Dynamic monitoring method and system for elderly people living alone based on smart community

Abstract

This application discloses a method for dynamic monitoring of elderly people living alone in a smart community. The method includes: obtaining individual portrait data of the elderly person; generating customized monitoring indicators and customized warning thresholds for the elderly person based on the individual portrait data; constructing a behavior association matrix based on the customized monitoring indicators using a preset method; obtaining the elderly person's monitoring data in real time and determining whether the monitoring data triggers the behavior association matrix; if so, updating the customized warning threshold based on the behavior association matrix; determining whether the monitoring data reaches the customized warning threshold; and outputting a warning prompt message if the monitoring data reaches the customized warning threshold. By applying a customized monitoring strategy to the elderly person's individual portrait data and dynamically adjusting the monitoring warning threshold according to the elderly person's real-time behavior patterns, more accurate monitoring can be achieved.

Description

Technical Field

[0001] The present application relates to the field of intelligent monitoring technology, and in particular to a method and system for dynamic monitoring of elderly people living alone based on a smart community. Background Art

[0002] Intelligent monitoring for elderly people living alone based on smart communities is a three-dimensional monitoring network built for elderly people living alone by deeply integrating community public service resources with intelligent technologies, providing them with safer and more convenient life guarantees.

[0003] Existing monitoring methods for elderly people living alone mainly collect multi-dimensional data on the elderly by installing intelligent monitoring equipment, setting fixed warning thresholds, and combining them with emergency alarm systems to provide timely assistance to the elderly. However, due to the differences in lifestyles and physical conditions of different elderly people, and the impact of their current behavioral status on monitoring indicators, the monitoring method with fixed warning thresholds is prone to misjudgment, which can have a negative impact on the lives of elderly people living alone and even cause them to lose their safety. Summary of the Invention

[0004] The purpose of this application is to provide a dynamic monitoring method and system for elderly people living alone based on a smart community, which can provide customized monitoring strategies based on the individual portrait data of the elderly, and dynamically adjust the monitoring warning threshold according to the real-time behavior patterns of the elderly, thereby achieving more accurate monitoring.

[0005] In the first aspect, the present application provides a method for dynamic monitoring of elderly people living alone based on a smart community, which adopts the following technical solutions:

[0006] Obtain individual portrait data of elderly people living alone;

[0007] Generate customized monitoring indicators and customized warning thresholds for elderly people living alone based on individual portrait data;

[0008] Based on customized monitoring indicators, a behavioral correlation matrix is ​​constructed through a preset method;

[0009] Obtain monitoring data of elderly people living alone in real time and determine whether the monitoring data triggers the behavioral association matrix;

[0010] If so, the customized warning threshold is updated based on the behavior correlation matrix;

[0011] Determine whether the monitoring data reaches the customized warning threshold;

[0012] If the monitoring data reaches the customized warning threshold, a warning prompt message will be output.

[0013] Through the above technical solution, customized monitoring indicators and warning thresholds can be established based on the individual portrait data of the elderly, and a behavioral correlation matrix can be constructed based on the impact of the elderly's behavior patterns on the monitoring indicators. The warning thresholds of monitoring can be dynamically adjusted according to the real-time behavior patterns of the elderly, thereby achieving more accurate monitoring.

[0014] Optionally, the behavior association matrix is ​​constructed based on the customized monitoring indicators by a preset method, including:

[0015] Based on individual portrait data, obtain a set of behavior tags;

[0016] Traverse the behavior tag set and obtain the correlation coefficient of the behavior tag to each monitoring indicator through the preset method;

[0017] After the traversal is completed, a behavior correlation matrix is ​​generated based on the correlation coefficients of all behavior labels and various monitoring indicators.

[0018] Optionally, the monitoring data includes behavioral data, and determining whether the monitoring data triggers a correlation matrix includes:

[0019] Based on the current behavior data, obtain the current behavior label;

[0020] Determine whether the current behavior label matches the behavior association matrix;

[0021] If not, the current behavior tag is recorded as a new behavior tag;

[0022] If so, then determine the associated monitoring indicators based on the behavior tags;

[0023] Determine whether the correlation coefficient of the associated monitoring indicator is greater than a preset correlation threshold;

[0024] If the correlation coefficient of the associated monitoring indicator is greater than the preset correlation threshold, a trigger response message is output.

[0025] Optionally, after recording the current behavior tag as a new behavior tag, the step further includes:

[0026] Obtain monitoring data when the new behavior occurs and record it as control monitoring data;

[0027] Compare the control monitoring data with the monitoring data under normal conditions and calculate the monitoring indicator deviation;

[0028] Determine whether the monitoring indicator deviation reaches the preset indicator deviation threshold;

[0029] If so, the behavior association matrix is ​​updated based on the monitoring indicator deviation.

