CN117197998A - Sensor integrated nursing system of thing networking - Google Patents

Abstract

The invention discloses a sensor integrated nursing system of the Internet of things, which comprises: the data acquisition unit is used for acquiring sensing data in real time through the sensor network of the Internet of things, wherein the sensing data comprises current environmental factors, the life state and the behavior gesture of a monitored person; the data analysis unit is used for carrying out real-time scene recognition and analysis on the sensing data based on the guardian nursing judgment model to obtain an analysis result; and the early warning response unit is used for immediately triggering an early warning mechanism when the analysis result shows abnormality, notifying a caretaker in real time and simultaneously automatically starting corresponding safety measures in the guardian. The sensing data is acquired in real time through the sensor network of the Internet of things, so that the life state and environmental factors of monitored personnel can be known in time, and the real-time performance and accuracy of monitoring are improved.

Description

Sensor integrated nursing system of thing networking

Technical Field

The invention relates to the technical field of sensors, in particular to a sensor integrated nursing system of the Internet of things.

Background

Current care systems are mainly based on traditional monitoring technologies such as Closed Circuit Television (CCTV) cameras, infrared sensors, door and window sensors, etc. These systems are mainly used for monitoring and alerting, but their functions are relatively single.

The application number is: the invention of CN202110904260 discloses a home care system for old people based on monitoring of the Internet of things, wherein an environment monitoring module is used for carrying out environment analysis on a care site of a cared person to obtain a ring shadow value, the environment monitoring module feeds back the ring shadow value to a server, the server sends the ring shadow value to a body condition monitoring module along with real-time body data, the body condition monitoring module receives the ring shadow value of the care site of the cared person and the real-time body data of the cared person sent by the server, and the body condition monitoring module is used for carrying out body condition monitoring analysis on the body condition data of the cared person and sending a care early warning signal to the server; the server receives the nursing early warning signal and then sends the nursing early warning signal to the user terminal. The defects include: multiple data transmissions (from the environmental monitoring module to the server, back to the body monitoring module, back to the server, and finally to the user terminal) can lead to delays in system response, which can affect the safety of caregivers in emergency situations; the transfer of data between multiple modules and a server may increase the risk of interception or tampering of the data; the interaction between multiple modules and servers increases the complexity of the system, which may lead to maintenance difficulties and higher error rates; if the server fails, the entire system may be affected, resulting in failure of the monitoring and early warning functions.

Therefore, there is an urgent need for a sensor integrated care system for the internet of things.

Disclosure of Invention

The invention provides a sensor integrated nursing system of the Internet of things, which aims to solve the problem that the system response delay is caused by multiple data transmission (from an environment monitoring module to a server, then to a body condition monitoring module, then to the server and finally to a user terminal) in the prior art, so that the safety of a nursed person can be influenced under emergency conditions; the transfer of data between multiple modules and a server may increase the risk of interception or tampering of the data; the interaction between multiple modules and servers increases the complexity of the system, which may lead to maintenance difficulties and higher error rates; if the server fails, the entire system may be affected, resulting in the above-mentioned problem of failure of the monitoring and early warning functions.

In order to achieve the above purpose, the present invention provides the following technical solutions:

a sensor integrated care system of the internet of things, comprising:

the data acquisition unit is used for acquiring sensing data in real time through the sensor network of the Internet of things, wherein the sensing data comprises current environmental factors, the life state and the behavior gesture of a monitored person;

The data analysis unit is used for carrying out real-time scene recognition and analysis on the sensing data based on the guardian nursing judgment model to obtain an analysis result;

and the early warning response unit is used for immediately triggering an early warning mechanism when the analysis result shows abnormality, notifying a caretaker in real time and simultaneously automatically starting corresponding safety measures in the guardian.

Wherein the data acquisition unit includes:

the self-adaptive configuration module is used for automatically configuring the positions and the number of the sensors of the Internet of things according to the size, the layout and the object distribution of the indoor space;

the environment sensing module is used for acquiring corresponding environment data based on a plurality of environment sensors according to real-time environment factors, adaptively adjusting the sensitivity and the data acquisition frequency of the sensors, and immediately adjusting corresponding sensor parameters when the environment factors are suddenly changed so as to adapt to new environment conditions;

and the vital state and behavior gesture sensing module is used for capturing vital sign changes and behavior gestures of the monitored person in real time based on the vital sign monitor and the behavior gesture recognition sensor.

Wherein the data analysis unit includes:

the nursing judging module is used for dynamically constructing and optimizing a nursing judging model according to the historical data, the physiological characteristics and the daily habits of the monitored person;

The scene recognition module is used for carrying out feature extraction and situation classification on the sensing data based on a deep learning technology, and carrying out real-time situation annotation on the classified sensing data, wherein the situation annotation comprises cooking, reading and resting;

the comprehensive evaluation module is used for comprehensively evaluating the current state of the monitored person based on the scene recognition result and the nursing judgment model to obtain a comprehensive evaluation result;

the trend prediction module is used for carrying out trend analysis on the behavior and physiological states of the guardian based on a time sequence analysis technology, and predicting the corresponding states and requirements of the guardian in the future through the trend analysis.

Wherein, early warning response unit includes:

the early warning triggering module is used for immediately triggering an early warning mechanism when an abnormal condition is detected, wherein the abnormal condition comprises sudden physiological sign change or unusual behavior mode;

the multi-channel notification module is used for simultaneously sending early warning information through a plurality of channels based on an information redundancy mechanism and notifying a caretaker in real time, wherein the channels comprise short messages, telephones and application pushing;

the safety measure starting module is used for automatically selecting and starting corresponding safety measures according to the type and severity of the abnormal condition, and corresponding safety measure equipment comprises emergency light, an automatic door lock and an oxygen supply system.

Wherein, life state and action gesture perception module includes: a vital sign monitor and a behavioral gesture recognition sensor;

a vital sign monitor for capturing key vital signs in real time, the key vital signs including heart rate, blood oxygen level and body temperature;

the behavior gesture recognition sensor is used for recognizing the behaviors and gestures of the monitored person in real time, wherein the behaviors and gestures comprise standing, sitting, walking or falling.

