 = 5 times / hour;

[0175] S13. Physiological feature extraction:

[0176] Extracted features include: heart rate (average heart rate HR avg , heart rate variability), blood oxygen saturation, blood pressure, body temperature;

[0177] The physiological monitoring unit collects the heart rate, blood oxygen, blood pressure, and body temperature data of the elderly. The specific extraction steps are as follows:

[0178] S131, Heart rate characteristics:

[0179] The heart rate characteristics of the elderly are extracted through a heart rate sensor (ECG or PPG), including the average heart rate and heart rate variability;

[0180] The raw data that can be directly obtained through ECG (electrocardiogram) or PPG (photoplethysmogram) sensors are usually the time interval between heartbeats (RR interval); specifically:

[0181] Directly collected raw data: RR intervals are the time interval between two consecutive heartbeats, in seconds (s);

[0182] RR interval collection method: The ECG sensor detects the heart electrical signal (i.e., the R wave of the electrocardiogram signal) and accurately calculates the time interval between two consecutive R waves;

[0183] The PPG sensor detects blood flow fluctuations through light, collects pulse signals, and then obtains the time interval between pulse beats;

[0184] The RR interval collected by the sensor is represented as a series of time data: t1, t2, t3, …, t n (Unit: seconds)

[0185] Among them, t i is the interval between the i-th heartbeats;

[0186] Calculate the mean and standard deviation of the heart rate signal to obtain the average heart rate and heart rate fluctuation range of the elderly;

[0187] Heart rate variability (HRV) assesses heart health by analyzing the time differences between heart beats;

[0188] Calculate average heart rate HR avg : The average heart rate is calculated from the RR interval data collected directly by the sensor, that is:

[0189]

[0190] Where, t i is the time between the i-th heartbeat and the i+1-th heartbeat directly collected by the sensor (i.e., the RR interval), in seconds (s), and n is the number of heartbeats;

[0191] Calculate heart rate variability (HRV):

[0192] Heart rate variability is calculated using the RR interval data collected by the sensor, and the main indicators include SDNN (standard deviation) and RMSSD (root mean square of the difference between adjacent RR intervals);

[0193] SDNN (standard deviation of RR intervals): SDNN is one of the most commonly used HRV indicators for evaluating heart rate variability by calculating the standard deviation of the RR intervals. In the embodiment, SDNN is used as an indicator of heart rate variability HRV, and its value is calculated according to the following formula:

[0194]

[0195] Where, t i is the interval between each heartbeat collected by the sensor (i.e., RR interval), h is the number of RR intervals, is the average value of the RR interval, and its value is calculated according to the following formula:

[0196]

[0197] Where, t i is the RR interval collected by the sensor, h is the number of RR intervals;

[0198] RMSSD (root mean square of differences between adjacent RR intervals) is calculated according to the following formula:

[0199]

[0200] Where h is the number of RR intervals, t i is the RR interval of the i-th heartbeat (the time interval between two consecutive heartbeats, usually in milliseconds), t i+1 is the RR interval of the i+1th heartbeat.

[0201] S132, blood oxygen saturation:

[0202] Extract the average and minimum values ​​of blood oxygen saturation to identify fluctuations in blood oxygen levels of the elderly during monitoring;

[0203] Analyze blood oxygen sensor data and extract the average blood oxygen saturation of the elderly. Low blood oxygen levels may indicate that the elderly have physical problems;

[0204] In this embodiment, based on the blood oxygen saturation value (SpO2) continuously measured by the sensor, the average value of the elderly during the monitoring period is further calculated to analyze the fluctuation of their blood oxygen level;

[0205] At the same time, in this embodiment, the average value of blood oxygen saturation SpO2 avg as a characteristic of blood oxygen saturation;

[0206] Among them, the average blood oxygen saturation refers to the average level of blood oxygen saturation of the elderly during a period of monitoring time, and its value is calculated according to the following formula:

[0207]

[0208] Where m is the total number of SpO2 data points collected, SpO2(t i ) is at the i-th time point t i Measured blood oxygen saturation value;

[0209] S133, blood pressure characteristics:

[0210] Extract blood pressure characteristics of the elderly, including the mean and fluctuation range of systolic and diastolic blood pressure;

[0211] Through the blood pressure sensor, the mean and standard deviation of blood pressure over a period of time are calculated to identify whether the blood pressure fluctuation is abnormal;

[0212] S134, Body temperature characteristics:

[0213] Extract the body temperature characteristics of the elderly and determine whether the body temperature is abnormal;

[0214] Through the body temperature sensor, the average body temperature is calculated; abnormal body temperature may indicate fever or hypothermia;

