Old people body health early diagnosis and risk assessment system based on artificial intelligence

Abstract

The invention provides an old people body health early diagnosis and risk assessment system based on artificial intelligence, and relates to the technical field of old people health monitoring and assessment. Old people health parameters are collected in real time according to a set time interval, and a statistical method is used to calculate a mean value and a standard deviation of cardiovascular parameters; then, based on acceleration data, heart rate change, blood pressure fluctuation and position change information, sleep, movement and resting states of the old people are judged, and the day and night blood pressure change rate, the sleep state heart rate standard deviation, the movement state pulse wave conduction speed change quantity and the resting state heart stroke output average value are calculated; four risk assessment formulas of blood pressure risk, heart rate risk, pulse wave conduction velocity risk and heart stroke output quantity are constructed according to the data, a comprehensive health risk value is calculated through weighted merging, and the health state of the old people is judged to be healthy, medium-low risk or high risk according to a set threshold value.

Description

Technical Field

[0001] The present invention relates to the technical field of elderly health monitoring and assessment, and particularly to a system for early diagnosis and risk assessment of the physical health of the elderly based on artificial intelligence. Background Art

[0002] With the intensification of population aging, the health management of the elderly has become increasingly important. At present, in the field of elderly physical health monitoring and risk assessment, there are many limitations in the existing technologies.

[0003] In terms of data collection, traditional methods mostly rely on single-index monitoring. For example, only blood pressure or heart rate data is concerned. This single-index collection method cannot comprehensively reflect the complex physical state of the elderly. The decline of the physical functions of the elderly is often reflected in multiple physiological systems. Single indexes are difficult to cover the changes in multiple aspects such as the cardiovascular system, nervous system, and motor system, resulting in one-sided evaluation results and prone to missing other potential health risks.

[0004] From the perspective of risk assessment, most of the existing technologies lack real-time and comprehensiveness. Most assessment methods are based on regular physical examination data and cannot dynamically track the physiological indexes of the elderly in their daily lives. This makes it difficult to timely capture some short-term but abnormal physiological fluctuations, missing the best opportunity for early detection of diseases. And because a comprehensive and accurate risk assessment model cannot be constructed, it is difficult to detect abnormal changes in the body at the early stage of the disease, and thus early diagnosis and prevention cannot be effectively achieved.

[0005] In summary, the existing elderly health monitoring and risk assessment technologies have obvious deficiencies in the comprehensiveness of data collection and the real-time nature of risk assessment, and there is an urgent need for an innovative method to improve this situation to better protect the health of the elderly.

[0006] The above information disclosed in the background art section is only used to strengthen the understanding of the background of the present disclosure, and therefore it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0007] The purpose of the present invention is to provide a system for early diagnosis and risk assessment of the physical health of the elderly based on artificial intelligence to solve the problems raised in the above background art.

[0008] To achieve the above purpose, the present invention provides the following technical solutions:

[0009] A system for early diagnosis and risk assessment of the physical health of the elderly based on artificial intelligence, comprising:

[0010] A data acquisition module, which is used to collect the cardiovascular parameters of the elderly at the same time interval, remove outliers and smooth the cardiovascular parameters through preprocessing. The cardiovascular parameters include blood pressure, heart rate, pulse wave velocity, and cardiac stroke volume data;

[0011] A state judgment module, which is used to collect the state classification parameters of the elderly at the same time interval, and then judge the sleep, exercise and resting states of the elderly according to the state classification parameters, heart rate change rate, and blood pressure. The state classification parameters include acceleration and position data;

[0012] A health risk construction module, which is used to construct a risk formula according to the cardiovascular parameters and state classification parameters in different states. The risk formula includes the diurnal blood pressure change rate, the standard deviation of heart rate in the sleep state, the change amount of pulse wave velocity in the exercise state, and the average value of cardiac stroke volume in the resting state;

[0013] A comprehensive risk assessment and judgment module, which is used to construct a comprehensive risk assessment formula by using weighted summation according to the risk formula, and evaluate the overall health status and risk of the elderly through the result of the comprehensive risk assessment formula.

[0014] Further, the set time interval is t jg , and collect the blood pressure, heart rate, pulse wave velocity, and cardiac stroke volume data of the elderly at the same time interval t jg ;

[0015] Record the blood pressure as [SBP i , DBP i , where SBP represents the systolic blood pressure in the blood pressure data, DBP represents the diastolic blood pressure in the blood pressure data, record the heart rate as HR i , record the pulse wave velocity as PWV i , and the cardiac stroke volume as SV i , where i represents the acquisition serial number, and the records of each day are renumbered starting from 6:00 in the morning. i = 1, 2, 3,..., N, and N is a positive integer ≥ 0, representing the total number of acquisitions from 6:00 in the morning of each day to 6:00 in the morning of the next day, and a time stamp is attached.

[0016] Further, select the historical data of the most recent complete 3 days as the data source for calculating the mean and standard deviation, that is, from 6:00 three days ago to 3 days after this starting time point. The total duration of the data source is:

[0017] T z = 4320

[0018] In the formula, T zDenoted as the total duration of the data source;

[0019] Then calculate the number of data collected within the most recent complete 3 days:

[0020]

[0021] In the formula, n g Denoted as the number of data collected within the most recent complete 3 days;

[0022] Calculate the mean value of the cardiovascular parameters:

[0023]

[0024] In the formula, μ is denoted as the mean value of the cardiovascular parameters, and x i Denoted as the value of the cardiovascular physiological parameter at the i-th data point;

[0025] Calculate the standard deviation of the cardiovascular parameters:

[0026]

[0027] In the formula, σ is denoted as the standard deviation of the cardiovascular parameters;

[0028] For each cardiovascular parameter x i , when |x i - μ| > 3σ, then it is determined that the cardiovascular parameter x i is an outlier and is replaced with the mean value μ of the cardiovascular parameters;

[0029] Adopt the moving average filtering method to smooth each cardiovascular parameter. First, set the moving average window size:

[0030]

[0031] In the formula, ω is denoted as the sliding window size and is a positive integer;

[0032] Calculate the moving average value:

[0033]

[0034] In the formula, Denoted as the cardiovascular parameter corresponding to the i-th data point after smoothing, and x j Denoted as the cardiovascular parameter corresponding to the j-th data point directly collected. For the data that cannot satisfy the complete window calculation, the original value is retained. After that, the smoothed data replaces the original data x i .