[0030] Optionally, updating the customized warning threshold based on the behavior association matrix includes:

[0031] Based on the behavior association matrix, determine the correlation coefficient of the current behavior label to the associated monitoring indicator;

[0032] Based on the correlation coefficient, calculate and obtain the expected value of the indicator through the preset calculation rules;

[0033] Update customized warning thresholds based on expected indicator values.

[0034] Optionally, after outputting the warning prompt information, the method further includes:

[0035] Push the warning information to the manual verification department for verification and obtain the verification feedback results;

[0036] If the verification feedback result is a valid warning, the monitoring data associated with the warning will be added to the preset warning database;

[0037] If the amount of data in the early warning database reaches the preset threshold, a risk prediction model is constructed based on the data in the early warning database through big data analysis and modeling;

[0038] If the verification feedback result is a false alarm, the customized warning threshold is updated based on the associated monitoring data.

[0039] Optionally, after obtaining the monitoring data of the elderly living alone in real time, the method further includes:

[0040] Set the time period, perform data statistics on the historical monitoring data in units of time periods, and obtain the monitoring data statistics results within each time period;

[0041] Taking two adjacent time periods as a group, calculate the monitoring data deviation under the two time periods;

[0042] Based on the deviation of monitoring data, identify the different behavior patterns under two time periods;

[0043] Based on the monitoring data deviation, the impact factors of different behavior patterns on related monitoring indicators are obtained;

[0044] Update customized warning thresholds based on the impact factors of distinguishing behavioral patterns on related monitoring indicators.

[0045] Optionally, the individual portrait data includes residence information, and after obtaining the individual portrait data of the elderly living alone, the following steps are included:

[0046] Based on the residential information of elderly people living alone, a guardianship network distribution map is constructed;

[0047] Based on the monitoring network distribution map, set up associated community monitoring points.

[0048] In a second aspect, the present application provides a dynamic monitoring system for elderly people living alone based on a smart community, comprising:

[0049] The data acquisition module 101 is used to obtain individual portrait data of elderly people living alone;

[0050] The monitoring plan customization module 102 is used to generate customized monitoring indicators and customized warning thresholds for elderly people living alone based on individual portrait data;

[0051] A behavior correlation matrix construction module 103 is used to construct a behavior correlation matrix based on customized monitoring indicators using a preset method;

[0052] The warning threshold updating module 104 is used to obtain monitoring data of elderly people living alone in real time and determine whether the monitoring data triggers the behavior association matrix. If so, it updates the customized warning threshold based on the behavior association matrix;

[0053] The monitoring and early warning judgment module 105 is used to judge whether the monitoring data reaches the customized early warning threshold. If the monitoring data reaches the customized early warning threshold, an early warning prompt message is output.

[0054] On the third aspect, the present application provides a computer-readable storage medium storing a computer program that can be loaded by a processor and execute the above-mentioned method for dynamic monitoring of elderly people living alone based on a smart community.

[0055] To sum up, this application first provides customized monitoring strategies for the individual portrait data of elderly people living alone, and dynamically adjusts the warning threshold of monitoring according to the real-time behavior patterns of elderly people living alone, so as to achieve more accurate monitoring; in addition, by recording the warning monitoring data and constructing a risk prediction model, it is possible to predict the warning risk based on the current behavior patterns of the elderly, so as to make preparations for the warning in advance and further enhance the safety monitoring protection; in addition, historical monitoring data is also periodically statistically analyzed to further measure the impact of behavior patterns on monitoring indicators, so as to realize dynamic adjustment of the warning threshold from multiple dimensions and provide a more suitable monitoring model for elderly people living alone. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This is a flow chart of a method for dynamic monitoring of elderly people living alone based on a smart community provided in an embodiment of the present application;

[0057] Figure 2 This is a flowchart of constructing a behavior association matrix provided by an embodiment of the present application;

[0058] Figure 3 This is a flow chart of determining whether monitoring data triggers a behavior association matrix provided by an embodiment of the present application;

[0059] Figure 4This is a flow chart of updating customized warning thresholds based on a behavior association matrix provided in an embodiment of the present application;

[0060] Figure 5 This is a schematic diagram of a dynamic monitoring system for elderly people living alone based on a smart community provided in an embodiment of the present application. DETAILED DESCRIPTION

[0061] The following is combined with Figure 1 -Attached Figure 5 , further details of this application are given.

[0062] This application provides a method for dynamic monitoring of elderly people living alone based on smart communities. Figure 1 , including the following steps:

[0063] S100. Obtain individual portrait data of elderly people living alone.

[0064] S200. Based on individual portrait data, generate customized monitoring indicators and customized warning thresholds for elderly people living alone.

[0065] Among them, individual portrait data includes personal basic information, health data, and daily behavior data. Personal basic information includes age, gender, family status, education level, occupational background, residential address, contact information, etc.; health data includes past medical history, current medication status and basic physical examination data; daily behavior data includes work and rest patterns, dietary preferences, exercise habits, and daily behavior (behavior at home, outdoor activities), etc.