[0215] S14, Environmental feature extraction:

[0216] The environmental sensing unit collects the temperature, humidity, light intensity, air quality (PM2.5, VOC), and air pressure data of the elderly's environment. The environmental feature extraction steps are as follows:

[0217] S141, Ambient temperature characteristics:

[0218] Extract the temperature characteristics of the elderly's environment and calculate the average temperature;

[0219] Monitor temperature data and extract the average ambient temperature Tem avg , its value is calculated according to the following formula:

[0220]

[0221] Where, Temi is the temperature sensor reading at the i-th time point, and a is the number of samples collected by the temperature sensor;

[0222] S142, Humidity characteristics:

[0223] Extract environmental humidity characteristics to ensure that the humidity is within a comfortable range;

[0224] Monitor the ambient humidity through the humidity sensor and extract the average ambient humidity H avg , its value is calculated according to the following formula:

[0225]

[0226] In the formula, H i is the humidity sensor reading at the i-th time point (unit: %RH), b is the number of samples collected by the humidity sensor;

[0227] S143, Light intensity characteristics:

[0228] Extract the mean value of light intensity to determine whether the light is appropriate;

[0229] Analyze the light sensor data and extract the average light intensity Light avg , its value is calculated according to the following formula:

[0230]

[0231] Where, L i is the light intensity sensor reading at the i-th time point (unit: lux) c is the number of samples collected by the light intensity sensor;

[0232] S144. Air quality characteristics:

[0233] Extract key features of air quality, focusing on PM2.5 and VOC concentrations;

[0234] The average value of air pollutant concentration is extracted through PM2.5 and VOC sensors to identify whether the air quality meets the standard; in this embodiment, if a PM2.5 sensor is used, the average PM2.5 avg Calculate according to the following formula:

[0235]

[0236] Where, PM i is the PM2.5 sensor reading at the i-th time point (unit: μg / m 3 ), d is the number of samples collected by the PM2.5 sensor;

[0237] In this embodiment, if a VOC sensor is used, the average VOC of POV avg Calculate according to the following formula:

[0238]

[0239] Where VOCi is the VOC sensor reading at the i-th time point (unit: ppm), and E is the number of samples collected by the VOC sensor;

[0240] S145, Air pressure characteristics:

[0241] Extract the ambient air pressure and determine whether it is within the normal air pressure range;

[0242] Analyze the air pressure sensor data and extract the mean air pressure P avg , its value is calculated according to the following formula:

[0243]

[0244] Where P i is the pressure sensor reading at the ith time point (unit: hPa), and f is the number of samples collected by the pressure sensor;

[0245] S15, constructing the extracted key features into a core feature data to represent the current state of the elderly;

[0246] In this embodiment, the core characteristic data includes the step speed StepSpeed, the gait stability S 稳定性 , Action frequency A f , average heart rate HR avg , RR interval standard deviation SDNN, average blood oxygen saturation SpO2 avg 、Ambient temperature mean Tem avg 、Ambient humidity average H avg , the average light intensityLight avg 、The average PM2.5 avg , the mean value of air pressure P avg ;

[0247] The time synchronization component includes a data caller and a synchronization processor, wherein the data caller calls the core characteristic data processed by the preprocessing component, and the synchronization processor performs time alignment and synchronization processing on the called core characteristic data;

[0248] The synchronization processor performs time alignment and synchronization processing on the core characteristic data according to the following steps:

[0249] S21. It is necessary to obtain feature data at different times from each sensor; each sensor may have a different sampling frequency and timestamp; for example:

[0250] Gait speed and gait stability may be acquired by accelerometers and gyroscopes at high frequency (e.g., multiple times per second);

[0251] Heart rate and blood oxygen saturation are usually collected by a heart rate sensor and a blood oxygen sensor every few seconds or tens of seconds;

[0252] Environmental data (temperature, humidity, light, PM2.5, air pressure) are collected every few seconds by respective environmental sensors;

[0253] The data collection time of different sensors may not be synchronized, so it is necessary to align these data to a unified time axis;

[0254] S22. Define a unified timeline:

[0255] A unified timeline needs to be defined, usually based on the system timestamp (e.g., once a sample per second); all sensor feature data needs to be aligned to this timeline;

[0256] For example, assuming that the time interval you define is 1 second, then you need to get a core feature data vector x every 1 second;

[0257] The time axis is t1, t2, t3, ..., t n Indicates the 1st second, the 2nd second, ..., the nth second;

[0258] S23, Time Alignment:

[0259] Time alignment is performed for the characteristic data of each sensor; the specific processing method is as follows:

[0260] Interpolation: For sensor data with low sampling frequency (such as heart rate, blood oxygen saturation, and environmental data), use an interpolation algorithm to interpolate it to a unified time axis; if the heart rate data has a value every 5 seconds, but the system needs a heart rate value every second, use linear interpolation to estimate the heart rate data per second;

[0261]

[0262] Where t is the interpolation time point, t i and t i+1 is the time between adjacent known heart rate data points, HR avg (t) is the average heart rate at time t, HR avg (t i ) is the time point t i Average heart rate measured at avg (ti+1 ) is the time point t i+1 The average heart rate measured at

[0263] Assume that the sampling frequency of the blood oxygen saturation (SpO2) sensor is low, and the sensor provides data at a certain interval (for example, every 5 seconds), and needs to be aligned to the data point per second; in this case, linear interpolation is used to estimate the blood oxygen saturation value per second; for two adjacent blood oxygen saturation data points SpO2 (t i ) and SpO2(t i+1 ), assuming that it is necessary to i and t i+1 Insert a new data point between

[0264] The linear interpolation formula of blood oxygen saturation is:

[0265]

[0266] Where t is the interpolation time point, t i and t i+1 is the time between adjacent known heart rate data points, SpO2 avg (t) is the average blood oxygen saturation value at time t, SpO2 avg (t i ) is the time point t i The average blood oxygen saturation value measured at avg (t i+1 ) is the time point t i+1 The average blood oxygen saturation value measured at

[0267] Similar to blood oxygen saturation, environmental data (such as temperature, humidity, light intensity, PM2.5, and air pressure) are usually collected through low-frequency sensors; in order to align to the time point of each second, linear interpolation method is still used to estimate the missing data;

[0268] For each environmental characteristic data, the formula is similar to that of blood oxygen saturation;

[0269] Assuming that X(t) represents the data of a certain environmental characteristic (such as temperature, humidity, light intensity, PM2.5 and air pressure), the formula for linear interpolation of the data of a certain environmental characteristic is:

[0270]

[0271] Where t is the interpolation time point, t i and t i+1 is the time of the adjacent known heart rate data points, X(t) is the interpolated value at time point t (for example: temperature, humidity, light, PM2.5, air pressure), X(t i) and X(t i+1 ) is the adjacent known environmental data value;

[0272] Resampling: For data sampled at a high frequency (such as pace and gait stability), downsampling, downsampling or averaging is performed; if there are multiple sampling points in a certain second (for example, 10 samplings), the average value of these sampling points is taken as the representative value of that second;

[0273]

[0274] Where, X avg (t) is the average value after resampling (such as walking speed, gait stability), X i is the value of the i-th sampling point, and k is the number of sampling times in this time period;

[0275] When the average value of the pace after resampling is calculated, the data obtained by pace collection is substituted into the above formula to calculate the average value of the pace after resampling;

[0276] S24, synchronous processing:

[0277] After time alignment, all sensor data will be processed synchronously at the same time point; at this time, each time point (such as t1, t2, ..., t n ) have feature data from different sensors:

[0278] S25. Construct core feature data vector:

[0279] At each time point t i After time alignment and synchronization, all feature data are combined into an elderly person at t i The core feature data vector x(t i ):

[0280] x(t i )=[StepSpeed(t i ),S 稳定性 (t i ),A f (t i ),HR avg (t i ),SDNN(t i ),SpO2 avg (t i ),Tem avg (t i ),H avg (t i ),Light avg (t i ),PM avg (t i),P avg (t i )];