[0035] Furthermore, collect the acceleration of the elderly through the accelerometer and calculate the amplitude of the acceleration based on the acceleration:

[0036]

[0037] Wherein, A i represents the total amount of acceleration collected for the i-th time, i.e., the acceleration amplitude, a x , a y , a z represent the components of the acceleration on the three coordinate axes;

[0038] Statistically analyze the acceleration amplitude A jg within each time interval t i to calculate the number of times n l that the change exceeds 0.05, and calculate the acceleration change frequency:

[0039]

[0040] Wherein, f represents the acceleration change frequency, and n l represents the number of times that the acceleration amplitude A i changes by more than 0.05;

[0041] Calculate the change rate of the heart rate:

[0042]

[0043] Wherein, ΔHR i represents the average change of the heart rate within each t jg time period, i.e., the heart rate change rate, and HR i+1 represents the heart rate value collected for the (i + 1)-th time, and HR i represents the heart rate value collected for the i-th time;

[0044] According to the position data of the state division parameter, record the position coordinates (x1, y1) and (x2, y2) before and after each time interval t jg and calculate the displacement distance:

[0045]

[0046] Wherein, d represents the displacement distance within the time interval T q ;

[0047] Calculate the moving speed:

[0048]

[0049] Wherein, v i represents the moving speed within the i-th time interval;

[0050] Calculate the fluctuation amount of the blood pressure:

[0051]

[0052] Wherein, ΔSBP i represents the average change in systolic blood pressure within each t jg time period, that is, the systolic blood pressure fluctuation amount of blood pressure, SBP i+1 represents the systolic blood pressure collected at the (i + 1)-th time, SBP i represents the systolic blood pressure collected at the i-th time; ΔDBP i represents the average change in diastolic blood pressure within each t jg time period, that is, the diastolic blood pressure fluctuation amount of blood pressure, DBP i+1 represents the diastolic blood pressure collected at the (i + 1)-th time, DBP i represents the diastolic blood pressure collected at the i-th time;

[0053] The sleep state judgment conditions are as follows:

[0054] For a certain time interval t jg , it is necessary to simultaneously satisfy that the acceleration amplitude A i <0.1, the acceleration change frequency f < 0.5 and the position movement speed v i <0.1. On this basis, it is also necessary to satisfy that the heart rate change rate ΔHR i <5, or simultaneously satisfy that the systolic blood pressure fluctuation amount of blood pressure ΔSBP i <10 and the diastolic blood pressure fluctuation amount of blood pressure ΔDBP i <5;

[0055] The exercise state judgment conditions are as follows:

[0056] Record the age K of the elderly. For a certain time interval t jg , when the acceleration amplitude satisfies 0.3 < A i <1.5, the acceleration change frequency 2 < f < 10, the heart rate change rate ΔHR i > 10, the position movement speed 5 > v i > 0.5. When it is judged as the exercise state, when the first heart rate change rate ΔHR i ≤10, when the heart rate HR i > (220 - K) × 0.6, it is still judged as the exercise state. When the heart rate HR i < (220 - K) × 0.6, it is judged that the exercise state ends;

[0057] The resting state judgment conditions are as follows:

[0058] For a certain time interval t jg , when the acceleration amplitude satisfies 0.1 < A i< 0.3, acceleration change frequency 0.5 < f < 2, heart rate change rate 10 > ΔHR i > 5, position movement speed 0.5 ≥ v i ≥ 0.1;

[0059] If within a certain time interval t jg the data does not meet the determination conditions for sleep, exercise, and resting states, then the state corresponding to this time interval is determined as other state.

[0060] Further, first divide the data of 24 hours a day into day and night: Daytime period: 6:00 - 22:00, Nighttime period: 22:00 - 6:00 of the next day;

[0061] Record the blood pressure during the daytime period as SBP day，i , DBP day，i ; Record the blood pressure during the nighttime period as SBP night，i , DBP night，i ;

[0062] Calculate the mean systolic blood pressure of the blood pressure during the daytime period:

[0063]

[0064] In the formula, μ SBP，day represents the mean systolic blood pressure of the blood pressure during the daytime period, n day represents the number of acquisitions during the daytime period;

[0065] Calculate the mean diastolic blood pressure of the blood pressure during the daytime period:

[0066]

[0067] In the formula, μ DBP，day represents the mean diastolic blood pressure of the blood pressure during the daytime period;

[0068] Calculate the mean systolic blood pressure of the blood pressure during the nighttime period:

[0069]

[0070] In the formula, μ SBP，day represents the mean systolic blood pressure of the blood pressure during the nighttime period, n nigh represents the number of acquisitions during the nighttime period;

[0071] Calculate the mean diastolic blood pressure of the blood pressure during the nighttime period:

[0072]

[0073] In the formula, μ DBP，night represents the mean diastolic blood pressure of the blood pressure during the nighttime period;

[0074] Calculate the circadian blood pressure change rate:

[0075]

[0076] In the formula, NDBPR SBP represents the systolic blood pressure change rate in the circadian blood pressure change rate, and NDBPR DBP represents the diastolic blood pressure change rate in the circadian blood pressure change rate;

[0077] Extract the heart rate data in the sleep state and record it as HR sleep，ai' , ai' = 1, 2, 3,..., n sleep , where n sleep represents the number of heart rate records in the sleep state, and calculate the average heart rate in the sleep state:

[0078]

[0079] In the formula, μ HR，sleep represents the average heart rate in the sleep state;

[0080] Calculate the standard deviation of the heart rate in the sleep state:

[0081]

[0082] In the formula, σ HR，sleep represents the standard deviation of the heart rate in the sleep state;

[0083] Extract the pulse wave velocity data in the exercise state, that is, PWV data, and record it as PWV exercise，bi' , bi' = 1, 2, 3,..., n exercise , where n exercise represents the number of PWV records in the exercise state;

[0084] Calculate the change in pulse wave velocity in the exercise state:

[0085] ΔPWV exercise = max(PWW exercise , bi') - min(PWV exercise , bi')

[0086] In the formula, ΔPWV exercise represents the change in pulse wave velocity in the exercise state, and max(PWV exercise , bi') represents all the PWV exercise，bi' recorded in the exercise state, and min(PWV exercise , bi') represents all the PWV exercise，bi' recorded in the exercise state;

[0087] Extract the stroke volume data of the heart in the resting state, that is, the SV data is recorded as SV rest，ci' , ci' = 1, 2, 3, …, n rest , where, n rest represents the number of SV times recorded in the resting state;

[0088] Calculate the average stroke volume of the heart in the resting state:

[0089]

[0090] In the formula, μ SV，rest represents the average stroke volume of the heart in the resting state.

[0091] Furthermore, set the range of the diurnal variation rate of normal systolic blood pressure in the elderly to be 10 - 20%, the range of the diurnal variation rate of normal diastolic blood pressure to be 8 - 15%, the range of the standard deviation of heart rate in the normal sleep state to be 30 - 60 ms; the range of the change in pulse wave velocity in the normal exercise state to be 1 - 2 m / s, and the range of the average stroke volume of the heart in the normal resting state to be 50 - 80 ml;

[0092] Construct the risk formula for the diurnal blood pressure variation rate:

[0093]

[0094] In the formula, R bloodpressure represents the blood pressure risk, α, β represent the weight coefficients, α ≥ 0, β ≥ 0;

[0095] When NDBPR SBP ∈[10 - 20], the weight coefficient α = 0 in the blood pressure risk formula;

[0096] When NDBPR DBP ∈[8 - 15], the weight coefficient β = 0 in the blood pressure risk formula;

[0097] When NDBPR SBP ∈[10 - 20], NDBPR DBP ∈[8 - 15], the weight coefficient α > β > 0 in the blood pressure risk formula;

[0098] Construct the risk formula for the standard deviation of heart rate:

[0099]

[0100] In the formula, R hartrate represents the risk of the standard deviation of heart rate in the sleep state, γ represents the weight coefficient;

[0101] When γ > 0, when σHR，sleep When it ∈ [30 - 60], γ = 0;

[0102] Construct a risk formula for the change in pulse wave velocity:

[0103]

[0104] In the formula, R PWV represents the risk of the change in pulse wave velocity, and δ represents the weight coefficient;

[0105] When , δ > 0. When ΔPWV exercise ∈ [1 - 2], δ = 0;

[0106] Construct a risk formula for the average value of cardiac stroke volume:

[0107]

[0108] In the formula, R SV represents the risk of the average value of cardiac stroke volume, and ε represents the weight coefficient;

[0109] When , ε > 0. When μ SV，rest ∈ [50 - 80], ε = 0.