[0066] First, with the authorization of the elderly living alone and their families, the individual portrait data of the elderly will be obtained through multiple channels.

[0067] Because the physical health status, daily routines and habits of different elderly people living alone are different, the corresponding monitoring indicators should also be different. For example, according to the medical information of the elderly living alone, the elderly need to take medicine on time every day, and whether the medicine is taken on time will also be used as a monitoring indicator; for example, if the elderly have the habit of morning exercises every day, outdoor monitoring and the time spent outdoors will be increased as monitoring indicators.

[0068] Therefore, in the embodiment of the present application, customized monitoring indicators will be generated based on the individual portrait data of the elderly living alone. The monitoring indicators will also be divided into several dimensions according to the monitoring means, including physiological indicators, behavioral indicators and environmental indicators. Physiological indicators, such as heart rate, blood sugar, sleep conditions, etc.; behavioral indicators, such as exercise duration and intensity, length of stay in a region, water intake, frequency of going to the toilet, frequency of going out, length of stay outside, falls, etc.; environmental indicators, such as abnormal water use, abnormal electricity use, gas leaks, smoke alarms, etc.

[0069] Specifically, generating customized monitoring indicators for elderly people living alone means using a universal standard monitoring indicator template, combined with the personalized characteristics of elderly people living alone, such as "frequently getting up at night" and "taking medicine on time", etc. The individual portrait data of the elderly can be broken down into dimensions, personalized characteristics can be extracted from them, and corresponding monitoring indicators can be generated. Then, customized monitoring indicators can be generated by combining with the standard monitoring indicator template.

[0070] Customized monitoring indicators are equivalent to converting the individual portrait data of elderly people living alone into quantifiable and monitorable monitoring dimensions, thereby forming a set of safety monitoring rules. For customized monitoring indicators, corresponding customized warning thresholds will also be generated. For example, for heart rate, the heart rate range under normal conditions is: 60-100 beats / minute. If it exceeds the normal range, an early warning will be triggered. For example, for gas leaks, the normal concentration is <50ppm. If the concentration is ≥50ppm, an alarm will be triggered. Of course, the early warning can also be divided into levels, for example, into mild early warning, moderate early warning and severe early warning, and the corresponding early warning thresholds can be set respectively. The setting can be made according to the actual situation, and this application does not make specific restrictions.

[0071] S300. Based on customized monitoring indicators, a behavior association matrix is ​​constructed through a preset method.

[0072] Taking into account that the behavior patterns of the elderly will also have an impact on monitoring indicators, for example, an increase in heart rate after exercise is a normal phenomenon, but if it passes through a fixed warning threshold, it is easy to make a misjudgment. Therefore, the impact of the elderly's behavior patterns on monitoring indicators can be modeled. In this way, the warning threshold can be re-determined based on the impact of the elderly's current behavior on the monitoring indicators, which can greatly reduce the false alarm rate of the warning.

[0073] Therefore, in an embodiment of the present application, after generating customized monitoring indicators based on the individual portrait data of the elderly, a behavior association matrix will also be generated based on the customized monitoring indicators. The so-called behavior association matrix represents the correlation and influence degree between the elderly's behavior patterns and various monitoring indicators.

[0074] Specifically, see Figure 2 , based on customized monitoring indicators, a behavior association matrix is ​​constructed through a preset method, including the following steps:

[0075] S310. Obtain a set of behavior tags based on individual portrait data.

[0076] S320: traverse the behavior tag set and obtain the correlation coefficient of the behavior tag to each monitoring indicator through a preset method.

[0077] S330: After the traversal is completed, a behavior correlation matrix is ​​generated based on the correlation coefficients of all behavior tags and various monitoring indicators.

[0078] Among them, behavioral labels are similar to the above-mentioned behavioral indicators. They represent the activity status of the elderly. Simply put, they are "what the elderly are currently doing." Based on the individual portrait data of the elderly and combined with the daily activity record data, all possible behavioral patterns of the elderly are determined and converted into structured features; such as behavior category, intensity, and duration. By combining all behavioral patterns, a set of behavioral labels can be generated.

[0079] First, based on the individual portrait data, a set of behavior labels can be obtained, and then the set of behavior labels can be traversed. For each behavior label, the correlation coefficient of the behavior label to each monitoring indicator can be obtained through a preset method. The preset method here adopts various correlation coefficient analysis methods. The so-called correlation coefficient analysis method refers to data statistical analysis based on historical data, and obtains the correlation between behavior labels and monitoring indicators through corresponding calculations. For example, the Pearson correlation coefficient is used, with behavior labels and monitoring indicators as variables, and the covariance of the two variables is calculated through historical statistical data. Then, their respective standard deviations are calculated to obtain the degree of correlation between the two variables.