[0281] Where, StepSpeed(t i ) is the elderly at time point t i The pace at t is given by i The pace data in a certain time window before is obtained by downsampling, usually expressed as the average pace in the time window, S 稳定性 (t i ) is an indicator for evaluating the gait stability of the elderly based on the fluctuations of acceleration and angular velocity when they walk. It reflects whether the gait is stable. It is usually calculated using the standard deviation of acceleration and angular velocity. The larger the standard deviation, the more unstable the gait. f (t i ) refers to the elderly at time point t i The frequency of corresponding actions (such as standing, walking, sitting, bending) performed within a certain time window before, HR avg (t i ) is the time point t i The average heart rate of the elderly in the previous time window. The heart rate sensor usually collects heart rate data at regular intervals. The average heart rate after downsampling represents the stability of the elderly’s heart activity. SDNN(t i ) is the standard deviation of the RR interval (i.e., heartbeat interval) of the elderly within the time window. SDNN reflects the variability of heart rate and is used to assess the health of the heart. A larger SDNN indicates a larger heart rate variability, and a smaller SDNN indicates a more stable heart rate. SpO2 avg (t i ) is the elderly at time point t i The average blood oxygen saturation in a certain time window before, indicating the percentage of oxygenated hemoglobin in the blood to the total hemoglobin. The blood oxygen saturation sensor usually collects data every few seconds, and the average value in the time window is obtained after downsampling. Tem avg (t i ) is the time point t i The average temperature of the ambient temperature measured in a certain time window before. The temperature sensor usually has a low acquisition frequency. The average temperature of the window is obtained by interpolation or downsampling. avg (t i ) is the elderly at time point t i The average value of the air humidity measured in a certain time window before. The acquisition frequency of the humidity sensor is low. After interpolation or downsampling, the humidity mean of the window is calculated. avg (t i ) is the time point t iThe average light intensity measured in a certain time window before. The light intensity sensor usually collects data at intervals and obtains the average light intensity in the time window through interpolation or downsampling. PM avg (t i ) is the time point t i The PM2.5 sensor collects data every few seconds and calculates the average value of the time window by interpolation or downsampling. Pavg(t i ) is the time point t i The average value of the ambient air pressure measured in a certain time window before. The air pressure sensor has a low acquisition frequency, and the average air pressure of the window is calculated after interpolation or downsampling.

[0282] Optionally, the evaluation unit obtains the elderly person at t after being processed by the fusion unit. i The core data vector x(t i ), the state index S(x) of the elderly is calculated according to the following formula:

[0283]

[0284] In the formula, x(t i ) for the elderly in t i The core feature data vector at the moment, x0 is the core feature data vector of the elderly in a normal state, that is, the reference value of the normal state, D(x(t i ), x0) is t i The Euclidean distance between the feature data at the moment and the normal state feature is calculated according to the following formula:

[0285]

[0286] In the formula, x j (t i ) is the time point t i The data value of the jth feature at x 0,j is the reference value of the jth feature under normal conditions;

[0287] In this embodiment, time point t i The core feature data vector at and the core feature data vector in the normal state contain the same n eigenvalues;

[0288] In this embodiment, after the elderly wear the vest, the elderly enter the set recording mode, and in the recording mode, the core feature data vector x0 of the elderly in a normal state is recorded;

[0289] x0 is the core feature data vector of the elderly in a normal state, i.e., the reference value of the normal state, which is determined according to the following steps;

[0290] S31. The entry mode is used to establish a normal state benchmark for the elderly:

[0291] In this mode, the system continuously collects multi-dimensional status data of the elderly in their daily activities through sensor modules to ensure that the data reflects the normal behavior and physiological characteristics of the elderly in different environments and activities. This process usually requires continuous collection over a period of time to cover the daily status of enough elderly people.

[0292] The purpose of the recording mode is to record the movement behavior (such as walking speed, gait stability, movement frequency), physiological characteristics (such as heart rate, blood oxygen saturation) and environmental conditions (such as temperature, humidity, air pressure) of the elderly in their daily life;

[0293] The duration of the recording mode is that the recording process usually needs to last for several days or even weeks to ensure that the status of the elderly in different activities and environments is covered;

[0294] S32, sensor module collects status data:

[0295] The sensor module collects daily status data of the elderly through various sensors (such as gait monitoring unit, physiological monitoring unit, and environmental perception unit); these data will include:

[0296] Movement status data: walking speed, gait stability, movement frequency;

[0297] Physiological status data: average heart rate, standard deviation of RR interval, blood oxygen saturation;

[0298] Environmental data: temperature, humidity, light, PM2.5 and air pressure;

[0299] At each time point ti, a core feature data vector x(ti) is generated;

[0300] S33, fusion unit processing data:

[0301] The collected high-frequency, multi-dimensional data needs to be processed by the fusion unit. The specific steps include:

[0302] Data preprocessing: cleaning noise in the data, dealing with outliers and missing values;

[0303] Feature extraction: Extract key features such as pace, gait stability, heart rate, and blood oxygen from the preprocessed raw data;

[0304] Time alignment and synchronization: align the feature data collected by different sensors to the same time axis to ensure that they are analyzed at the same time point;

[0305] Construct the core feature data vector x(t i ): All the features extracted at each time point t i The above are combined into a core feature data vector;

[0306] Through the above processing, a core feature data vector x(t i );

[0307] S34. Calculate the core feature data vector x0 of the elderly in normal state:

[0308] In the recording mode, the system needs to determine the reference vector x0 of the normal state by analyzing and processing all the state data collected by the elderly in the normal state;

[0309] S35. Set x0 as reference:

[0310] After processing in the input mode, the system finally determines the core feature data vector x0 of the elderly in a normal state and stores it as a reference value; this vector will serve as a benchmark for subsequent monitoring of whether the elderly's behavior is abnormal; when the system is in monitoring mode, the feature data vector x(t) collected in real time will be compared with the reference vector x0 to determine whether the elderly is in a normal state;

[0311] In this embodiment, the core feature data vector of the elderly in a normal state is the state data collected by the sensor module when the elderly is in a normal state and in an input mode (or a learning state), and the state data is processed by the fusion unit to form the core feature data vector in a normal state;

[0312] In addition, in the input mode, all the status data of the elderly in their daily normal state need to be recorded, and the core feature data vector in the normal state is set as x0;

[0313] If the state index S(x) of the elderly is less than the fence warning threshold T set by the system, it indicates that the elderly's behavior is abnormal, and the behavior analysis module is triggered to analyze the elderly's behavior;

[0314] If the state index S(x) of the elderly is greater than or equal to the fence warning threshold T set by the system, it means that the elderly is close to normal and there is no abnormality;

[0315] Among them, the fence warning threshold T set by the system is set by the system or the administrator / guardian and is input from the human-computer interaction interface. This is a technical means well known to those skilled in the art. Those skilled in the art can consult relevant technical manuals to learn about this technology, so it will not be described in detail in this embodiment;

[0316] Through the cooperation of the sensor module and the evaluation module, the behavior data of the elderly can be analyzed, ensuring that the system can obtain a full range of elderly status data, and through the processing of the evaluation module, the health and behavior status of the elderly can be dynamically monitored, so that the system has a powerful data fusion capability and real-time monitoring function;

[0317] Optionally, the behavior analysis module includes a behavior feature extraction unit and a behavior analysis unit, the behavior feature extraction unit extracts at least four pre-processed behavior features of the elderly from the fusion unit in real time to form a multi-dimensional behavior feature vector, and the behavior analysis unit performs analysis based on the multi-dimensional behavior feature vector to identify abnormal behavior of the elderly;

[0318] The behavior feature extraction unit extracts at least four pre-processed behavior features of the elderly from the fusion unit in real time to form a current behavior feature vector b=[b1, b2, ..., b n ], where each b i Specific behavioral characteristics that represent older adults;

[0319] For example: b1: gait speed;

[0320] b2: movement frequency (standing up, sitting down);

[0321] b3: duration of immobility (standing still for a long time);

[0322] b4: stride change;

[0323] The above-mentioned behavior characteristics are collected in real time by the sensor module;

[0324] In addition, the habitual behavior pattern of the elderly is constructed through historical data or individualized behavior baselines; in this embodiment, it is assumed that the characteristic vector of the habitual behavior pattern is b0=[b 0，1 ,b 0，2 ,...,b 0，n ], where b i Represents a historical average or individualized benchmark value for a characteristic;

[0325] For example: 0，1 is the historical average gait speed;

[0326] b 0，2 is the historical action frequency;

[0327] b 0，3 is the length of time that is static in history;

[0328] The behavior analysis unit obtains the current behavior feature vector b and the feature vector b0 of the usual behavior pattern, and calculates the behavior abnormality index Behavioral:

[0329]

[0330] Where, T b is the abnormal behavior threshold set by the system. If D(b,b0) exceeds this abnormal behavior threshold T b , which means that the behavior may be abnormal. D(b, b0) is the Euclidean distance between the current behavior feature vector b and the usual behavior feature vector b0, and its value is calculated according to the following formula:

[0331]

[0332] Where b j is the value of the jth feature in the current behavior feature vector b, b 0,j is the value of the jth feature in the habitual behavior pattern feature vector b0, indicating the reference value of the feature in the normal state of the elderly; m is the dimension of the behavior feature vector, indicating the number of features contained in the current behavior and habitual behavior vectors; for example, if the monitored behavior features include four features: walking speed, movement frequency, gait stability, and heart rate, then m = 4;

[0333] In this implementation, through the cooperation between the evaluation module and the behavior analysis module, the system can accurately analyze the behavior of the elderly, identify abnormal situations, detect changes in the elderly's body and environment, and provide personalized abnormal evaluation based on the elderly's daily habits, thereby enhancing the intelligence and personalized monitoring capabilities of the system;

[0334] Optionally, the early warning module includes an early warning unit and an interactive unit, wherein the early warning unit triggers early warning information according to the analysis result, and the interactive unit interactively prompts the manager with the early warning information;

[0335] The early warning unit includes a comparator and an early warning trigger. The comparator obtains the analysis result and a monitoring threshold Worn set by the system, and compares the analysis result with the set monitoring threshold to form a comparison result. The early warning trigger triggers the early warning information according to the comparison result.