[0110] Furthermore, construct a comprehensive risk assessment formula:

[0111] R total = ω1 × R bloodpressure + ω2 × R hartrate + ω3 × R PWV + ω4 × R SV

[0112] In the formula, R total represents the comprehensive risk, ω1, ω2, ω3, ω4 represent the weight coefficients, and ω1 + ω2 + ω3 + ω4 = 1, ω1 × ω2 × ω3 × ω4 > 0;

[0113] When R total ≤ 1.2, it is determined to be in a healthy state. When 1.2 < R total ≤ 1.5, it is determined that the physical health state is in a medium - low risk. When 1.5 < R total , it is determined that the physical health state is in a high - risk state.

[0114] Compared with the prior art, the beneficial effects of the present invention are:

[0115] The present invention continuously collects data at fixed time intervals to collect multi-dimensional data of the blood pressure, heart rate, pulse wave velocity, stroke volume, acceleration, and position of the elderly, more comprehensively reflecting the physical state of the elderly and avoiding one-sidedness caused by single-index evaluation;

[0116] The present invention also calculates the diurnal blood pressure change rate, the standard deviation of heart rate during sleep, the change in pulse wave velocity during exercise, and the average value of stroke volume at rest in real time by collecting data, and constructs a comprehensive risk assessment formula, which can timely evaluate the degree of health risk, help detect abnormal changes in the body at the early stage of the disease, and achieve early diagnosis and prevention. BRIEF DESCRIPTION OF THE DRAWINGS

[0117] Figure 1 It is a schematic diagram of the overall module process of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0118] In order to make the purpose, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to specific embodiments.

[0119] It should be noted that unless otherwise defined, the technical terms or scientific terms used in the present invention should be the general meanings understood by those with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not represent any order, quantity or importance, but are only used to distinguish different components. Words such as "including" or "comprising" mean that the elements or objects appearing before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Words such as "connected" or "coupled" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right", etc. are only used to represent relative positional relationships, and when the absolute position of the object being described changes, the relative positional relationship may also change accordingly.

[0120] Embodiment:

[0121] Please refer to Figure 1 , the present invention provides a technical solution:

[0122] An early diagnosis and risk assessment system for the physical health of the elderly based on artificial intelligence, comprising:

[0123] A data acquisition module, which is used to acquire the cardiovascular parameters of the elderly at the same time interval, remove outliers and smooth the cardiovascular parameters through preprocessing, and the cardiovascular parameters include blood pressure, heart rate, pulse wave velocity and stroke volume data;

[0124] Cardiovascular diseases are common among the elderly. Persistent high blood pressure can damage important organs such as the heart, brain, and kidneys, leading to complications such as coronary heart disease, cerebral hemorrhage, and renal failure. Therefore, by paying attention to and monitoring cardiovascular health, not only can diseases such as hypertension, coronary heart disease, and myocardial infarction be effectively managed, but also many diseases that affect the health of the elderly can be prevented and reduced.

[0125] Using a wearable device, such as the Huawei HUAWEI WATCH B9-D10, set the time interval as t jg at the same time interval t jg The time interval t jg is defaulted to 1 minute and can be adjusted according to actual needs. Collect the blood pressure, heart rate, and pulse wave velocity of the elderly, and cooperate with the wearable cardiac ultrasound imaging patch to collect the data of stroke volume of the heart, and attach a time stamp. It can accurately match different types of data in the time dimension. For example, through the time stamp, the specific heart rate, blood pressure and other cardiovascular parameters at a certain moment during sleep can be determined, so as to more accurately analyze the relationship between the sleep state and cardiovascular parameters;

[0126] Systolic blood pressure mainly reflects the pressure in the artery during heart contraction, while diastolic blood pressure reflects the pressure generated by the elastic recoil of the arterial blood vessel during heart relaxation. It is an important basis for diagnosing hypertension and judging the severity of hypertension. Through these two indicators, the pressure load borne by the cardiovascular system can be understood, and the health status of the blood vessel wall can be evaluated; Heart rate refers to the number of times the heart beats per minute. Too fast or too slow may indicate problems such as cardiac arrhythmia, hyperthyroidism, and sinoatrial node dysfunction. Monitoring the change of heart rate is of great significance for evaluating the physical health status; Pulse wave velocity (PWV) is an important indicator reflecting arterial stiffness. As people age, the arterial wall will gradually become hardened. The increase in PWV is closely related to the risk of cardiovascular diseases such as coronary heart disease and stroke. Therefore, measuring PWV can detect arterial lesions at an early stage; Stroke volume (SV) of the heart refers to the amount of blood ejected by the heart each time it contracts. By monitoring SV, the pumping ability of the heart can be understood. For example, in patients with heart failure, due to the weakened myocardial contractility, the stroke volume usually decreases. Therefore, collecting data such as blood pressure, heart rate, pulse wave velocity, and stroke volume of the heart can comprehensively monitor the cardiovascular system and overall health status of the elderly from multiple dimensions.

[0127] Take 6:00 in the morning as the starting time point for daily monitoring every day. This is because at this time, the elderly usually have completed a night's sleep and their physical state is relatively stable, and relatively basic and representative physiological data can be obtained. Record the collected blood pressure as [SBP i , DBP i, where SBP represents the systolic blood pressure in the blood pressure data, DBP represents the diastolic blood pressure in the blood pressure data, and the recorded heart rate is HR i , and the recorded pulse wave velocity is PWV i , and the stroke volume of the heart is SV i , where i represents the acquisition serial number, and the records for each day are renumbered starting from 6:00 in the morning. i = 1, 2, 3,..., N, that is, the data collected for the first time is marked as i = 1, where N is a positive integer ≥ 0, representing the total number of acquisitions from 6:00 in the morning of each day to 6:00 in the morning of the next day.