Wherein, the scene recognition module includes: a deep learning classifier;

based on the cyclic neural network, carrying out feature extraction on the sensing data, wherein the feature extraction process automatically identifies and extracts key modes and structures in the sensing data so as to facilitate subsequent situation classification; based on the extracted features, the sensing data is subjected to situation classification through a deep learning classifier, the sensing data is classified into different situations by the deep learning classifier, the situations comprise cooking, reading and resting, after the data is classified, the classified data is subjected to situation annotation in real time, and each section of sensing data is ensured to be matched with the corresponding situation.

Wherein, the security measure start module includes: an adaptive learning module;

continuously monitoring the incoming data to identify any abnormal situation in real time, automatically classifying the type and severity of the abnormal situation based on an abnormality classification model, the type including falls and asphyxia, the severity including mild, moderate and severe; according to the type and severity of the abnormal situation, corresponding safety measures are automatically selected, emergency light is started to remind surrounding people of falling event, and an oxygen supply system is started for suffocation or dyspnea; the self-adaptive learning module automatically adjusts and optimizes the selection and starting strategy of the safety measures according to the historical data and the user feedback.

Initializing a main control board level in the vital sign monitor, abstracting and packaging hardware in the vital sign monitor, and providing an interface for a node program of the vital sign monitor; starting an application task to execute a core function task of the vital sign monitor node, and creating a corresponding client when the application task is started, wherein the client sends a registration request through the service of the edge intelligent gateway and establishes connection with the edge intelligent gateway; entering a vital sign monitor node main circulation program, judging whether a counter value reaches a threshold value, if so, reading body temperature, acceleration and heart rate data through an interface, calculating a body temperature value, human activity intensity and heart rate value, updating data and a flag bit, and sending the data to an edge intelligent gateway; if the counter value does not reach the threshold value, the counter value is incremented by 1.

Wherein, carry out the situation classification through the deep learning classifier to the sensing data, include:

constructing a feature extraction information base corresponding to the sensing data;

acquiring a plurality of preset feature extraction nodes, and traversing each feature extraction node in sequence;

each time of traversing, acquiring a preset feature extraction template and a preset verification template corresponding to the node type of the traversed feature extraction node;

Extracting required characteristic data from the sensing data based on the characteristic extraction template;

based on the verification template, performing feature verification according to the extracted feature data;

when the feature verification is passed, a preset training template of the deep learning classifier is obtained;

training by using the marked sensing data based on a training template of the deep learning classifier to obtain the deep learning classifier;

based on the deep learning classifier, carrying out situation classification on the extracted characteristic data;

classifying the sensing data into different situations according to the classification result, wherein the situations comprise cooking, reading and resting;

storing the classification result in a situation classification result library;

repeating the steps when new sensing data is input until all the sensing data are classified;

the situation classification result is updated in real time by comparing the new sensing data with the data in the situation classification result library so as to ensure the accuracy and real-time performance of the situation classification;

after determining the context, executing corresponding operations or reminders according to different contexts, wherein the corresponding operations or reminders comprise: under the cooking situation, the user is reminded to pay attention to the safety of the fire source, under the reading situation, indoor illumination is adjusted to protect eyesight, and under the rest situation, indoor temperature is adjusted to ensure comfort.

Wherein, according to the type and severity of the abnormal situation, the corresponding security measures are automatically selected, including:

acquiring abnormal condition data of a user;

based on a preset abnormal feature extraction template, carrying out feature extraction on abnormal condition data to obtain an abnormal feature set;

extracting a plurality of groups of standard abnormality feature sets and abnormality types and severity which are in one-to-one correspondence from a preset abnormality classification experience library;

matching the abnormal feature set with any standard abnormal feature set to obtain an abnormal matching degree;

determining the type and severity of the abnormal condition based on the degree of abnormal matching;

types include falls and asphyxia, and severity includes mild, moderate, and severe, among others;

based on a preset abnormal severity classification standard, classifying severity grades of abnormal conditions;

extracting a corresponding safety measure scheme from a preset safety measure experience library according to the determined abnormality type and severity level;

based on the security measure scheme, corresponding security measures are automatically selected and executed to ensure the security of the user.

Compared with the prior art, the invention has the following advantages:

a sensor integrated care system of the internet of things, comprising: the data acquisition unit is used for acquiring sensing data in real time through the sensor network of the Internet of things, wherein the sensing data comprises current environmental factors, the life state and the behavior gesture of a monitored person; the data analysis unit is used for carrying out real-time scene recognition and analysis on the sensing data based on the guardian nursing judgment model to obtain an analysis result; and the early warning response unit is used for immediately triggering an early warning mechanism when the analysis result shows abnormality, notifying a caretaker in real time and simultaneously automatically starting corresponding safety measures in the guardian. The sensing data is acquired in real time through the sensor network of the Internet of things, so that the life state and environmental factors of monitored personnel can be known in time, and the real-time performance and accuracy of monitoring are improved.

Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention.

The technical scheme of the invention is further described in detail through the drawings and the embodiments.

Drawings

The accompanying drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate the invention and together with the embodiments of the invention, serve to explain the invention. In the drawings:

FIG. 1 is a block diagram of an integrated sensor care system for Internet of things in an embodiment of the invention;

FIG. 2 is a block diagram of a data acquisition unit in an embodiment of the invention;

fig. 3 is a block diagram of a data analysis unit according to an embodiment of the present invention.

Detailed Description

The preferred embodiments of the present invention will be described below with reference to the accompanying drawings, it being understood that the preferred embodiments described herein are for illustration and explanation of the present invention only, and are not intended to limit the present invention.

The embodiment of the invention provides a sensor integrated nursing system of the Internet of things, which comprises:

the data acquisition unit is used for acquiring sensing data in real time through the sensor network of the Internet of things, wherein the sensing data comprises current environmental factors, the life state and the behavior gesture of a monitored person;

The data analysis unit is used for carrying out real-time scene recognition and analysis on the sensing data based on the guardian nursing judgment model to obtain an analysis result;

and the early warning response unit is used for immediately triggering an early warning mechanism when the analysis result shows abnormality, notifying a caretaker in real time and simultaneously automatically starting corresponding safety measures in the guardian.