[0080] The correlation coefficient includes correlation and influence coefficient, and its value range is [-1,1]. Correlation is divided into positive correlation, negative correlation and no correlation, that is, the numerical symbol, + indicates positive correlation, - indicates negative correlation, and 0 indicates no correlation; the influence coefficient represents the degree of influence of the behavior pattern on the monitoring indicator, that is, the numerical size.

[0081] After generating the correlation coefficient between each behavior and each monitoring indicator, the correlation coefficients of all behavior labels and each monitoring indicator are arranged and combined to generate a behavior correlation matrix.

[0082] S400: Acquire monitoring data of elderly people living alone in real time, and determine whether the monitoring data triggers a behavior association matrix.

[0083] S500: If the monitoring data triggers the behavior association matrix, the customized warning threshold is updated based on the behavior association matrix.

[0084] S600: Determine whether the monitoring data reaches a customized warning threshold.

[0085] S700: If the monitoring data reaches the customized warning threshold, output warning prompt information.

[0086] In an embodiment of the present application, after determining the monitoring indicators and warning thresholds based on the individual portrait data of the elderly living alone, real-time monitoring will be carried out to obtain monitoring data. The monitoring data includes physiological data (heart rate, blood pressure, blood oxygen, body temperature, body movement frequency, etc.), behavioral data (work and rest patterns, activity trajectories, movement patterns, etc.) and environmental data (indoor temperature, humidity, air quality, water and electricity usage, gas, smoke, etc.).

[0087] After the monitoring data is collected, judgments can be made based on the collected monitoring data. If the collected monitoring data reaches the customized warning threshold, a warning prompt message will be output to notify relevant family members or community personnel for assistance.

[0088] Because of the behavioral association matrix, during the real-time monitoring of elderly people living alone, after confirming the elderly person’s current behavioral label, the behavioral association matrix will be used to make a judgment, that is, to determine whether the monitoring data triggers the behavioral association matrix.

[0089] Specifically, see Figure 3 , to determine whether the monitoring data triggers the behavior association matrix, including the following steps:

[0090] S410: Obtain a current behavior label based on the current behavior data.

[0091] S420: Determine whether the current behavior label matches the behavior association matrix.

[0092] S430: If the current behavior label does not match the behavior association matrix, the current behavior label is recorded as a new behavior label.

[0093] S440: If the current behavior label matches the behavior association matrix, determine the associated monitoring indicator based on the behavior label.

[0094] S450: Determine whether the correlation coefficient of the associated monitoring indicator is greater than a preset correlation threshold.

[0095] S460: If the correlation coefficient of the associated monitoring indicator is greater than a preset correlation threshold, a trigger response message is output.

[0096] First, based on the current behavior data, the current behavior label is obtained. Because the behavior association matrix itself represents the association between the behavior label and various monitoring indicators, it is necessary to determine whether the current behavior label matches the behavior association matrix, that is, to determine whether the current behavior label appears in the behavior association matrix.

[0097] If the current behavior label matches the behavior association matrix, the associated monitoring indicator can be determined based on the current behavior label. Considering that not all behavior patterns will affect the monitoring indicator, it is also necessary to combine the intensity, duration and degree of impact of the current behavior on the monitoring indicator to determine it. Therefore, a correlation threshold will be set in advance for further judgment. Only when the correlation coefficient of the associated monitoring indicator is greater than the preset correlation threshold, it will be considered that the current behavior has a greater impact on the monitoring indicator and has reached a level that cannot be ignored. At this time, the trigger response information will be output to indicate that the current behavior data has triggered the behavior association matrix.

[0098] It is worth noting that the comparison between the correlation coefficient and the preset correlation threshold is based on the absolute value of the correlation coefficient.

[0099] In addition, if the current behavior label does not match the behavior association matrix, it means that the current behavior label does not exist in the behavior association matrix, that is, the behavior pattern has not appeared before. At this time, the current behavior label will be recorded as a new behavior label for subsequent supplementation and updating of the behavior association matrix. This will also help to maximize the accurate monitoring of elderly people living alone.

[0100] Therefore, in the embodiment of the present application, after recording the current behavior tag as a newly added behavior tag, the following steps are further included:

[0101] S431. Obtain monitoring data when the new behavior occurs, and record it as control monitoring data.

[0102] S432. Compare the control monitoring data with the monitoring data under normal conditions, and calculate the monitoring indicator deviation.

[0103] S433: Determine whether the monitoring indicator deviation reaches a preset indicator deviation threshold.

[0104] S434. If yes, then based on the monitoring indicator deviation, the behavior association matrix is ​​updated.

[0105] In the process of real-time monitoring of elderly people living alone, after determining that the elderly person's current behavior label is a newly added behavior label, the monitoring data of the new behavior occurrence stage will be recorded and then compared with the monitoring data under normal conditions. Because the monitoring data and the monitoring indicators correspond to each other, the monitoring data of the new behavior occurrence stage will be compared with the monitoring data under normal conditions. The monitoring indicators will be used as the benchmark, and then the monitoring data of the two states of the single monitoring indicators will be compared one by one. If there is a difference in the comparison results, the difference between the two will be calculated and recorded as the monitoring indicator deviation.