[0128] Taking the current time as the reference, tracing back to 6:00 in the morning three days ago, and selecting the historical data of the most recent complete three days as the data source for calculating the mean and standard deviation, that is, 6:00 three days ago, lasting until three days after this starting time point, the calculation method of the total duration of the data source is as follows:

[0129] There are 24 hours in a day and 60 minutes in each hour, so the number of minutes in a day is 24 × 60 = 1440. Therefore, the total duration of three days is:

[0130] T z = 3 × 1440 = 4320

[0131] In the formula, T z represents the total duration of the data source. By calculating the mean and standard deviation of the data collected within 4320 minutes, it is possible to accurately determine whether the subsequently collected data is within the normal fluctuation range. When the data deviates from the mean by more than a certain multiple of the standard deviation, it can be determined as an outlier, and corresponding processing can be carried out accordingly;

[0132] Then calculate the number of data collected within the most recent complete three days:

[0133]

[0134] In the formula, n g represents the number of data collected within the most recent complete three days. For example, when the time interval t jg takes a value of 1 minute, then the number of data collected within three days n g is 4320;

[0135] Calculate the mean of the cardiovascular parameters:

[0136]

[0137] In the formula, μ represents the mean of the cardiovascular parameters, and x i represents the value of the cardiovascular physiological parameter at the i-th data point;

[0138] Calculate the standard deviation of the cardiovascular parameters:

[0139]

[0140] Wherein, σ represents the standard deviation of cardiovascular parameters;

[0141] Due to errors in the measurement device, changes in the temporary physiological state of the measured person, the influence of environmental factors, etc., the collected data may contain outliers. Outliers can cause the results to deviate from the true average level and normal fluctuation range, thereby affecting the analysis of the overall data and the judgment of the health status of the elderly. Therefore, for each cardiovascular parameter x i , when |x i -μ|>3σ, it is determined that the cardiovascular parameter x i is an outlier and is replaced with the mean μ of the cardiovascular parameters;

[0142] The collected cardiovascular parameter data often has certain noise and fluctuations, which may mask the true trend of the data and is not conducive to the accurate analysis of the health status of the elderly. The moving average filtering method can effectively remove this noise and short-term fluctuations, make the data smoother, and highlight the long-term trend and true change law of the data. Therefore, the moving average filtering method is used to smooth each cardiovascular parameter. First, set the moving average window size:

[0143]

[0144] Wherein, ω represents the sliding window size and is a positive integer;

[0145] Calculate the moving average value:

[0146]

[0147] Wherein, represents the cardiovascular parameter corresponding to the i-th data point after smoothing, and x j represents the cardiovascular parameter corresponding to the j-th data point directly collected. For data that cannot satisfy the complete window calculation, the original value is retained. After that, the smoothed data replaces the original data x i .

[0148] A state judgment module, which is used to collect the state classification parameters of the elderly at the same time interval, and then judge the sleep, exercise and resting states of the elderly according to the state classification parameters, heart rate change rate, and blood pressure. The state classification parameters include acceleration and position data;

[0149] Using an acceleration sensor and a position positioning device at the same time interval t jgCollect the acceleration in the x, y, and z directions and the position information with an accuracy of up to several meters. By collecting the acceleration and position information in real time, it is possible to determine whether the elderly are in a moving, resting, or sleeping state, which is an important factor in the assessment of health status, especially for evaluating cardiovascular health and sleep quality during exercise.

[0150] Collect the acceleration of the elderly through an accelerometer and calculate the amplitude of the acceleration based on the acceleration:

[0151]

[0152] In the formula, A i represents the total amount of acceleration collected in the i-th collection, that is, the acceleration amplitude, and a x , a y , a z represent the components of the acceleration on the three coordinate axes. By the change of the acceleration amplitude, different activity states (exercise, rest, sleep) can be effectively distinguished. For example, in the exercise state, the acceleration amplitude will increase with the increase of exercise intensity; in the resting state, the amplitude is close to zero;

[0153] Count the number n jg of times when the acceleration amplitude A i changes by more than 0.05 within each time interval t l , and calculate the acceleration change frequency:

[0154]

[0155] In the formula, f represents the acceleration change frequency, and n l represents the number of times when the acceleration amplitude A i changes by more than 0.05. The acceleration change frequency can intuitively reflect the frequency of activity changes of the elderly within a certain period of time, so as to analyze the activity state more comprehensively. According to a large number of studies and practical experiences on the daily exercise of the elderly, an acceleration amplitude change of more than 0.05 can better reflect some exercise changes with obvious physiological significance.

[0156] During different activities, the metabolic demands of the body will change. For example, when changing from a quiet state to an exercise state, the muscles need more oxygen and energy supply. To meet this demand, the heart will increase the heart rate and cardiac output. Therefore, the heart rate will increase with the increase of activity intensity. The change of the heart rate can reflect the change of the body activity state;

[0157] Calculate the change rate of the heart rate:

[0158]

[0159] In the formula, ΔHRi is expressed as the average change in heart rate within each t jg time period, that is, the heart rate change rate, HR i+1 is expressed as the heart rate value collected at the (i + 1)-th time, HR i is expressed as the heart rate value collected at the i-th time;

[0160] By calculating the moving speed, it is possible to intuitively judge whether the elderly are in different activity states such as stationary, slow walking, fast walking, or running. Therefore, first calculate the displacement distance based on the position data, and then calculate the moving speed based on the time interval and the displacement distance to provide basic data for judging daily activities;

[0161] According to the position data of the state division parameters, record the position coordinates (x1, y1) and (x2, y2) before and after each time interval t jg and calculate the displacement distance:

[0162]

[0163] In the formula, d represents the displacement distance within the time interval T q ;

[0164] Calculate the moving speed:

[0165]

[0166] In the formula, v i represents the moving speed within the i-th time interval.

[0167] When the human body is performing different activities, there will be corresponding physiological changes in blood pressure. When changing from a quiet state to an active state such as exercise, the peripheral blood vessels will also perform corresponding contraction or relaxation adjustments according to the needs of the activity to ensure the reasonable distribution of blood. Generally speaking, blood pressure may increase slightly during mild activities, and the increase in blood pressure will be more obvious during medium- to high-intensity activities;

[0168] Calculate the fluctuation amount of blood pressure:

[0169]

[0170] In the formula, ΔSBP i represents the average change in the systolic blood pressure of blood pressure within each t jg time period, that is, the systolic blood pressure fluctuation amount of blood pressure, SBP i+1 is expressed as the systolic blood pressure collected at the (i + 1)-th time, SBP i is expressed as the systolic blood pressure collected at the i-th time; ΔDBP i represents the average change in the diastolic blood pressure of blood pressure within each t jg time period, that is, the diastolic blood pressure fluctuation amount of blood pressure, DBP i+1Denoted as the diastolic blood pressure of the (i + 1)-th collection, DBP i Denoted as the diastolic blood pressure of the i-th collection

[0171] The sleep state judgment conditions are as follows:

[0172] Since the human body is basically in a static state, the acceleration of the body is very small, and the change in acceleration is also very small. Therefore, for a certain time interval t jg , it is necessary to simultaneously satisfy the acceleration amplitude A i <0.1, the acceleration change frequency f < 0.5, and the position movement speed v i <0.1, reflecting the relative static state of the body during sleep; normally, the heart rate change during sleep is usually relatively small, and the blood pressure of the human body will show a certain degree of decline. Moreover, there are differences in the physiological manifestations of different individuals during sleep, and when the elderly have certain cardiovascular diseases, the cardiovascular parameters may be abnormal. Therefore, on this basis, it is also necessary to satisfy the heart rate change rate ΔHR i <5, or simultaneously satisfy the systolic blood pressure fluctuation amount ΔSBP of the blood pressure i <10 and the diastolic blood pressure fluctuation amount ΔDBP of the blood pressure i <5;

[0173] The exercise state judgment conditions are as follows:

[0174] Record the age K of the elderly. When the human body is exercising, there will be obvious displacement and movement changes in the body, such as walking, running, jumping, etc. The heart needs to increase blood pumping to meet the body's needs, so the heart rate change rate is relatively large, and the acceleration will change frequently. Therefore, for a certain time interval t jg , when the acceleration amplitude satisfies 0.3 < A i <1.5, the acceleration change frequency 2 < f < 10, the heart rate change rate ΔHR i >10, and the position movement speed 5 > v i >0.5, it is judged as the exercise state;

[0175] In the initial stage of exercise, the heart rate rises rapidly, and the heart rate change rate is obvious. As the exercise time prolongs, the body gradually adapts to the exercise intensity, and the heart rate will enter a relatively stable state. The exercise end state needs to be further judged in combination with the heart rate; when it is judged as the exercise state, when the first heart rate change rate ΔHR i ≤10, when the heart rate HR i >(220 - K) × 0.6, it is still judged as the exercise state. When the heart rate HR i <(220 - K) × 0.6, it is judged as the end of the exercise state. (220 - K) × 0.6 is a reference threshold for distinguishing the heart rate between the exercise and non-exercise states, which is universal for the elderly;

[0176] The resting state judgment conditions are as follows:

[0177] When the human body is at rest, it is not completely stationary, but the movement is smaller than during exercise, and the frequency and amplitude of the movement are also smaller. Although the body is relatively calm at rest, it is higher than during sleep, and the speed is also relatively slow. Therefore, for a certain time interval t jg , when the acceleration amplitude satisfies 0.1 < A i < 0.3, the acceleration change frequency 0.5 < f < 2, the heart rate change rate 10 > ΔHR i > 5, the position movement speed 0.5 ≥ v i ≥ 0.1, it is judged as the resting state;

[0178] Under certain special physiological or psychological stress states, such as sudden fright, extreme emotional fluctuations, etc., the body's various indicators will also show atypical changes. If within a certain time interval t jg , the data does not meet the judgment conditions of the sleep, exercise, and resting states, then the state corresponding to this time interval is judged as other states, which can accommodate these individual differences. When there are more atypical changes, it is also conducive to discovering abnormal situations and arranging further examinations.

[0179] A health risk construction module, which is used to construct a risk formula based on cardiovascular parameters and state division parameters in different states. The risk formula includes the diurnal blood pressure change rate, the standard deviation of heart rate during sleep, the change amount of pulse wave velocity during exercise, and the average value of cardiac stroke volume during the resting state;

[0180] Clinically, the diurnal blood pressure change situation is an important indicator for evaluating cardiovascular diseases such as hypertension, providing a basis for judging whether the blood pressure is normal and whether there is abnormal blood pressure rhythm. The 24-hour data of a day is divided into daytime and nighttime. Daytime period: 6:00 - 22:00, nighttime period: 22:00 - 6:00 the next day;

[0181] Record the blood pressure during the daytime period as SBP day，i , DBP day，i ; Record the blood pressure during the nighttime period as SBP night，i , DBP night，i ;

[0182] Calculate the mean systolic blood pressure of the blood pressure during the daytime period:

[0183]

[0184] In the formula, μ SBP，day represents the mean systolic blood pressure of the blood pressure during the daytime period, and n day represents the number of acquisitions during the daytime period;

[0185] Calculate the mean diastolic blood pressure during the daytime period:

[0186]

[0187] Where, μ DBP，day represents the mean diastolic blood pressure during the daytime period;

[0188] Calculate the mean systolic blood pressure during the nighttime period:

[0189]

[0190] Where, μ SBP，day represents the mean systolic blood pressure during the nighttime period, and n nigh represents the number of acquisitions during the nighttime period;

[0191] Calculate the mean diastolic blood pressure during the nighttime period:

[0192]

[0193] Where, μ DBP，night represents the mean diastolic blood pressure during the nighttime period;

[0194] Calculate the diurnal blood pressure variation rate:

[0195]

[0196] Where, NDBPR SBP represents the systolic blood pressure variation rate in the diurnal blood pressure variation rate, and NDBPR DBP represents the diastolic blood pressure variation rate in the diurnal blood pressure variation rate. Under normal circumstances, the human blood pressure shows a "dipper" rhythm, that is, the blood pressure is relatively high during the daytime and relatively low at night. If the blood pressure at night fails to decrease significantly or even increases, it may indicate problems in the cardiovascular system, which is usually related to problems such as hypertension, heart diseases, and even sleep apnea;

[0197] Extract the heart rate data in the sleep state and record it as HR sleep，ai' , ai' = 1, 2, 3,..., n sleep , where, n sleep represents the number of heart rates recorded in the sleep state, and calculate the mean heart rate in the sleep state:

[0198]

[0199] Where, μ HR，sleep represents the mean heart rate in the sleep state;

[0200] Calculate the standard deviation of the heart rate in the sleep state:

[0201]

[0202] In the formula, σ HR，sleep represents the standard deviation of the heart rate during sleep. Many diseases can affect the heart rate changes during sleep, which in turn leads to abnormalities in the standard deviation of the heart rate. For example, patients with sleep apnea hypopnea syndrome will repeatedly experience apnea and hypopnea during sleep, causing periodic changes in the heart rate and increasing the standard deviation of the heart rate. Another example is that patients with cardiovascular diseases such as coronary heart disease and arrhythmia may also experience abnormal fluctuations in the heart rate during sleep, manifested as changes in the standard deviation of the heart rate;

[0203] Extract the pulse wave velocity data in the exercise state, that is, PWV data, and record it as PWV exercise，bi' , bi' = 1, 2, 3,..., n exercise , where n exercise represents the number of PWV records in the exercise state;

[0204] Calculate the change in pulse wave velocity in the exercise state:

[0205] ΔPWV exercise = max(PWV exercise , bi') - min(PWV exercise , bi')

[0206] In the formula, ΔPWV exercise represents the change in pulse wave velocity in the exercise state, max(PWV exercise , bi') represents the maximum value of all PWV exercise，bi' data recorded in the exercise state, min(PWV exercise , bi') represents the minimum value of all PWV exercise，bi' data recorded in the exercise state. The change in PWV can indirectly reflect the pumping function of the heart and the compliance of blood vessels. Under normal exercise conditions, the change in PWV will be relatively small, indicating that the blood vessel elasticity can maintain good stability during exercise and the cardiovascular function is strong. Otherwise, the change in PWV in the exercise state will be abnormal, which may be an early warning signal for cardiovascular diseases such as hypertension and arteriosclerosis;

[0207] Extract the stroke volume data of the heart in the resting state, that is, SV data, and record it as SV rest，ci' , ci' = 1, 2, 3,..., n rest , where n rest represents the number of SV records in the resting state;

[0208] Calculate the average value of the stroke volume of the heart in the resting state:

[0209]

[0210] where μ SV，rest represents the average value of the cardiac stroke volume at rest. Under the resting state, the heart is not strongly interfered by external factors such as exercise. At this time, the stroke volume can more purely reflect the pumping ability and functional state of the heart itself. If the measured value deviates from the normal range for a long time, whether it is too high or too low, it may indicate that there are certain problems with the heart function.