The working principle of the technical scheme is as follows: the data acquisition unit acquires sensing data of environmental factors, life states and behavior postures of guardianship personnel through an Internet of things sensor network, and the sensors comprise a temperature sensor, a humidity sensor, a heart rate sensor and a motion sensor, and selects corresponding sensors according to requirements of the guardianship personnel; transmitting the sensing data to a data analysis unit through a wireless communication technology, and storing the sensing data in real time so as to facilitate subsequent analysis and processing; a data analysis unit: based on a guardian nursing judging model, carrying out real-time scene recognition and analysis on the sensing data, wherein the scene recognition is to classify the sensing data through a machine learning algorithm, judge whether the current environment state and the life state and behavior posture of the guardian are normal or not and recognize whether the guardian is interacting with other people or not and whether specific activities (such as cooking and reading) are being carried out or not; judging whether an abnormal condition exists according to the analysis result, triggering an early warning mechanism and immediately informing a caretaker if the analysis result shows the abnormal condition, wherein the early warning mechanism can be informed by means of a mobile phone short message, a telephone call, APP pushing and terminal alarming, and simultaneously, corresponding safety measures in a guardian room are automatically started, such as starting an alarm, calling an emergency medical rescue system, adjusting indoor temperature and the like, so that the safety and health of the guardian can be ensured.

The beneficial effects of the technical scheme are as follows: the sensing data is acquired in real time through the sensor network of the Internet of things, so that the life state and environmental factors of monitored personnel can be known in time, and the real-time performance and accuracy of monitoring are improved; the sensing data is identified and analyzed in real time through the data analysis unit, so that abnormal conditions can be found in time, and measures are taken in advance to intervene and process; when the analysis result shows abnormality, an early warning mechanism is immediately triggered, a caretaker is notified, and safety measures are automatically started, so that emergency can be timely dealt with, and the safety and health of the guarded personnel are ensured; by automated data acquisition and analysis, the workload of caregivers is reduced, the care efficiency and quality are improved, and more accurate monitoring results and advice can be provided.

In another embodiment, the data acquisition unit includes:

the self-adaptive configuration module is used for automatically configuring the positions and the number of the sensors of the Internet of things according to the size, the layout and the object distribution of the indoor space;

the environment sensing module is used for acquiring corresponding environment data based on a plurality of environment sensors according to real-time environment factors, adaptively adjusting the sensitivity and the data acquisition frequency of the sensors, and immediately adjusting corresponding sensor parameters when the environment factors are suddenly changed so as to adapt to new environment conditions;

And the vital state and behavior gesture sensing module is used for capturing vital sign changes and behavior gestures of the monitored person in real time based on the vital sign monitor and the behavior gesture recognition sensor.

The working principle of the technical scheme is as follows: the self-adaptive configuration module automatically configures the positions and the quantity of the sensors of the Internet of things according to the size, the layout and the object distribution of the indoor space through an algorithm and a rule, the module can determine the optimal sensor positions through the layout information of an indoor map or a sensor network and combining the technologies such as object detection, distance measurement and the like, the quantity of the sensors is automatically adjusted according to the requirements, and the self-adaptive configuration can ensure the uniformity of the coverage of the sensors, avoid blind areas and overlapping areas and improve the accuracy and the efficiency of monitoring;

the environment sensing module senses environment factors by collecting environment data in real time based on a plurality of environment sensors, and adaptively adjusts the sensitivity and data collection frequency of the sensors according to the real-time data;

The vital state and behavior gesture sensing module is based on a vital sign monitor and a behavior gesture recognition sensor, captures vital sign changes and behavior gestures of monitored personnel in real time, can acquire vital sign data of the monitored personnel through vital sign monitors such as a heart rate sensor, a blood pressure sensor and a respiratory sensor, acquires behavior gesture data of the monitored personnel through behavior gesture recognition sensors such as a camera and an acceleration sensor, and then analyzes and processes the data through an algorithm and a model to realize sensing and monitoring of the vital state and the behavior gesture of the monitored personnel;

the operation modes of the modules are realized by software or hardware; in the aspect of software, the functions of self-adaptive configuration, environment perception and life state and behavior gesture perception are realized through programming and algorithm design; in terms of hardware, the operation and control of the sensors are realized through the arrangement and connection of the sensors and corresponding data acquisition and processing equipment.

The method for capturing the vital sign change and the behavior gesture of the monitored person in real time comprises the following steps of:

acquiring initial vital sign data of a monitored person;

capturing vital sign data of heart rate, blood oxygen saturation and respiratory frequency of the monitored person in real time through at least one vital sign monitor arranged in the environment of the monitored person;

Based on a preset first feature extraction template, carrying out feature extraction on vital sign data to obtain a vital sign feature set;

the method comprises the steps that through at least one behavior gesture recognition sensor arranged in an environment where a monitored person is located, behavior gesture data of walking, sitting and lying and falling of the monitored person are captured in real time;

based on a preset second feature extraction template, feature extraction is carried out on the behavior gesture data, and a behavior gesture feature set is obtained;

matching the vital sign feature set with a preset standard vital sign feature set to obtain a first matching degree;

matching the behavior gesture feature set with a preset standard behavior gesture feature set to obtain a second matching degree;

based on the first matching degree and the second matching degree, the health state and the safety risk of the monitored personnel are evaluated in real time;

and when the first matching degree or the second matching degree exceeds a preset safety threshold, notifying a guardian or a related medical institution in a preset notification mode.

The beneficial effects of the technical scheme are as follows: the intelligent and automatic level of the system is improved, and the manual intervention and management cost is reduced; the accuracy and the efficiency of the layout and the configuration of the sensor are improved, and the monitoring and controlling effects are optimized; sensing and monitoring environmental factors, life states and behavior postures in real time, and timely finding abnormal conditions and risks; according to the environmental change and individual requirements, the sensor parameters are automatically adjusted, and the accuracy and stability of data acquisition are improved.

In another embodiment, the data analysis unit comprises:

the nursing judging module is used for dynamically constructing and optimizing a nursing judging model according to the historical data, the physiological characteristics and the daily habits of the monitored person;

the scene recognition module is used for carrying out feature extraction and situation classification on the sensing data based on a deep learning technology, and carrying out real-time situation annotation on the classified sensing data, wherein the situation annotation comprises cooking, reading and resting;

the comprehensive evaluation module is used for comprehensively evaluating the current state of the monitored person based on the scene recognition result and the nursing judgment model to obtain a comprehensive evaluation result;

the trend prediction module is used for carrying out trend analysis on the behavior and physiological states of the guardian based on a time sequence analysis technology, and predicting the corresponding states and requirements of the guardian in the future through the trend analysis.