[0106] When the monitoring indicator deviation is very small, it is considered that the impact of the new behavior on the monitoring indicator can be ignored. In this case, the behavior association matrix will not be updated. Therefore, an indicator deviation threshold will be set in advance for judgment. Only when the monitoring indicator deviation reaches the preset indicator deviation threshold will the behavior association matrix be updated based on the monitoring indicator deviation.

[0107] Specifically, the behavior association matrix is ​​updated, that is, the association coefficient of the new behavior label to the monitoring indicator is calculated according to the monitoring indicator deviation, and after obtaining the association coefficient of the new behavior label to each monitoring indicator, it is added to the current behavior association matrix.

[0108] In an embodiment of the present application, if the current behavior data is monitored to trigger the behavior association matrix, the customized warning threshold will be updated based on the behavior association matrix.

[0109] Specifically, see Figure 4 , based on the behavior association matrix, update the customized warning threshold, including the following steps:

[0110] S510: Determine the correlation coefficient of the current behavior label to the associated monitoring indicator based on the behavior correlation matrix.

[0111] S520. Based on the correlation coefficient, calculate and obtain the expected value of the indicator through the preset calculation rules.

[0112] S530. Update the customized warning threshold based on the expected value of the indicator.

[0113] First, based on the behavior association matrix, the correlation coefficient of the current behavior label to the associated monitoring indicator is determined.

[0114] Then, using the preset calculation rules, the expected value of the indicator is calculated and obtained. The preset calculation rules here mainly use historical data statistics to determine the historical impact of the behavior label on the monitoring indicator. Then, combining the correlation coefficient and the normal state value of the monitoring indicator, the expected value of the indicator is calculated. The so-called expected value of the indicator is a form of expression that quantifies the impact of the current behavior on the monitoring indicator.

[0115] The correlation coefficient between behavior label and monitoring index is , the historical impact of behavior labels on monitoring indicators is , the normal state value of the monitoring indicator is , then the expected value of the indicator It can be expressed as:

[0116]

[0117] Finally, based on the expected value of the indicator, the customized warning threshold can be updated. For example, if an elderly person is performing light exercise, the correlation coefficient of "light exercise" to "heart rate" is determined to be 0.8 through the behavioral association matrix. The normal heart rate value is 60-100 beats / minute. In historical data, light exercise causes the heart rate to increase by an average of 20 beats / minute. By substituting the above expression for calculation, the expected heart rate value can be calculated to be 76-116 beats / minute. The corresponding customized warning threshold can be adjusted to: an alarm will be triggered outside the range of 76-116 beats / minute. In this way, if the elderly person's heart rate is monitored to be over 100, an alarm will be triggered, but after dynamic adjustment, it may be within the normal range, effectively avoiding the impact of false alarms.

[0118] Although the false alarm rate can be reduced to a certain extent by constructing a behavioral association matrix and dynamically adjusting the warning threshold according to the real-time behavioral patterns of the elderly, false alarms may still exist. Therefore, monitoring the elderly should be a continuous learning process, and each false alarm is also an opportunity to learn. By effectively analyzing the causes of false alarms, it is helpful to optimize and improve the current monitoring model.

[0119] In addition, for effective early warning, we can also judge which factors are likely to cause early warning based on the specific circumstances of the effective early warning. By analyzing a large amount of early warning record data, we can build a corresponding data model to predict the early warning situation during real-time monitoring.

[0120] Therefore, in the embodiment of the present application, after outputting the warning prompt information, the following steps are also included:

[0121] S710: Push the warning prompt information to the manual verification location for verification, and obtain the verification feedback result.

[0122] S720: If the verification feedback result is a valid warning, the monitoring data associated with the warning is added to the preset warning database.

[0123] S730. If the amount of data in the early warning database reaches a preset threshold, a risk prediction model is constructed based on the data in the early warning database through big data analysis and modeling.

[0124] S740: If the verification feedback result is a false alarm, the customized warning threshold is updated based on the associated monitoring data.

[0125] Among them, the manual verification department, that is, the related community staff and the corresponding terminal, will push the warning prompt information to the related staff for timely arrangement and processing after outputting it, and the related staff will determine whether the current warning information is valid, that is, determine the verification feedback result; the verification feedback results are divided into two types, namely effective warnings and false alarms.

[0126] If the verification feedback result is an effective warning, the monitoring data associated with the current warning will be added to the preset warning database. When the data in the warning database reaches a certain level, it can be based on all historical warning data and analyzed through big data modeling. It mainly collects data statistics on the occurrence of effective warnings, analyzes the reasons for the warnings, the environment in which the elderly are located, and the elderly’s current behavior patterns and physical conditions. Through data modeling, a risk prediction model can be constructed. This model can predict possible warning risks and probabilities based on the elderly’s current behavior patterns, environmental conditions, and physical conditions. With the help of this model, it can help predict potential risks so that response measures can be prepared in advance.