[0211] A comprehensive risk assessment and judgment module, which is used to construct a comprehensive risk assessment formula by weighted summation according to the risk formula, and evaluate the overall health status and risk of the elderly through the result of the comprehensive risk assessment formula;

[0212] Through a large number of medical research, clinical practice, and statistical analysis of the physiological data of a large-scale elderly population, the normal range of the circadian change rate of systolic blood pressure in the elderly is set to 10 - 20%, the normal range of the circadian change rate of diastolic blood pressure is 8 - 15%, and the normal range of the standard deviation of heart rate during sleep is 30 - 60 ms; the normal range of the change amount of pulse wave velocity during exercise is 1 - 2 m / s, and the normal range of the average value of the cardiac stroke volume at rest is 50 - 80 ml;

[0213] Construct a circadian blood pressure change rate risk formula:

[0214]

[0215] where R bloodpressure represents the blood pressure risk, and α, β represent the weight coefficients;

[0216] When , it means that the circadian change rate of systolic blood pressure is within the normal range at this time, and the risk brought by diastolic blood pressure abnormality should be considered; conversely when, the risk brought by systolic blood pressure abnormality should be considered;

[0217] Therefore, when NDBPR SBP ∈[10 - 20], the weight coefficient α = 0 in the blood pressure risk formula, so that it does not affect the evaluation result in the subsequent comprehensive risk assessment. Generally speaking, the impact of systolic blood pressure on the cardiovascular system is more direct and significant. Too high or too low systolic blood pressure may lead to problems such as increased heart burden and increased blood vessel wall pressure. In contrast, the impact of diastolic blood pressure is slightly weaker. Therefore, β = 0.6;

[0218] When NDBPR DBP ∈[8 - 15], the weight coefficient β = 0 in the blood pressure risk formula, so that it does not affect the evaluation result in the subsequent comprehensive risk assessment. At this time, the abnormal circadian change rate of systolic blood pressure is the main risk factor and the diastolic blood pressure is normal. Therefore, α = 0.7;

[0219] When NDBPR SBP ∈ [10 - 20], NDBPR DBP ∈ [8 - 15], since both systolic blood pressure and diastolic blood pressure are abnormal and both have important impacts on blood pressure risk, but the abnormality of systolic blood pressure may be slightly dominant in the risk of cardiovascular diseases. Therefore, in the blood pressure risk formula, the weight coefficients α > β > 0. Thus, α = 0.4 and β = 0.3;

[0220] Construct the heart rate standard deviation risk formula:

[0221]

[0222] In the formula, R hartrate represents the risk of heart rate standard deviation during sleep, and γ represents the weight coefficient;

[0223] When the heart rate standard deviation exceeds the normal range, indicating that the heart rate variability is abnormal. γ > 0, and the value is γ = 0.4 to give corresponding consideration in relevant evaluations. When σ HR，sleep ∈ [30 - 60], it indicates that the cardiac autonomic nerve regulation is relatively normal, and γ = 0, indicating that it does not affect the evaluation result in the subsequent comprehensive risk assessment;

[0224] Construct the risk formula for the change in pulse wave velocity:

[0225]

[0226] In the formula, R PWV represents the risk of the change in pulse wave velocity, and δ represents the weight coefficient;

[0227] When it means that there may be abnormal changes in vascular function. δ > 0, and the value is δ = 0.35, indicating that a weight coefficient greater than 0 is given to increase the role of this factor in the comprehensive risk assessment. When ΔPWV exercise ∈ [1 - 2], it indicates that the elasticity and function of the blood vessels are in a relatively stable and healthy state under normal exercise stimulation, and δ = 0, which does not affect the evaluation result in the subsequent comprehensive risk assessment;

[0228] Construct the risk formula for the average value of cardiac stroke volume:

[0229]

[0230] In the formula, R SV represents the risk of the average value of cardiac stroke volume, and ε represents the weight coefficient;

[0231] When When the pumping function of the heart may be abnormal, it is necessary to reflect the impact of this abnormality on the overall risk in the comprehensive risk assessment. ε > 0, and the value is ε = 0.25. When μ SV，rest ∈[50 - 80], the pumping function of the heart is in a relatively stable and normal state, ε = 0, and it does not affect the assessment result in the subsequent comprehensive risk assessment.

[0232] In the actual data collection process, cardiovascular parameters (blood pressure, heart rate, pulse wave velocity, stroke volume) and state classification parameters (acceleration, position data) are all collected at equal time intervals to ensure the regularity and comparability of the data, facilitating accurate data processing and analysis in the follow-up. At the same time, to ensure the consistency and accuracy between different parameters, it is necessary to align the timestamps of the collected data. Therefore, after collecting the data, it is necessary to calibrate the timestamps of each parameter to make them precisely corresponding in time.

[0233] After that, based on the diurnal variation rate of systolic and diastolic blood pressure, the standard deviation of heart rate during sleep, the change in pulse wave velocity during exercise, and the average value of stroke volume at rest, a comprehensive risk assessment formula is constructed by weighted summation:

[0234] R total = ω1×R bloodpressure + ω2×R hartrate + ω3×R PWV + ω4×R SV

[0235] In the formula, R total represents the comprehensive risk, ω1, ω2, ω3, ω4 represent the weight coefficients, and ω1 + ω2 + ω3 + ω4 = 1, ω1×ω2×ω3×ω4 > 0. Since it varies due to various factors such as individual characteristics, disease types, and research purposes, the importance of each index in the comprehensive risk assessment can be flexibly adjusted according to the specific situation to make the assessment result more in line with the actual situation and improve the accuracy and pertinence of the assessment. As a quantitative manifestation of the comprehensive risk, the higher its value, the greater the degree of deviation of indicators such as the diurnal variation rate of systolic and diastolic blood pressure, the standard deviation of heart rate during sleep, the change in pulse wave velocity during exercise, and the average value of stroke volume at rest from the normal range, comprehensively reflecting a higher risk faced by the physical health of the elderly;

[0236] For the health risk assessment of the general elderly population, considering that the elderly have a relatively high risk of cardiovascular diseases, and indicators such as blood pressure, heart rate, vascular function, and heart pumping function are all crucial for cardiovascular health;

[0237] The default setting is ω1 = 0.4. The reason is that the circadian variation of blood pressure has a greater impact on the cardiovascular system. Long-term abnormal fluctuations in blood pressure are closely related to the occurrence and development of various cardiovascular diseases. For example, hypertension increases the burden on the heart, leading to changes in the structure and function of the heart, and also damages the vascular endothelium, triggering arteriosclerosis, etc. For the elderly, blood pressure problems are more common and more harmful, so a relatively higher weight is given;

[0238] ω2 = 0.2. The standard deviation of sleep heart rate reflects the cardiac autonomic nervous regulation function. Unstable heart rate during sleep may indicate an increased risk of cardiovascular disease and also affects sleep quality, thereby affecting overall health. However, compared with abnormal blood pressure, its impact on the comprehensive risk is slightly lower, so a relatively smaller weight is assigned;

[0239] ω3 = 0.3. The change in pulse wave velocity can reflect vascular elasticity and function. Vascular health is crucial for the elderly to prevent cardiovascular diseases. However, in the general elderly population, it does not affect health risks as directly and commonly as abnormal blood pressure, so the weight is slightly lower than that of blood pressure risk;

[0240] ω4 = 0.1. For the general elderly population, the resting stroke volume reflects the pumping function of the heart. It is important for the assessment of heart health, but compared with other factors, its relative contribution to the overall risk is relatively small, so a lower weight is assigned.