The working principle of the technical scheme is as follows: the nursing judging module dynamically builds and optimizes a nursing judging model based on historical data, physiological characteristics and daily habits of the guardian, the module can analyze and model the data of the guardian through machine learning and data mining technology, key characteristics are extracted, judgment is carried out according to the historical data and the model, and the accuracy and adaptability of the nursing judging model can be improved through continuous learning and optimization;

The scene recognition module is based on a deep learning technology, performs feature extraction and situation classification on the sensing data, analyzes and processes the sensing data through a neural network and other deep learning models, extracts key features, classifies the data into different situations, and can mark the sensing data through real-time situation labeling, such as cooking, reading, resting and the like, so that the behavior and state of a monitored person can be better understood;

the working principle of the comprehensive evaluation module is to comprehensively evaluate the current state of the monitored person based on the scene recognition result and the nursing judgment model, the module can comprehensively evaluate the state of the monitored person by integrating the scene recognition result and the output of the nursing judgment model, and the comprehensive evaluation result of the monitored person, such as health condition, emotion state and the like, can be obtained through comprehensive evaluation;

the working principle of the trend prediction module is that the trend analysis is carried out on the behavior and physiological state of the guardian based on a time sequence analysis technology, the module can model and analyze the historical data of the guardian through a time sequence analysis method, extract trend information, predict the corresponding state and the requirement of the guardian in the future through trend analysis, and can carry out corresponding adjustment and intervention in advance through trend prediction, thereby improving the nursing effect and timeliness.

Wherein, dynamically constructing and optimizing a nursing judgment model according to historical data, physiological characteristics and daily habits of a guardian, comprising:

obtaining a preset model construction strategy set, wherein the model construction strategy set comprises the following steps: a plurality of first build strategies;

splitting the first construction strategy into a plurality of first strategy items (for example, the first construction strategy is "constructing a health prediction model based on physiological characteristics", and the split first strategy items are "extracting heart rate characteristics", "extracting blood pressure characteristics", etc.);

acquiring historical data, physiological characteristics and daily habits (wherein the historical data comprises health records of past time, the physiological characteristics comprise heart rate, blood pressure and the like, and the daily habits comprise diet, exercise and the like) of a monitored person;

based on a preset first feature extraction template (the first feature extraction template defines how to extract heart rate variability, blood pressure waveform and other features from physiological features), carrying out feature extraction on historical data, physiological features and daily habits to obtain a plurality of second features;

acquiring a preset feature-dynamic weight library (the feature-dynamic weight library designates the importance of heart rate variability to a health prediction model, and the system determines the weight of heart rate variability according to information in the library), determining the dynamic weight corresponding to the second feature based on the feature-dynamic weight library, and correlating with a corresponding first strategy item;

Accumulating and calculating the dynamic weights associated with the first strategy items to obtain a weight sum (if the dynamic weights associated with the first strategy items are 0.6 and 0.4, the weight sum is 0.6+0.4=1.0);

if the weight sum is equal to or greater than the preset weight sum threshold, taking the corresponding first construction strategy as a second construction strategy (if the weight sum is 1.0 and the preset weight sum threshold is 0.8, the first construction strategy is determined to be the second construction strategy);

dynamically constructing and optimizing a nursing judgment model based on a second construction strategy;

acquiring real-time data of a monitored person;

the real-time data is input into the nursing judgment model to obtain a nursing judgment result (for example, the system acquires the current heart rate data of the monitored person and inputs the current heart rate data into the health prediction model to obtain a health prediction result).

Trend analysis is carried out on the behavior and physiological state of the guardian, and the trend analysis comprises the following steps:

constructing an information database corresponding to the behavior and physiological state of the guardian, wherein the database comprises historical behavior data and physiological parameter data of the guardian;

acquiring behavior and physiological state data of the monitored person in real time through at least one physiological and behavior data acquisition device arranged in a living or activity area of the monitored person;

Storing the data acquired in real time into an information database, and marking a corresponding time stamp to form time sequence data;

carrying out trend analysis on behavior and physiological state data of a guardian in an information database by using a time sequence analysis technology, wherein the trend analysis comprises seasonal analysis, periodic analysis and abnormal value detection;

predicting the behavior and physiological state of the monitored person within a specific time period in the future based on the trend analysis, wherein the prediction comprises the activity range, the physiological parameter change and the potential health risk of the monitored person;

comparing the prediction result with preset behavior and physiological state thresholds, and determining whether the future state and the demand of the guardian are in a normal range;

if the future state and the demand of the guardian exceed the normal range, notifying the guardian or a related medical institution through a preset notification mode, such as mobile application notification, a short message or a telephone;

based on the prediction result, the guardian or the related medical institution takes corresponding intervention measures to prevent the potential health risk of the guardian or meet the specific requirements of the guardian.

The beneficial effects of the technical scheme are as follows: the nursing accuracy and individuation degree are improved, and judgment and evaluation are carried out according to the characteristics and habits of the guardian; identifying and marking the situation of the guardian in real time, and better understanding the behavior and state of the guardian; comprehensively evaluating the current state of the guardian and providing comprehensive nursing information and advice; through trend analysis and prediction, possible problems of intervention are prevented in advance, and nursing effect and timeliness are improved; and decision support and guidance are provided for nursing staff, and the efficiency and quality of nursing work are improved.

In another embodiment, the early warning response unit includes:

the early warning triggering module is used for immediately triggering an early warning mechanism when an abnormal condition is detected, wherein the abnormal condition comprises sudden physiological sign change or unusual behavior mode;

the multi-channel notification module is used for simultaneously sending early warning information through a plurality of channels based on an information redundancy mechanism and notifying a caretaker in real time, wherein the channels comprise short messages, telephones and application pushing;

the safety measure starting module is used for automatically selecting and starting corresponding safety measures according to the type and severity of the abnormal condition, and corresponding safety measure equipment comprises emergency light, an automatic door lock and an oxygen supply system.

The working principle of the technical scheme is as follows: the working principle of the early warning triggering module is that an early warning mechanism is triggered immediately when an abnormal condition is detected. The module can be used for comparing the physiological sign and the behavior pattern of the monitored person with a preset normal range, and triggering early warning immediately once sudden physiological sign change or unusual behavior pattern is found. Through real-time monitoring and analysis, potential risks and problems can be quickly found.