[0127] If the verification feedback result is a false alarm, the customized warning threshold will be updated based on the associated monitoring data.

[0128] Since it is a false alarm, the first thing we can determine is that the elderly person is still in a normal and safe state, that is, the current monitoring value has reached the customized warning threshold, but has not actually reached the warning state, that is, the current customized warning threshold is too high or too low. At this time, we can first determine the warning threshold of the associated monitoring indicator and the monitoring value that triggers the warning, compare the difference between the two values, and the time when the value changes, and combine the elderly person’s current behavior pattern to judge whether new behaviors have occurred, or the correlation coefficient of the behavior correlation matrix is ​​improperly set, so as to update the behavior correlation matrix in a targeted manner, and then use the behavior correlation matrix to update the customized warning threshold accordingly.

[0129] Some of the impacts of elderly behavior patterns on monitoring indicators are direct and manifest quickly, such as the effect of exercise on heart rate; others are more pronounced over a longer period of time, such as the effect of medication on blood sugar. While using a behavioral association matrix can effectively improve early warning accuracy for short-term behavioral patterns, it remains insufficient for long-term patterns.

[0130] To address this situation, in an embodiment of the present application, a time period is also set, and the historical monitoring data of the elderly are statistically analyzed in units of periods. The monitoring data and warning information in each period are compared, and the differentiated data are used to reversely infer the influencing factors of this differentiation. The influencing factors may be that the elderly have new behaviors in a certain period of time, or that certain behavioral patterns have changed. For example, the elderly have a record of taking medicine during the time period, which makes the elderly's blood sugar concentration lower than before. For example, the elderly have participated in community activities during the time period, and emotional changes indirectly affect physiological indicators.

[0131] Therefore, in the embodiment of the present application, after obtaining the monitoring data of the elderly living alone in real time, the following steps are also included:

[0132] S810: Set a time period, perform data statistics on the historical monitoring data in units of time periods, and obtain statistical results of the monitoring data in each time period.

[0133] S820: Taking two adjacent time periods as a group, calculate the monitoring data deviation in the two time periods.

[0134] S830. Determine the distinguishing behavior patterns in two time periods based on the monitoring data deviation.

[0135] S840. Based on the monitoring data deviation, obtain the impact factor of the distinguishing behavior pattern on the associated monitoring indicator.

[0136] S850. Update customized warning thresholds based on the impact factors of the distinguished behavior patterns on the associated monitoring indicators.

[0137] In an embodiment of the present application, after obtaining the real-time monitoring data of the elderly living alone, the monitoring data will be stored and recorded. When the amount of data reaches a certain level, a time period will be set first. It can be in days, weeks or months, and can be set according to actual conditions.

[0138] The historical monitoring data is then statistically analyzed in units of time periods to obtain the statistical results of the monitoring data within each time period. The monitoring data, as described above, includes physiological data, behavioral data, and environmental data. Statistics are performed on the numerical indicator data. Because the units are periodic, the data statistics are also standardized by period, such as calculating the mean, standard deviation, and variance.

[0139] Next, take two adjacent time periods as a group and calculate the monitoring data deviation under the two time periods. When the monitoring data deviation reaches a certain level, the distinguishing behavior patterns under the two time periods can be determined based on the monitoring data deviation. The so-called distinguishing behavior pattern is actually a statistical result, which can be the emergence of a new behavior pattern or the difference in the intensity and duration of the behavior pattern itself.

[0140] Based on the monitoring data deviation, the impact factor of the different behavior patterns on the related monitoring indicators can be calculated. The impact factor is similar to the correlation coefficient mentioned above. It is also a form of quantitative expression of the degree of influence. It can be calculated by combining the monitoring data deviation and the statistical mean of the monitoring data.

[0141] Finally, based on the impact factors of different behavioral patterns on related monitoring indicators, customized warning thresholds are updated. For example, during this period, the activity level of the elderly has decreased significantly, so the warning threshold for heart rate can be slightly lowered. In this way, the accuracy of monitoring and warning for the elderly can be further improved.

[0142] In addition, considering the living environment of the elderly themselves, as well as their current behavior and warning conditions, if the relevant community personnel are far away from the location of the elderly living alone after receiving the warning, rescue may not be timely. Therefore, it is necessary to simulate the assistance situation after the warning appears based on the living conditions of the elderly, and conduct a feasibility analysis to make corresponding adjustments.

[0143] Therefore, in the embodiment of the present application, after obtaining the individual portrait data of the elderly, the following steps are also included:

[0144] S910. Construct a monitoring network distribution map based on the residential information of elderly people living alone.

[0145] S920. Based on the monitoring network distribution map, set associated community monitoring points.