[0336] Specifically, the main task of the comparator is to obtain the abnormal behavior index Behavioral and the monitoring threshold Worn set by the system, and judge whether the behavior of the elderly is abnormal by comparing these two values;

[0337] Specific comparison method:

[0338] Obtaining the behavioral abnormality index Behavioral: The behavioral abnormality index calculated by the analysis module indicates the degree of deviation between the current behavior and the usual behavior pattern; it is a dynamically changing value that is updated as the elderly person's behavior changes;

[0339] Get the monitoring threshold Worn set by the system: This threshold is a critical value pre-set in the system and is used to determine whether the behavior exceeds the normal range. Usually, Worn is adjusted by the system administrator or based on historical data.

[0340] The trigger mechanism of the early warning trigger includes: obtaining a comparison result: the early warning trigger first obtains the output result of the comparator; if the comparison result shows that Behavioral≥Worn, the trigger is ready to send an early warning;

[0341] Generation of warning information: The trigger will generate corresponding warning information based on the abnormal behavior index, which may include the current status of the elderly, the nature and severity of the abnormal behavior; for example:

[0342] Warning content: including behavioral deviations of the elderly, such as "unstable gait", "abnormal heart rate", "sudden changes in behavior" and other specific prompts;

[0343] Priority classification: The warnings are classified into different priorities. According to the size of the abnormal behavior index, it is determined whether it is a minor warning, a general warning or an emergency warning;

[0344] Send warning information: The trigger sends the information to the manager's terminal (such as mobile phone, computer, special equipment) through the warning system, notifying the manager to take corresponding actions;

[0345] Optionally, the interaction unit includes a communicator and an identity verifier, wherein the communicator is used to send the warning information to the manager, and the identity verifier identifies the identity of the manager and allows the communicator to send the warning information to the manager after the identity is verified;

[0346] Optionally, the identity verifier compares the identity data collected in real time with the verification data pre-stored in the database of the server. If the identity data and the verification data are the same, the verification is passed, otherwise the verification fails;

[0347] When setting up or using the system for the first time, the administrator needs to register his / her identity; at this time, the system will collect information related to the administrator's identity and store it in the server's database; the identity data takes various forms, such as passwords, biometrics (such as fingerprints, facial recognition, iris), and multi-factor authentication information;

[0348] Storage process:

[0349] 1) The administrator provides identity information by entering a user name and password, or by scanning a fingerprint or face;

[0350] 2) If it is a password, the system will use a hash algorithm to encrypt it; if it is a biometric feature, the system will generate a feature template through a feature extraction algorithm;

[0351] 3) The processed identity verification data (such as encrypted passwords or biometric templates) is stored in the server's database; this database is specifically used to store and manage identity information;

[0352] In this embodiment, through the cooperation between the behavior analysis module and the early warning module, the manager or guardian can dynamically control the behavior of the elderly, ensuring that the entire system has the advantages of good personalized monitoring capabilities, high intelligence, strong multi-dimensional data fusion capabilities, reliable early warning analysis capabilities, high ease of use and strong interactive capabilities.

[0353] Embodiment 2: This embodiment should be understood to include all the features of any of the above embodiments, and further improve upon them. Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 , Figure 6 , Figure 7 As shown, the intelligent behavior care system also includes a wake-up module, which is arranged in the vest and performs a state result according to the physiological data collected by the sensor module, and performs a wake-up operation on the elderly according to the state result;

[0354] In this embodiment, the wake-up module is allowed to work when the elderly are in a non-sleeping period, and is not allowed to be turned on when the elderly are in a sleeping period, so as to prevent unnecessary influence on the elderly during their sleep; at the same time, the sleeping period needs to be set according to the habits of the elderly to improve the accuracy and reliability of care;

[0355] In addition, a start switch is provided on the vest. When the start switch is turned on, the awakening module is allowed to be in working state and to care for and awaken the elderly.

[0356] The awakening module includes an awakening unit and an awakening analysis unit. The awakening analysis unit analyzes the state of the elderly according to the physiological data collected by the sensor module and forms a state result according to the physiological data. The awakening unit performs an awakening operation on the elderly according to the state result.