[0241] When R total ≤ 1.2, it is determined to be in a healthy state. When 1.2 < R total ≤ 1.5, it is determined that the physical health state is at a medium - low risk. When 1.5 < R total , it is determined that the physical health state is at a high - risk state.

[0242] For example, for a 70 - year - old elderly person, the calculated circadian variation rate of systolic blood pressure NDBPR SBP = 25%, the circadian variation rate of diastolic blood pressure NDBPR DBP = 12%, the standard deviation of heart rate σ HR，sleep = 70 during sleep, the change in pulse wave velocity ΔPWV exercise = 0.8 during movement, and the average value of the resting stroke volume of the heart μ SV，rest = 45. Then

[0243]

[0244] Therefore, the comprehensive risk assessment formula:

[0245] R total = ω1×R bloodpressure +ω2×Rhartrate +ω3×R PWV +ω4×R SV

[0246] = 0.4 × 1.35 + 0.2 × 1.27 + 0.3 × 1.28 + 0.1 × 1.21 = 1.299 ≈ 1.3

[0247] Since 1.2 < R total ≤ 1.5, the physical health status of this elderly person is in a medium - low risk state.

[0248] The above - mentioned formulas are all dimensionless and only take their numerical values for calculation. The formula is obtained by collecting a large amount of data and performing software simulation to get a formula closest to the real situation. The preset parameters in the formula are set by those skilled in the art according to the actual situation.

[0249] The above - mentioned embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination. When implemented using software, the above - mentioned embodiments can be implemented in whole or in part in the form of a computer program product. Those skilled in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in this article can be implemented by electronic hardware or the combination of computer software and electronic hardware. Whether these functions are executed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0250] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units. They can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0251] The above - mentioned is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all should be covered by the protection scope of this application.

Claims

1. An artificial intelligence-based early diagnosis and risk assessment system for the health of the elderly, characterized in that the system include: A data acquisition module, the data acquisition module is used to collect cardiovascular parameters of the elderly at equal time intervals, remove abnormal values ​​and smooth cardiovascular parameters through preprocessing, the cardiovascular parameters include blood pressure, heart rate, pulse wave velocity and heart rate per stroke data; A state judgment module, which is used to collect state classification parameters of the elderly at equal time intervals, and then judge the sleep, exercise and resting states of the elderly based on the state classification parameters, heart rate change rate and blood pressure. The state classification parameters include acceleration and position data; A health risk construction module, wherein the health risk construction module is used to construct a risk formula based on cardiovascular parameters and state classification parameters under different states, wherein the risk formula includes the diurnal blood pressure change rate, the standard deviation of the heart rate in the sleeping state, the change of the pulse wave velocity in the exercise state, and the average heart stroke volume in the resting state; The comprehensive risk assessment and judgment module is used to construct a comprehensive risk assessment formula by weighted summation based on the risk formula, and evaluate the overall health status and risk of the elderly through the results of the comprehensive risk assessment formula.

2. According to claim 1, an artificial intelligence-based early diagnosis and risk assessment system for the health of the elderly, characterized in that: The data acquisition module collects cardiovascular parameters of the elderly at the same time interval, including the following steps: Set the time interval to t jg , at the same time interval t jg Collect blood pressure, heart rate, pulse wave velocity and cardiac output data of the elderly; Blood pressure was recorded as [SBP i , DBP i ], where SBP is the systolic pressure in the blood pressure data, DBP is the diastolic pressure in the blood pressure data, and the recorded heart rate is HR i , record the pulse wave velocity as PWV i SV is the cardiac output per stroke. i , where i represents the serial number of the collection. The records of each day are renumbered starting from 6:00 a.m., i = 1, 2, 3, ..., N, where N is a positive integer ≥ 0, representing the total number of collections from 6:00 a.m. each day to 6:00 a.m. the next day, with a timestamp attached.

3. The artificial intelligence-based early diagnosis and risk assessment system for the health of the elderly according to claim 2 is characterized in that: The method of removing outliers and smoothing cardiovascular parameters by preprocessing in the data acquisition module is: The historical data of the most recent three complete days is selected as the data source for calculating the mean and standard deviation, that is, 6:00 three days ago, and continues until three days after the starting time point. The total duration of the data source is: T z ＝4320 Where, T z It is expressed as the total duration of the data source; Then calculate the number of data collected in the last 3 complete days: Where n g It is expressed as the number of data collected in the last 3 complete days; Calculate the means of cardiovascular parameters: In the formula, μ represents the mean value of cardiovascular parameters, x i It is expressed as the value of the cardiovascular physiological parameter at the i-th data point; Calculate the standard deviation of cardiovascular parameters: In the formula, σ represents the standard deviation of cardiovascular parameters; For each cardiovascular parameter x i , when |x is satisfied i -μ|>3σ, the cardiovascular parameter x is determined i are outliers and are replaced by the mean μ of cardiovascular parameters; The moving average filtering method is used to smooth each cardiovascular parameter. First, the moving average window size is set: In the formula, ω represents the sliding window size and is a positive integer; Calculate the moving average: In the formula, It is represented as the cardiovascular parameter corresponding to the smoothed i-th data point, x j It is represented as the cardiovascular parameter corresponding to the jth data point directly collected. For the data that cannot meet the complete window calculation, the original value is retained. Then, the smoothed data is Replace the original data x i .

4. The artificial intelligence-based early diagnosis and risk assessment system for the health of the elderly according to claim 1 is characterized by: The method by which the state judgment module judges the sleep, exercise and resting states of the elderly based on the state classification parameters, heart rate change rate and blood pressure is as follows: The acceleration of the elderly is collected by an accelerometer, and the amplitude of the acceleration is calculated based on the acceleration: In the formula, A i It is expressed as the total amount of acceleration collected for the i-th time, that is, the acceleration amplitude, a x 、a y 、a z Expressed as the components of acceleration on three coordinate axes; Statistics for each time interval t jg Internal acceleration amplitude A i The number of times the change exceeds 0.05 n l , calculate the frequency of acceleration change: Where f is the frequency of acceleration change, n l Expressed as acceleration amplitude A i The number of changes exceeding 0.05; Calculate the rate of change of heart rate: In the formula, ΔHR i It is expressed as jg The average change of heart rate in a period of time, that is, the heart rate change rate, HR i+1 Represents the heart rate value collected at the i+1th time, HR i Represents the heart rate value collected for the i-th time; According to the position data of the state division parameters, record each time interval t jg The position coordinates before and after (x1, y1) and (x2, y2), calculate the displacement distance: In the formula, d represents the q Displacement distance within Calculate movement speed: In the formula, v i It is expressed as the moving speed in the i-th time interval; Calculate the fluctuation of blood pressure: In the formula, ΔSBP i It is expressed as jg The average change in systolic blood pressure during a period of time, that is, the systolic blood pressure fluctuation, SBP i+1 It is represented by the systolic blood pressure collected at the i+1th time, SBP i Expressed as the systolic blood pressure collected at the i-th time; ΔDBP i It is expressed as jg The average change in diastolic blood pressure during a period of time, that is, the diastolic blood pressure fluctuation, DBP i+1 It is expressed as the diastolic pressure collected at the i+1th time, DBP i It is represented as the diastolic pressure collected at the i-th time; The sleep status judgment conditions are as follows: For a certain time interval t jg , the acceleration amplitude A must be satisfied at the same time i <0.1, acceleration change frequency f<0.5 and position movement speed v i <0.1, on this basis, the heart rate change rate ΔHR must also be met i <5, or the systolic blood pressure fluctuation ΔSBP is satisfied at the same time i <10 and diastolic blood pressure fluctuation ΔDBP i <5; The conditions for judging the motion state are as follows: Record the age K of the elderly for a certain time interval t jg When the acceleration amplitude satisfies 0.3 < A i < 1.5, the acceleration change frequency 2 < f < 10, the heart rate change rate ΔHR i > 10, the position movement speed 5 > v i > 0.