[0146] Among them, individual portrait data includes residential information. In addition to the elderly’s living location, residential information also includes the elderly’s living environment, such as the floor situation, whether there is an elevator, the surrounding environment and other information.

[0147] First, based on the residential information of all elderly people living alone, a guardianship network distribution map can be constructed. The so-called guardianship network distribution map can be regarded as a location distribution map, which takes the entire guardianship area as a map, marks the location of the elderly, and determines the path to each location.

[0148] After determining the locations of all elderly people living alone, when an early warning prompt appears, it is necessary to ensure that the associated community personnel can provide timely assistance. Associated community monitoring points will be set, and the distance or arrival time from the associated community monitoring point to each elderly person living alone will not exceed the set threshold as a constraint condition. The minimum number of associated community monitoring points will be determined, and then the monitoring network distribution map will be divided according to the associated community monitoring points. The divided monitoring network distribution map is used as a unit, and the distance or arrival time from the associated community monitoring point to each elderly person living alone is minimized as the objective function. The appropriate associated community monitoring points can be determined through calculation.

[0149] After determining the associated community monitoring point, in addition to receiving early warning information and arranging assistance in a timely manner, based on the current risk warning forecast of the elderly, messages can be pushed to the nearest associated community monitoring point in advance so that timely assistance preparations can be made. In this way, an additional layer of protection can be added to the safety monitoring of elderly people living alone.

[0150] The present application also provides a dynamic monitoring system for elderly people living alone based on a smart community, see Figure 5 The system includes: a data acquisition module 101, a monitoring plan customization module 102, a behavior association matrix construction module 103, a warning threshold update module 104, and a monitoring and warning judgment module 105.

[0151] Among them, the data acquisition module 101 is used to obtain individual portrait data of elderly people living alone.

[0152] The monitoring plan customization module 102 is used to generate customized monitoring indicators and customized warning thresholds for elderly people living alone based on individual portrait data.

[0153] The behavior association matrix construction module 103 is used to construct a behavior association matrix based on customized monitoring indicators using a preset method.

[0154] The warning threshold updating module 104 is used to obtain the monitoring data of the elderly living alone in real time and determine whether the monitoring data triggers the behavior association matrix. If the behavior association matrix is ​​triggered, the customized warning threshold is updated based on the behavior association matrix.

[0155] The monitoring and early warning judgment module 105 is used to judge whether the monitoring data reaches the customized early warning threshold. If the monitoring data reaches the customized early warning threshold, an early warning prompt message is output.

[0156] In the embodiment of the present application, the data acquisition module 101 is specifically used to obtain individual portrait data of elderly people living alone.

[0157] The monitoring plan customization module 102 is specifically used to generate customized monitoring indicators and customized warning thresholds for elderly people living alone based on the individual portrait data obtained by the data acquisition module 101.

[0158] The behavior association matrix construction module 103 is specifically used to construct a behavior association matrix through a preset method based on the customized monitoring indicators generated by the monitoring plan customization module 102;

[0159] The warning threshold updating module 104 is specifically used to obtain the monitoring data of the elderly living alone in real time based on the behavior association matrix constructed by the behavior association matrix construction module 103, and to determine whether the monitoring data triggers the behavior association matrix. If the behavior association matrix is ​​triggered, the customized warning threshold is updated based on the behavior association matrix.

[0160] The monitoring and early warning judgment module 105 is specifically used to judge whether the monitoring data reaches the customized early warning threshold. If the monitoring data reaches the customized early warning threshold, an early warning prompt message is output.

[0161] An embodiment of the present application also provides a computer-readable storage medium storing a computer program that can be loaded by a processor and execute any of the above-mentioned methods for dynamic monitoring of elderly people living alone based on a smart community.

[0162] The embodiments of this specific implementation method are all preferred embodiments of the present application and are not intended to limit the scope of protection of the present application. Therefore, all equivalent changes made based on the principles of the present application should be included in the scope of protection of the present application.

Claims

1. A method for dynamic monitoring of elderly people living alone based on a smart community, characterized in that: include: Obtain individual portrait data of elderly people living alone; Generate customized monitoring indicators and customized warning thresholds for elderly people living alone based on individual portrait data; Based on customized monitoring indicators, a behavioral correlation matrix is ​​constructed through a preset method; The behavior association matrix is ​​constructed based on customized monitoring indicators through a preset method, including: Based on individual portrait data, obtain a set of behavior tags; Traverse the behavior tag set and obtain the correlation coefficient of the behavior tag to each monitoring indicator through the preset method; After the traversal is completed, a behavior correlation matrix is ​​generated based on the correlation coefficients of all behavior labels and various monitoring indicators; Obtain monitoring data of elderly people living alone in real time and determine whether the monitoring data triggers the behavioral association matrix; If so, the customized warning threshold is updated based on the behavior correlation matrix; Determine whether the monitoring data reaches the customized warning threshold; If the monitoring data reaches the customized warning threshold, the warning prompt information will be output; The monitoring data includes behavioral data, and determining whether the monitoring data triggers a correlation matrix includes: Based on the current behavior data, obtain the current behavior label; Determine whether the current behavior label matches the behavior association matrix; If not, the current behavior tag is recorded as a new behavior tag; If so, then determine the associated monitoring indicators based on the behavior tags; Determine whether the correlation coefficient of the associated monitoring indicator is greater than a preset correlation threshold; If the correlation coefficient of the associated monitoring indicator is greater than the preset correlation threshold, a trigger response message is output.