The working principle of the multi-channel notification module is based on an information redundancy mechanism, early warning information is sent simultaneously through a plurality of channels, and a caretaker is notified in real time. The module can send the early warning information to a caretaker through a plurality of communication channels such as short messages, telephones, application pushing and the like. The method of multi-channel notification can improve the reliability and timeliness of information transmission, and ensure that a caretaker can receive early warning information in time.

The working principle of the safety measure starting module is that corresponding safety measures are automatically selected and started according to the type and severity of abnormal conditions. The module can classify and evaluate abnormal conditions according to preset rules and strategies, and then select and start corresponding safety measures according to evaluation results. For example, when a serious change in physiological signs is found, emergency lights, automatic door locks, etc. can be activated to ensure the safety of the person under guardianship.

The beneficial effects of the technical scheme are as follows: monitoring and early warning abnormal conditions in real time, and timely finding potential risks and problems; through multi-channel notification, the caretaker can be ensured to receive early warning information in time, and the timeliness and accuracy of the response are improved; according to the severity of the abnormal condition, corresponding safety measures are automatically selected and started, so that the safety of a guardian is ensured; the efficiency and the quality of nursing work are improved, and the artificial omission and delay are reduced; the safety and the confidence of the nursing staff are enhanced, and the reliability and the satisfaction degree of the nursing service are improved.

In another embodiment, the life state and behavioral gesture sensing module includes: a vital sign monitor and a behavioral gesture recognition sensor;

A vital sign monitor for capturing key vital signs in real time, the key vital signs including heart rate, blood oxygen level and body temperature;

the behavior gesture recognition sensor is used for recognizing the behaviors and gestures of the monitored person in real time, wherein the behaviors and gestures comprise standing, sitting, walking or falling.

The working principle of the technical scheme is as follows: vital sign monitors capture key vital signs of a monitored person, including heart rate, blood oxygen level and body temperature, in real time via sensors. These sensors may acquire relevant data in a non-invasive or invasive manner; the heart rate sensor usually adopts a photoelectric sensing technology, detects the heart rate through an infrared light source and a photoelectric sensor, when light irradiates the skin, a part of the light is absorbed, a part of the light is reflected, and when the heart beats, the blood flow can cause weak change of the skin color, so that the intensity of the reflected light is changed, and the heart rate can be calculated by detecting the change of the light intensity; blood oxygen level sensors usually adopt the photoelectric measurement principle, the oxygen saturation of blood is measured through the absorption characteristics of infrared light and red light, the sensors are usually placed at the positions of fingertips, earlobes and the like, and the oxygen content in blood is measured through the transmission and reflection of light; the body temperature sensor can measure the body temperature in a contact or non-contact mode. Contact body temperature sensors typically measure body temperature by the temperature of the skin surface, while non-contact body temperature sensors measure body temperature by infrared radiation.

The behavior gesture recognition sensor is used for recognizing the behaviors and gestures of the monitored person in real time through sensors such as an accelerometer, a gyroscope and the like, and the sensors can detect the acceleration and angular velocity changes of the body, so that whether the monitored person stands, sits down, walks or falls can be judged.

The beneficial effects of the technical scheme are as follows: the monitor can monitor vital signs such as heart rate, blood oxygen level and body temperature of a monitored person in real time, and once abnormal conditions such as too high or too low heart rate, too low blood oxygen level or abnormal rise of body temperature occur, the monitored person can take measures in time to intervene and treat; the behavior gesture recognition sensor can monitor the behaviors and gestures of guardianship personnel in real time, and once accidents such as falling are found, the guardianship personnel can provide assistance in time, so that the occurrence of accidental injury is reduced; through monitoring and analyzing key vital signs and behavior postures, guardianship personnel can know the health condition and daily activity condition of guardianship personnel, so that personalized nursing and health management advice is provided; the monitor can record and store vital sign and behavior data of the monitored person, and the monitored person can know the health condition and behavior habit of the monitored person by analyzing the data, so that a reference basis is provided for health management and disease prevention.

In another embodiment, the context identification module comprises:

based on the cyclic neural network, carrying out feature extraction on the sensing data, wherein the feature extraction process automatically identifies and extracts key modes and structures in the sensing data so as to facilitate subsequent situation classification; based on the extracted features, the sensing data is subjected to situation classification through a deep learning classifier, the sensing data is classified into different situations by the deep learning classifier, the situations comprise cooking, reading and resting, after the data is classified, the classified data is subjected to situation annotation in real time, and each section of sensing data is ensured to be matched with the corresponding situation.

The working principle of the technical scheme is as follows: first, a recurrent neural network is used to perform feature extraction on the sensing data. The RNN has a memory function, can process sequence data and capture time dependence in the data, and key modes and structures in the sensing data can be automatically identified and extracted through the RNN, and the characteristics can be helpful for subsequent situation classification; next, based on the extracted features, classifying the sensing data by using a deep learning classifier, wherein the deep learning classifier can learn the feature representation under different situations by training a large amount of data, so as to classify the sensing data into different situations, such as cooking, reading, resting and the like; finally, the classified data are subjected to context labeling in real time, and each section of sensing data is ensured to be matched with the corresponding context, so that a guardian can know the behavior and the gesture of the guardian in real time, discover the abnormal situation in time and take corresponding measures.

The beneficial effects of the technical scheme are as follows: by adopting a circulating neural network and a deep learning classifier, the time dependence and complex characteristics in the sensing data can be fully utilized, and the accuracy and the efficiency of situation classification are improved; the method can realize real-time monitoring and personalized nursing, help guardianship personnel to better know behavior habits and health conditions of guardianship personnel, and provide accurate health management and disease prevention suggestions.

In another embodiment, the security measure initiation module comprises: an adaptive learning module;

continuously monitoring the incoming data to identify any abnormal situation in real time, automatically classifying the type and severity of the abnormal situation based on an abnormality classification model, the type including falls and asphyxia, the severity including mild, moderate and severe; according to the type and severity of the abnormal situation, corresponding safety measures are automatically selected, emergency light is started to remind surrounding people of falling event, and an oxygen supply system is started for suffocation or dyspnea; the self-adaptive learning module automatically adjusts and optimizes the selection and starting strategy of the safety measures according to the historical data and the user feedback.