[0357] The awakening analysis unit obtains the physiological data collected by the sensor module, and calculates the state index State of the elderly according to the following formula:

[0358]

[0359] Wherein, HR(t) is the heart rate at the current moment, and its value is obtained by the heart rate sensor; SpO2(t) is the blood oxygen saturation at the current moment, and its value is obtained by the blood oxygen saturation sensor; RR(t) is the respiratory rate at the current moment; ΔHR is the change between the current heart rate and the last heart rate measurement, that is, ΔHR=HR(t)-HR(t-1); ΔSpO2 is the change between the current blood oxygen saturation and the last blood oxygen saturation measurement, that is: ΔSpO2=pO2(t)-SpO2(t-1); ΔRR is the change between the current respiratory rate and the last respiratory rate measurement, that is: ΔRR=RR(t)-RR(t-1);

[0360] In this embodiment, the respiratory frequency is detected by measuring the expansion and contraction of the chest cavity through a chest strap sensor, which is based on the inductance principle. When the chest cavity expands, the area of ​​the inductance coil changes, generating a current change, and then calculating the breathing cycle and frequency;

[0361] Among them, the chest strap sensor integrates a flexible chest strap into the vest to achieve the goal of not affecting the activities of the elderly;

[0362] In addition, when the state index S of the elderly falls into the slight fluctuation monitoring threshold range [range1, range2] set by the system, it means that the physiological parameters of the elderly have slightly fluctuated, but not seriously abnormal; at this time, a slight reminder (such as vibration wake-up) operation is selected to remind the elderly to move or check their condition;

[0363] When the state index State of the elderly is lower than the lower limit range 1 of the slight fluctuation monitoring threshold range set by the system, it means that the physiological state of the elderly does not fluctuate much and is within the normal range;

[0364] When the state index S of the elderly exceeds the upper limit range2 of the monitoring threshold, it means that the physiological state of the elderly has fluctuated greatly and there may be potential health risks; at this time, the system should adopt a stronger awakening method, such as increasing the vibration intensity or combining sound awakening;

[0365] If the state index S exceeds or is equal to twice the upper limit of the monitoring threshold range2 set by the system, it means that the physiological state of the elderly is extremely abnormal, indicating that there is a danger (such as a sharp drop in heart rate or apnea, etc.), and the system should trigger an emergency wake-up operation;

[0366] In this embodiment, the slight fluctuation monitoring threshold range [range1, range2] set by the system is set by the manager or guardian according to the actual situation of the elderly and input from the human-computer interaction interface. This is a technical means well known to those skilled in the art, so it will not be described in detail in this embodiment;

[0367] The awakening unit includes a placement cavity, a vibration motor 10, an electrode patch 11, an audio device, an LED flashlight, and a battery. The placement cavity is arranged on the vest 1 for placing the vibration motor 10, the electrode patch 11, the audio device, and the LED flashlight. The vibration motor 10 is used to physically stimulate the elderly and generate a slight vibration to remind or wake up the elderly. The audio device wakes up by emitting a sound. The audio stimulation is an alarm, a voice prompt, or soft music. The electrode patch 11 is used to apply a slight electrical stimulation to wake up the elderly. The LED flashlight is used to flash to remind the elderly, and is particularly suitable for elderly people who are sensitive to vision. The battery is electrically connected to the vibration motor 10, the electrode patch 11, the audio device, and the LED flashlight, and supplies power to the vibration motor 10, the electrode patch 11, the audio device, and the LED flashlight.

[0368] When the system detects that the elderly person is in a long-term static or slightly abnormal state (when the state index S falls into the slight fluctuation monitoring threshold range [range1, range2] set by the system), the wake-up unit triggers the vibration motor 10 in the vest 1;

[0369] If the status index reaches a moderately abnormal or more serious state (status index S ≥ range2), the wake-up unit will send out a sound through the audio module to remind the elderly. At the same time, the LED light will remind the elderly by flashing quickly for a short time. If the vibration or sound cannot effectively wake up the elderly, the flash will be used as an auxiliary or enhanced reminder.

[0370] In an emergency state (the state index S exceeds or is equal to twice the upper limit of the monitoring threshold range2 set by the system), if the elderly person does not respond to the mild vibration or sound reminder, the awakening unit activates the electrode patch 11 to perform mild electrical stimulation;

[0371] In this embodiment, the awakening module is used to provide wake-up reminders to the elderly, which effectively improves the intelligent behavior care capabilities of the elderly, ensuring that the entire system has the advantages of good protection effect, high intelligence and good care effect.

[0372] The contents disclosed above are only preferred feasible embodiments of the present invention, and do not limit the protection scope of the present invention. Therefore, all equivalent technical changes made using the contents of the present invention specification and drawings are included in the protection scope of the present invention. In addition, the elements therein can be updated as technology develops.