5. After it is determined to be in the motion state, when the first heart rate change rate ΔHR i ≤ 10, when the heart rate HR i > (220 - K) × 0.6, it is still determined to be in the motion state. When the heart rate HR i < (220 - K) × 0.6, it is determined that the motion state ends; The resting state judgment conditions are as follows: For a certain time interval t jg , when the acceleration amplitude satisfies 0.1 < A i < 0.3, the acceleration change frequency 0.5 < f < 2, the heart rate change rate 10 > ΔHR i > 5, the position movement speed 0.5 ≥ v i ≥ 0.1; If at a certain time interval t jg If the data does not meet the judgment conditions of sleep, exercise and rest state within the time interval, the state corresponding to the time interval is judged as other states.

5. The artificial intelligence-based early diagnosis and risk assessment system for the health of the elderly according to claim 4 is characterized by: The method for constructing the risk formula according to the cardiovascular parameters and state classification parameters in different states by the state judgment module is: First, divide the 24-hour data into daytime and nighttime: daytime time period: 6:00-22:00, nighttime time period: 22:00-6:00 the next day; The blood pressure during the day is recorded as SBP day，i , DBP day，i ; Blood pressure during the night period is recorded as SBP night，i , DBP night，i ; Calculate the mean systolic blood pressure during the day: In the formula, μ SBP，day It is expressed as the mean systolic blood pressure during the daytime, n day It is expressed as the number of acquisitions during the daytime period; Calculate the mean diastolic blood pressure during the day: In the formula, μ DBP，day It is expressed as the mean diastolic blood pressure during the daytime period; Calculate the mean systolic blood pressure during the nighttime period: In the formula, μ SBP，day Expressed as the mean systolic blood pressure during the night period, n nigh It is expressed as the number of acquisitions during the night time period; Calculate the mean diastolic blood pressure during the night time period: In the formula, μ DBP，night It is expressed as the mean diastolic blood pressure during the night period; Calculate the diurnal blood pressure variation rate: Where, NDBPR SBP Expressed as the systolic blood pressure change rate in the diurnal blood pressure change rate, NDBPR DBP It is expressed as the diastolic blood pressure change rate in the diurnal blood pressure change rate; Extract heart rate data in sleeping state and record it as HR sleep，ai' , ai'=1,2,3,…,n sleep , where n sleep It is expressed as the number of heart rates recorded during sleep, and the average heart rate during sleep is calculated: In the formula, μ HR，sleep It is expressed as the mean heart rate in the sleeping state; Calculate the standard deviation of heart rate during sleep: In the formula, σ HR，sleep It is expressed as the standard deviation of heart rate in the sleeping state; Extract the pulse wave velocity data of the exercise state, that is, PWV data, and record it as PWV exercise，bi' , bi'=1,2,3,…,n exercise , where n exercise It is expressed as the number of PWV recorded during exercise; Calculate the change in pulse wave velocity during exercise: ΔPWV exercise =max(PWV exercise ,b)-min(PWV exercise ,the') Where ΔPWV exercise Expressed as the change in pulse wave velocity during exercise, max(PWV exercise , bi') represents all PWV recorded in the motion state exercise，bi' The maximum value of the data, min(PWV exercise , bi') represents all PWV recorded in the motion state exercise , bi' The minimum value of the data; Extract the heart's stroke volume data at rest, that is, SV data recorded as SV rest，ci' , ci'=1,2,3,…,n rest , where n rest It is expressed as the number of SVs recorded in the resting state; Calculate the average cardiac output per stroke at rest: In the formula, μ SV，rest It is expressed as the average cardiac output per stroke at rest.

6. The artificial intelligence-based early diagnosis and risk assessment system for the health of the elderly according to claim 5 is characterized by: The method for constructing the risk formula of the health risk construction module according to the cardiovascular parameters under different states and the state division parameters is as follows: The normal range of diurnal variation of systolic blood pressure in the elderly is set at 10-20%, the normal range of diurnal variation of diastolic blood pressure is set at 8-15%, the standard deviation of heart rate in normal sleep state is set at 30-60ms; the range of variation of pulse wave velocity in normal exercise state is set at 1-2m / s, and the average range of heart stroke volume in normal resting state is set at 50-80ml; Construct the risk formula of diurnal blood pressure change rate: In the formula, R bloodpressure It is expressed as blood pressure risk, α and β are weight coefficients, α≥0, β≥0; When NDBPR SBP ∈[10-20], When , the weight coefficient α in the blood pressure risk formula is 0; when NDBPR DBP When ∈[8-15], the weight coefficient β in the blood pressure risk formula is 0; When NDBPR SBP ∈[10-20], NDBPR DBP ∈[8-15], the weight coefficient α>β>0 in the blood pressure risk formula; Construct the heart rate standard deviation risk formula: In the formula, R hartrate It is expressed as the standard deviation risk of heart rate in sleeping state, and γ is expressed as the weight coefficient; when When γ>0, when σ HR，sleep ∈[30-60], γ=0; Construct the risk formula of pulse wave velocity change: In the formula, R PWW It is expressed as the risk of change in pulse wave velocity, and δ is expressed as the weight coefficient; when When δ>0, when ΔPWV exercise ∈[1-2], δ=0; Construct the risk formula for the mean stroke volume: In the formula, R SV It is expressed as the risk of the mean cardiac output per stroke, and ε is expressed as the weight coefficient; when When ε>0, when μ SV，rest When ∈[50-80], ε=0.

7. The artificial intelligence-based early diagnosis and risk assessment system for the health of the elderly according to claim 6 is characterized by: The comprehensive risk assessment and judgment module constructs a comprehensive risk assessment formula by weighted summation according to the risk formula, and the overall health status and risk assessment of the elderly are evaluated by the results of the comprehensive risk assessment formula, including the following steps: Constructing a comprehensive risk assessment formula: R total =ω1×R bloodpressure +ω2×R hartrate +ω3×R PWV +ω4×R SV In the formula, R total It is expressed as comprehensive risk, ω1, ω2, ω3, ω4 are expressed as weight coefficients, and ω1+ω2+ω3+ω4=1, ω1×ω2×ω3×ω4>0; When R total When ≤1.2, it is judged as healthy state, 1.2 <R total When the risk is ≤1.5, the health status is judged to be at medium to low risk. <R total When the patient's health condition is judged to be at high risk.

Images