The working principle of the technical scheme is as follows: first, the incoming data is continuously monitored, and any anomalies are identified in real time by an anomaly classification model. The abnormal classification model learns the characteristic representation of different abnormal conditions, such as falling and asphyxia, by training a large amount of data, and automatically classifies the type and severity of the abnormal condition when the abnormal condition is detected; depending on the type and severity of the anomaly, the system will automatically select the corresponding safety measures, for example, for a fall event, the system may activate emergency lights to alert surrounding persons, help provide timely assistance, for choking or dyspnea, the system may activate an oxygen supply system to ensure that the monitored person breathes smoothly; meanwhile, the system also comprises a self-adaptive learning module, the selection and starting strategy of the safety measures are automatically adjusted and optimized according to the historical data and the user feedback, the system can continuously improve an abnormal classification model through learning and analyzing the historical data, the accuracy and the reliability of abnormal conditions are improved, and the user feedback can help the system to better understand and meet the requirements of guardianship personnel, so that more personalized and effective safety measures are provided.

The beneficial effects of the technical scheme are as follows: by adopting an abnormal classification model and a self-adaptive learning module, the abnormal condition can be identified and processed in real time, and timely safety guarantee is provided; the system can select proper safety measures according to different abnormal conditions, so that guardianship personnel and guardianship personnel can be helped to cope with emergency conditions, and accidents and injuries are reduced; meanwhile, through continuous optimization of the self-adaptive learning module, the system can provide more intelligent and personalized safety protection, and the life quality and safety of guardianship personnel are improved.

In another embodiment, the main control board level in the vital sign monitor is initialized, hardware in the vital sign monitor is abstracted and packaged, and an interface is provided for the node program of the vital sign monitor; starting an application task to execute a core function task of the vital sign monitor node, and creating a corresponding client when the application task is started, wherein the client sends a registration request through the service of the edge intelligent gateway and establishes connection with the edge intelligent gateway; entering a vital sign monitor node main circulation program, judging whether a counter value reaches a threshold value, if so, reading body temperature, acceleration and heart rate data through an interface, calculating a body temperature value, human activity intensity and heart rate value, updating data and a flag bit, and sending the data to an edge intelligent gateway; if the counter value does not reach the threshold value, the counter value is incremented by 1.

The working principle of the technical scheme is as follows: firstly, the main control board level initialises to abstract and package the hardware in the vital sign monitor, and provides an interface for the node program, so that the hardware details can be hidden, the development and maintenance of the node program are simplified, and the expandability and maintainability of the system are improved; then, starting an application task to execute a core function task of the vital sign monitor node, creating a corresponding client when the application task is started, sending a registration request through the service of the edge intelligent gateway, and establishing connection with the edge intelligent gateway, so that communication with the edge intelligent gateway can be realized, and the monitored data is transmitted to the edge intelligent gateway for further processing and analysis; after entering a main circulation program of the vital sign monitor node, judging whether the counter value reaches a threshold value, if so, reading body temperature, acceleration and heart rate data through an interface, performing corresponding calculation, such as calculating the body temperature value, the activity intensity of a human body and the heart rate value, and finally updating the data and the zone bit and sending the data to an edge intelligent gateway; if the counter value does not reach the threshold value, the counter value is increased by 1, and the next cycle is continued to be waited.

The basic functions of the vital sign monitor can be realized by adopting the main control board level initialization and node program mode. The main control board level initialization abstracts and encapsulates the hardware, and simplifies the development and maintenance of the node program. The node program realizes the transmission and processing of data through the communication with the edge intelligent gateway. By running the circulation program, vital sign data can be monitored in real time and transmitted to the edge intelligent gateway for further analysis and application.

The beneficial effects of the technical scheme are as follows: the main control board level initialization abstracts and encapsulates the hardware, so that the development and maintenance of the node program are simpler and more efficient; by running a circulation program, vital sign data can be monitored in real time, and abnormal conditions can be found in time; through communication with the edge intelligent gateway, data transmission and processing are realized, and a foundation is provided for subsequent analysis and application; by adopting the mode of main control board level initialization and node program, the system has good expandability and can conveniently add new functions and modules.

In another embodiment, contextually classifying the sensor data by the deep learning classifier includes:

Constructing a feature extraction information base corresponding to the sensing data;

acquiring a plurality of preset feature extraction nodes, and traversing each feature extraction node in sequence;

each time of traversing, acquiring a preset feature extraction template and a preset verification template corresponding to the node type of the traversed feature extraction node;

extracting required characteristic data from the sensing data based on the characteristic extraction template;

based on the verification template, performing feature verification according to the extracted feature data;

when the feature verification is passed, a preset training template of the deep learning classifier is obtained;

training by using the marked sensing data based on a training template of the deep learning classifier to obtain the deep learning classifier;

based on the deep learning classifier, carrying out situation classification on the extracted characteristic data;

classifying the sensing data into different situations according to the classification result, wherein the situations comprise cooking, reading and resting;

storing the classification result in a situation classification result library;

repeating the steps when new sensing data is input until all the sensing data are classified;

the situation classification result is updated in real time by comparing the new sensing data with the data in the situation classification result library so as to ensure the accuracy and real-time performance of the situation classification;

After determining the context, executing corresponding operations or reminders according to different contexts, wherein the corresponding operations or reminders comprise: under the cooking situation, the user is reminded to pay attention to the safety of the fire source, under the reading situation, indoor illumination is adjusted to protect eyesight, and under the rest situation, indoor temperature is adjusted to ensure comfort.

The working principle of the technical scheme is as follows: firstly, an information base is required to be constructed, wherein the information base contains characteristic extraction information related to the sensing data, and the characteristic extraction information comprises useful characteristics in various sensing data and corresponding processing methods; the system defines a plurality of feature extraction nodes in advance, each node is used for extracting specific feature information, and then the system traverses the feature extraction nodes in sequence (the feature extraction nodes can comprise a temperature sensor, a humidity sensor, an illumination sensor and the like, and each node is responsible for extracting the features of corresponding sensing data);

the system extracts the required feature data from the sensed data according to a preset feature extraction template (if one feature extraction node is a temperature sensor, the feature extraction template specifies how to extract temperature information from the sensed data) each time the feature extraction node is traversed;

Based on a preset verification template, the system verifies the extracted characteristic data to ensure the quality and accuracy of the data (if the characteristic extraction node extracts temperature data, the verification template can comprise a verification step of checking whether the data is within a reasonable range);

once the feature data is validated, the system obtains a preset deep learning classifier training template and trains with the labeled sensing data to obtain a deep learning classifier (the deep learning classifier is a neural network for mapping the extracted feature data to different context categories);

the extracted characteristic data is subjected to situation classification by using a trained deep learning classifier, and the sensing data is classified into different situations, such as cooking, reading, resting and the like (the system classifies the sensing data into a 'cooking' situation according to the characteristics of the sensing data);

storing the classification result in a situation classification result library for subsequent use and reference; when new sensing data is input, the system repeats the steps, and the situation classification result is updated in real time so as to ensure the accuracy and the instantaneity of classification; depending on the determined context, the system performs the corresponding operation or alerts the user to take the appropriate action (if the context is classified as "cook," the system may alert the user to the safety of the fire source).