Claims

1. An intelligent behavior care system for monitoring elderly behavior, the intelligent behavior care system comprising a server, characterized in that: The intelligent behavior care system also includes a sensor module, an evaluation module, a behavior analysis module and an early warning module, and the server is connected to the sensor module, the evaluation module, the behavior analysis module and the early warning module respectively; The sensor module is used to collect status data of the elderly, the evaluation module evaluates the status of the elderly according to the status data collected by the sensor module to form an evaluation result, the behavior analysis module triggers analysis of the elderly's behavior according to the evaluation result to form a behavior analysis result, and the early warning module issues an early warning prompt to the guardian according to the behavior analysis result; The evaluation module includes a data fusion unit and an evaluation unit, wherein the data fusion unit performs fusion processing on the status data of the elderly, and the evaluation unit evaluates the status of the elderly according to the fused status data to form an evaluation result; The sensor module is worn on the elderly and moves along with the movement of the elderly.

2. The intelligent behavior care system for elderly behavior monitoring according to claim 1 is characterized in that: The sensor module includes a vest, a gait monitoring unit, a motion recognition unit, a physiological monitoring unit, an environmental perception unit and a data storage device. The gait monitoring unit, the motion recognition unit, the physiological monitoring unit and the environmental perception unit are arranged on the vest. The gait monitoring unit collects the body posture and gait information of the elderly, the motion recognition unit collects the stability data of the elderly, the physiological monitoring unit collects the physiological state data of the elderly, the environmental perception unit collects the environmental data of the environment in which the elderly are located, and the data storage device stores the collected body posture and gait information, stability data, physiological state data and environmental data; The state data includes the body posture and gait information, stability data, physiological state data, and environmental data.

3. The intelligent behavior care system for elderly behavior monitoring according to claim 2 is characterized in that: The fusion unit includes a data preprocessing component, a feature extraction component, and a time synchronization component. The data preprocessing component is used to preprocess the state data. The feature extraction component is used to extract gait, movement, physiology and environmental key features from the state data to form core feature data. The time synchronization component is used to perform time alignment and synchronization processing on the extracted core feature data to form the elderly in t i The core feature data vector x(t i ).

4. The intelligent behavior care system for elderly behavior monitoring according to claim 3 is characterized in that: The evaluation unit obtains the data processed by the fusion unit to form the elderly at t i The core data vector x(t i ), the state index S(x) of the elderly is calculated according to the following formula: Where x0 is the core feature data vector of the elderly in normal state, that is, the reference value of normal state, D(x(t i ), x0) is t i The Euclidean distance between the feature data at the moment and the normal state feature; If the state index S(x) of the elderly is less than the fence warning threshold T set by the system, it means that the elderly's behavior is abnormal, and the behavior analysis module is triggered to analyze the elderly's behavior.

5. The intelligent behavior care system for elderly behavior monitoring according to claim 4 is characterized in that: The behavior analysis module includes a behavior feature extraction unit and a behavior analysis unit. The behavior feature extraction unit extracts at least four pre-processed behavior features of the elderly from the fusion unit in real time to form a multi-dimensional behavior feature vector. The behavior analysis unit performs analysis based on the multi-dimensional behavior feature vector to identify abnormalities in the elderly's behavior.

6. The intelligent behavior care system for elderly behavior monitoring according to claim 5 is characterized in that: The early warning module includes an early warning unit and an interactive unit. The early warning unit triggers early warning information according to the analysis result, and the interactive unit interactively prompts the manager with the early warning information. Among them, the warning unit includes a comparator and a warning trigger. The comparator obtains the analysis result and the monitoring threshold Worn set by the system, and compares the analysis result with the set monitoring threshold to form a comparison result. The warning trigger triggers the warning information according to the comparison result.

7. The intelligent behavior care system for elderly behavior monitoring according to claim 6 is characterized in that: The interaction unit includes a communicator and an identity verifier. The communicator is used to send the warning information to the manager. The identity verifier identifies the identity of the manager and allows the communicator to send the warning information to the manager after the verification is passed.

8. The intelligent behavior care system for elderly behavior monitoring according to claim 7 is characterized in that: The gait monitoring unit is arranged in the lower back or waist area of ​​the vest and collects the body posture and gait data of the elderly; Wherein, the gait monitoring unit includes an accelerometer and a gyroscope, the accelerometer collects acceleration data of the elderly, and the gyroscope collects body posture change data of the elderly.

9. The intelligent behavior care system for elderly behavior monitoring according to claim 8 is characterized in that: The motion recognition unit is disposed on a shoulder of the vest and captures body stability data of the elderly.

10. The intelligent behavior care system for elderly behavior monitoring according to claim 9, characterized in that: The identity verifier compares the identity data collected in real time with the verification data pre-stored in the database of the server. If the identity data and the verification data are the same, the verification is passed, otherwise the verification fails.

Images