2. The method for dynamic monitoring of elderly people living alone based on a smart community according to claim 1, characterized in that: After recording the current behavior tag as a new behavior tag, the method further includes: Obtain monitoring data when the new behavior occurs and record it as control monitoring data; Compare the control monitoring data with the monitoring data under normal conditions and calculate the monitoring indicator deviation; Determine whether the monitoring indicator deviation reaches the preset indicator deviation threshold; If so, the behavior association matrix is ​​updated based on the monitoring indicator deviation.

3. The method for dynamic monitoring of elderly people living alone based on a smart community according to claim 1, characterized in that: The updating of customized warning thresholds based on the behavior association matrix includes: Based on the behavior association matrix, determine the correlation coefficient of the current behavior label to the associated monitoring indicator; Based on the correlation coefficient, calculate and obtain the expected value of the indicator through the preset calculation rules; Update customized warning thresholds based on expected indicator values.

4. The method for dynamic monitoring of elderly people living alone based on a smart community according to claim 1, characterized in that: After outputting the warning prompt information, the method further includes: Push the warning information to the manual verification department for verification and obtain the verification feedback results; If the verification feedback result is a valid warning, the monitoring data associated with the warning will be added to the preset warning database; If the amount of data in the early warning database reaches the preset threshold, a risk prediction model is constructed based on the data in the early warning database through big data analysis and modeling; If the verification feedback result is a false alarm, the customized warning threshold is updated based on the associated monitoring data.

5. The method for dynamic monitoring of elderly people living alone based on a smart community according to claim 1, characterized in that: After obtaining the monitoring data of the elderly living alone in real time, the method further includes: Set the time period, perform data statistics on the historical monitoring data in units of time periods, and obtain the monitoring data statistics results within each time period; Taking two adjacent time periods as a group, calculate the monitoring data deviation under the two time periods; Based on the deviation of monitoring data, identify the different behavior patterns under two time periods; Based on the monitoring data deviation, the impact factors of different behavior patterns on related monitoring indicators are obtained; Update customized warning thresholds based on the impact factors of distinguishing behavioral patterns on related monitoring indicators.

6. The method for dynamic monitoring of elderly people living alone based on a smart community according to claim 1, characterized in that: The individual portrait data includes residence information. After obtaining the individual portrait data of the elderly living alone, the following steps are included: Based on the residential information of elderly people living alone, a guardianship network distribution map is constructed; Based on the monitoring network distribution map, set up associated community monitoring points.

7. A dynamic monitoring system for elderly people living alone based on a smart community, characterized by: include: A data acquisition module (101) is used to obtain individual portrait data of elderly people living alone; A monitoring plan customization module (102) is used to generate customized monitoring indicators and customized warning thresholds for elderly people living alone based on individual portrait data; The behavior association matrix construction module (103) is used to construct a behavior association matrix based on customized monitoring indicators by a preset method, wherein the behavior association matrix is ​​constructed based on customized monitoring indicators by a preset method, including: Based on individual portrait data, obtain a set of behavior tags; Traverse the behavior tag set and obtain the correlation coefficient of the behavior tag to each monitoring indicator through the preset method; After the traversal is completed, a behavior correlation matrix is ​​generated based on the correlation coefficients of all behavior labels and various monitoring indicators; The warning threshold updating module (104) is used to obtain monitoring data of elderly people living alone in real time and determine whether the monitoring data triggers a behavior association matrix. If so, the customized warning threshold is updated based on the behavior association matrix. The monitoring data includes behavior data. The determination of whether the monitoring data triggers the association matrix includes: Based on the current behavior data, obtain the current behavior label; Determine whether the current behavior label matches the behavior association matrix; If not, the current behavior tag is recorded as a new behavior tag; If so, then determine the associated monitoring indicators based on the behavior tags; Determine whether the correlation coefficient of the associated monitoring indicator is greater than a preset correlation threshold; If the correlation coefficient of the associated monitoring indicator is greater than the preset correlation threshold, a trigger response message is output; The monitoring and early warning judgment module (105) is used to judge whether the monitoring data reaches the customized early warning threshold, and output early warning prompt information if the monitoring data reaches the customized early warning threshold.

8. A computer-readable storage medium storing a computer program that can be loaded by a processor and executed by a method for dynamic monitoring of elderly people living alone in a smart community based on any one of claims 1 to 6.

Images