The beneficial effects of the technical scheme are as follows: real-time classification of the sensing data can be realized, and the system is helped to know the behavior and the environment situation of the user in real time so as to take corresponding measures; through classification of different situations, the system can provide personalized reminding and operation, and user experience and life quality are improved; through the feature verification step, the extracted feature data can be ensured to have high quality and accuracy, and the reliability of classification and prediction is improved; the context classification result library may be a useful resource for subsequent data analysis and improved system performance.

In another embodiment, the corresponding security measures will be automatically selected according to the type and severity of the abnormal situation, including:

acquiring abnormal condition data of a user;

based on a preset abnormal feature extraction template, carrying out feature extraction on abnormal condition data to obtain an abnormal feature set;

extracting a plurality of groups of standard abnormality feature sets and abnormality types and severity which are in one-to-one correspondence from a preset abnormality classification experience library;

matching the abnormal feature set with any standard abnormal feature set to obtain an abnormal matching degree;

determining the type and severity of the abnormal condition based on the degree of abnormal matching;

Types include falls and asphyxia, and severity includes mild, moderate, and severe, among others;

based on a preset abnormal severity classification standard, classifying severity grades of abnormal conditions;

extracting a corresponding safety measure scheme from a preset safety measure experience library according to the determined abnormality type and severity level;

based on the security measure scheme, corresponding security measures are automatically selected and executed to ensure the security of the user.

The working principle of the technical scheme is as follows: firstly, acquiring abnormal condition data of a user, and carrying out feature extraction on the abnormal condition data based on a preset abnormal feature extraction template, wherein the abnormal feature extraction template defines a method and an algorithm for extracting features from the abnormal condition data, and the abnormal condition data can be converted into a representative feature set through feature extraction; extracting a plurality of groups of standard abnormal feature sets and abnormal types and severity which are in one-to-one correspondence from a preset abnormal classification experience library, wherein the standard feature sets and the corresponding abnormal types and severity of various abnormal conditions are stored in the abnormal classification experience library, and the types and severity of the abnormal conditions can be determined through matching with the abnormal feature sets; matching the abnormal feature set with any standard abnormal feature set to obtain abnormal matching degree, and calculating the abnormal matching degree by comparing the similarity or distance between the abnormal feature set and the standard abnormal feature set to measure the similarity between the abnormal condition and the standard abnormal condition; determining the type and severity of the abnormal condition based on the degree of abnormal match, and determining which type (such as fall or asphyxia) and severity level (such as mild, moderate or severe) the abnormal condition belongs to according to the degree of abnormal match;

Based on a preset abnormal severity classification standard, classifying the abnormal situation into different severity classes according to the preset abnormal severity classification standard so as to select corresponding safety measures later; extracting corresponding safety measure schemes from a preset safety measure experience library according to the determined abnormality type and severity level, wherein the safety measure schemes corresponding to various abnormality types and severity levels are stored in the safety measure experience library, and selecting the corresponding safety measure schemes according to the abnormality type and severity level; based on the safety measure scheme, corresponding safety measures are automatically selected and executed to ensure the safety of the user, and according to the selected safety measure scheme, the system can automatically execute corresponding safety measures, such as starting emergency light, providing oxygen supply and the like, so as to ensure the safety of the user.

By adopting a preset abnormal feature extraction template and an abnormal classification experience library, the abnormal condition can be identified and classified by adopting a feature extraction and matching mode, and a proper safety measure scheme can be selected according to the abnormal type and severity level by adopting a preset safety measure experience library.

The beneficial effects of the technical scheme are as follows: the abnormal conditions can be accurately identified and classified through a preset abnormal feature extraction template and an abnormal classification experience library, and the pertinence and the effectiveness of safety measures are ensured; by automatically selecting and executing the safety measures, the abnormal condition can be automatically processed, and the response speed and the response efficiency are improved; the preset abnormal characteristic extraction template, the abnormal classification experience library and the safety measure experience library have certain expandability, and new abnormal conditions and safety measure schemes can be conveniently added; according to the type and severity level of the abnormal condition, a corresponding safety measure scheme is selected, personalized safety protection can be provided, and the specific requirements of users are met.

It will be apparent to those skilled in the art that various modifications and variations can be made to the present invention without departing from the spirit or scope of the invention. Thus, it is intended that the present invention also include such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.

Claims (10)

1. The utility model provides a sensor integrated nursing system of thing networking which characterized in that includes:

the data acquisition unit is used for acquiring sensing data in real time through the sensor network of the Internet of things, wherein the sensing data comprises current environmental factors, the life state and the behavior gesture of a monitored person;

The data analysis unit is used for carrying out real-time scene recognition and analysis on the sensing data based on the guardian nursing judgment model to obtain an analysis result;

and the early warning response unit is used for immediately triggering an early warning mechanism when the analysis result shows abnormality, notifying a caretaker in real time and simultaneously automatically starting corresponding safety measures in the guardian.

2. The integrated sensor care system of claim 1, wherein the data acquisition unit comprises:

the self-adaptive configuration module is used for automatically configuring the positions and the number of the sensors of the Internet of things according to the size, the layout and the object distribution of the indoor space;

the environment sensing module is used for acquiring corresponding environment data based on a plurality of environment sensors according to real-time environment factors, adaptively adjusting the sensitivity and the data acquisition frequency of the sensors, and immediately adjusting corresponding sensor parameters when the environment factors are suddenly changed so as to adapt to new environment conditions;

and the vital state and behavior gesture sensing module is used for capturing vital sign changes and behavior gestures of the monitored person in real time based on the vital sign monitor and the behavior gesture recognition sensor.

3. The integrated sensor care system of claim 1, wherein the data analysis unit comprises:

The nursing judging module is used for dynamically constructing and optimizing a nursing judging model according to the historical data, the physiological characteristics and the daily habits of the monitored person;

the scene recognition module is used for carrying out feature extraction and situation classification on the sensing data based on a deep learning technology, and carrying out real-time situation annotation on the classified sensing data, wherein the situation annotation comprises cooking, reading and resting;

the comprehensive evaluation module is used for comprehensively evaluating the current state of the monitored person based on the scene recognition result and the nursing judgment model to obtain a comprehensive evaluation result;

the trend prediction module is used for carrying out trend analysis on the behavior and physiological states of the guardian based on a time sequence analysis technology, and predicting the corresponding states and requirements of the guardian in the future through the trend analysis.

4. The integrated sensor care system of claim 1, wherein the early warning response unit comprises:

the early warning triggering module is used for immediately triggering an early warning mechanism when an abnormal condition is detected, wherein the abnormal condition comprises sudden physiological sign change or unusual behavior mode;

the multi-channel notification module is used for simultaneously sending early warning information through a plurality of channels based on an information redundancy mechanism and notifying a caretaker in real time, wherein the channels comprise short messages, telephones and application pushing;

The safety measure starting module is used for automatically selecting and starting corresponding safety measures according to the type and severity of the abnormal condition, and corresponding safety measure equipment comprises emergency light, an automatic door lock and an oxygen supply system.

5. The integrated sensor care system of claim 2, wherein the life state and behavioral posture sensing module comprises: a vital sign monitor and a behavioral gesture recognition sensor;

a vital sign monitor for capturing key vital signs in real time, the key vital signs including heart rate, blood oxygen level and body temperature;

the behavior gesture recognition sensor is used for recognizing the behaviors and gestures of the monitored person in real time, wherein the behaviors and gestures comprise standing, sitting, walking or falling.

6. The integrated sensor care system of claim 3, wherein the scene recognition module comprises: a deep learning classifier;

based on the cyclic neural network, carrying out feature extraction on the sensing data, wherein the feature extraction process automatically identifies and extracts key modes and structures in the sensing data so as to facilitate subsequent situation classification; based on the extracted features, the sensing data is subjected to situation classification through a deep learning classifier, the sensing data is classified into different situations by the deep learning classifier, the situations comprise cooking, reading and resting, after the data is classified, the classified data is subjected to situation annotation in real time, and each section of sensing data is ensured to be matched with the corresponding situation.

7. The integrated sensor care system of claim 4, wherein the security initiation module comprises: an adaptive learning module;

continuously monitoring the incoming data to identify any abnormal situation in real time, automatically classifying the type and severity of the abnormal situation based on an abnormality classification model, the type including falls and asphyxia, the severity including mild, moderate and severe; according to the type and severity of the abnormal situation, corresponding safety measures are automatically selected, emergency light is started to remind surrounding people of falling event, and an oxygen supply system is started for suffocation or dyspnea; the self-adaptive learning module automatically adjusts and optimizes the selection and starting strategy of the safety measures according to the historical data and the user feedback.

8. The integrated sensor care system of claim 5, wherein a main control board level in the vital sign monitor is initialized, hardware in the vital sign monitor is abstracted and packaged, and an interface is provided for a vital sign monitor node program; starting an application task to execute a core function task of the vital sign monitor node, and creating a corresponding client when the application task is started, wherein the client sends a registration request through the service of the edge intelligent gateway and establishes connection with the edge intelligent gateway; entering a vital sign monitor node main circulation program, judging whether a counter value reaches a threshold value, if so, reading body temperature, acceleration and heart rate data through an interface, calculating a body temperature value, human activity intensity and heart rate value, updating data and a flag bit, and sending the data to an edge intelligent gateway; if the counter value does not reach the threshold value, the counter value is incremented by 1.

9. The integrated sensor care system of claim 6, wherein the context classification of the sensing data by the deep learning classifier comprises:

constructing a feature extraction information base corresponding to the sensing data;

acquiring a plurality of preset feature extraction nodes, and traversing each feature extraction node in sequence;

each time of traversing, acquiring a preset feature extraction template and a preset verification template corresponding to the node type of the traversed feature extraction node;

extracting required characteristic data from the sensing data based on the characteristic extraction template;

based on the verification template, performing feature verification according to the extracted feature data;

when the feature verification is passed, a preset training template of the deep learning classifier is obtained;

training by using the marked sensing data based on a training template of the deep learning classifier to obtain the deep learning classifier;

based on the deep learning classifier, carrying out situation classification on the extracted characteristic data;

classifying the sensing data into different situations according to the classification result, wherein the situations comprise cooking, reading and resting;

storing the classification result in a situation classification result library;

repeating the steps when new sensing data is input until all the sensing data are classified;

The situation classification result is updated in real time by comparing the new sensing data with the data in the situation classification result library so as to ensure the accuracy and real-time performance of the situation classification;

after determining the context, executing corresponding operations or reminders according to different contexts, wherein the corresponding operations or reminders comprise: under the cooking situation, the user is reminded to pay attention to the safety of the fire source, under the reading situation, indoor illumination is adjusted to protect eyesight, and under the rest situation, indoor temperature is adjusted to ensure comfort.

10. The integrated sensor care system of claim 7, wherein the automatic selection of the corresponding security measures based on the type and severity of the abnormal situation comprises:

acquiring abnormal condition data of a user;

based on a preset abnormal feature extraction template, carrying out feature extraction on abnormal condition data to obtain an abnormal feature set;

extracting a plurality of groups of standard abnormality feature sets and abnormality types and severity which are in one-to-one correspondence from a preset abnormality classification experience library;

matching the abnormal feature set with any standard abnormal feature set to obtain an abnormal matching degree;

determining the type and severity of the abnormal condition based on the degree of abnormal matching;

Types include falls and asphyxia, and severity includes mild, moderate, and severe, among others;

based on a preset abnormal severity classification standard, classifying severity grades of abnormal conditions;

extracting a corresponding safety measure scheme from a preset safety measure experience library according to the determined abnormality type and severity level;

based on the security measure scheme, corresponding security measures are automatically selected and executed to ensure the security of the user.