Personnel home safety monitoring method and device and computer readable storage medium

Abstract

The invention provides a personnel home safety monitoring method and device, and a computer readable storage medium, and the method comprises the steps: obtaining the effective behavior feature data of a target person, the target person being a home person needing safety monitoring, and the effective behavior feature data being a feature vector representing the behavior of the target person; data analysis is conducted on the effective behavior characteristic data through a data analysis technology, behavior key information of the target person is obtained, and the behavior key information at least comprises behavior modes, activity rules and association relationships among behaviors; evaluating the behavior key information to generate a corresponding evaluation report, wherein the evaluation report at least comprises the physical condition and the action ability of the target person and the environment condition of the target person; and carrying out monitoring and intelligent equipment cooperative early warning according to the evaluation report in combination with a set early warning rule. According to the invention, the problem of low intelligent monitoring and early warning efficiency of home safety of target persons such as old people and the like in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the field of safety monitoring technology, and in particular to a method for monitoring the safety of people at home, a device for monitoring the safety of people at home, and a computer-readable storage medium. Background Art

[0002] Without on-site supervision by children or medical staff, the elderly are prone to accidental falls and sudden illnesses that cannot be discovered and treated in time, endangering their personal safety. How to accurately monitor the indoor risks and vital signs of the elderly, and promptly notify caregivers and provide timely rescue in the event of an accident is a problem that plagues every family. At present, more are based on mobile Internet and mobile communication technologies, through smart wearable devices, cameras, smart door magnets, etc., to provide location services, emergency assistance, health management, life reminders and other functions for the elderly in the form of APP. However, existing smart elderly care products or services can only passively receive user instructions or abnormal information, cannot actively identify abnormal behaviors to prevent them in advance, cannot provide personalized service push for the elderly, cannot make intelligent decisions in emergency situations, cannot effectively link and respond to equipment, and cannot achieve remote centralized management of smart elderly care equipment.

[0003] In summary, the existing technology still has some problems:

[0004] 1. Unable to proactively identify abnormal behaviors, unable to timely discover and prevent risks, and lacking a functional linkage response mechanism;

[0005] 2. Unable to provide personalized service push for the elderly;

[0006] 3. Unable to make intelligent decisions in emergency situations;

[0007] 4. It is impossible to effectively link and respond to the equipment, and it is impossible to achieve remote centralized management of smart elderly care equipment;

[0008] 5. The access threshold for equipment and services is high, compatibility is poor, the failure rate is high, it is difficult to store data in a unified manner, and unified management is impossible. Summary of the invention

[0009] The main purpose of the present application is to provide a method for monitoring the safety of people at home, a device for monitoring the safety of people at home and a computer-readable storage medium, so as to at least solve the problem of low efficiency of intelligent monitoring and early warning of the home safety of target people such as the elderly in the prior art.

[0010] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a method for monitoring the safety of people at home is provided, including: obtaining effective behavioral characteristic data of a target person, wherein the target person is a person at home who needs to be monitored safely, and the effective behavioral characteristic data is a feature vector that characterizes the behavior of the target person; using data analysis technology to perform data analysis on the effective behavioral characteristic data to obtain key behavioral information of the target person, wherein the key behavioral information at least includes behavioral patterns, activity patterns, and correlations between behaviors; evaluating the key behavioral information to generate a corresponding evaluation report, wherein the evaluation report at least includes the physical condition, mobility, and environmental conditions of the target person; and performing monitoring and collaborative warning with intelligent equipment based on the evaluation report in combination with setting warning rules.

[0011] Optionally, obtaining effective behavioral characteristic data of the target person includes: obtaining real-time behavioral characteristic data of the target person through the intelligent device, the intelligent device at least including an intelligent voice assistant, a smart camera, a smart home device and a health monitoring sensor device, the health monitoring sensor device at least including an infrared sensor and a millimeter wave monitoring sensor; performing data cleaning, data denoising and format conversion on the real-time behavioral characteristic data in sequence to obtain the effective behavioral characteristic data.

[0012] Optionally, after obtaining the real-time behavioral characteristic data of the target person, the method further includes: classifying the real-time behavioral characteristic data from the time dimension and the space dimension to obtain multi-dimensional behavioral characteristic data; using data compression and encoding technology to compress and optimize the encoding of the multi-dimensional behavioral characteristic data to obtain compressed and optimized behavioral characteristic data; using a deep neural network to abstractly describe the compressed and optimized behavioral characteristic data, generate a corresponding feature vector and store the feature vector in a behavioral characteristic database, wherein the feature vector represents the behavioral characteristic data within a corresponding time period.

[0013] Optionally, data analysis technology is used to perform data analysis on the effective behavior characteristic data to obtain key behavior information of the target person, including: using a cluster analysis algorithm to perform clustering processing on the effective behavior characteristic data to identify the behavior pattern of the target person, and the behavior pattern at least includes a living pattern, a leisure pattern, and a dining pattern; using a time series analysis method to perform time processing on the effective behavior characteristic data to analyze the activity pattern of the target person, and the activity pattern at least includes the target person's waking time, sleeping time, and peak activity period; using association rule mining technology to perform data mining on the effective behavior characteristic data to analyze the correlation between the behaviors of the target person, and the correlation between behaviors at least includes the relationship between sleep quality and indoor environment.

[0014] Optionally, the key behavior information is evaluated to generate a corresponding evaluation report, including: inputting the key behavior information into a behavior evaluation model, analyzing the key behavior information using the behavior evaluation model, and outputting a behavior evaluation result of the target person, the behavior evaluation model being a convergence model obtained by iteratively training sample training data by inputting sample training data into a predetermined model using a machine learning algorithm, the sample training data including sample behavior key information and sample behavior evaluation results corresponding to the sample behavior key information, the behavior evaluation results including at least an activity volume index, a work and rest regularity index, and an environmental adaptation index, the activity volume index including at least the number of walking steps and activity time, the work and rest regularity index including at least the sleeping time and the waking up time, and the environmental adaptation index including at least the indoor temperature adaptation degree and the indoor humidity adaptation degree; generating the evaluation report according to the behavior evaluation result.

[0015] Optionally, the set warning rules include behavior pattern warning rules and environmental parameter dynamic warning rules, and monitoring and intelligent equipment collaborative warning are performed according to the evaluation report combined with the set warning rules, including: comparing the action ability in the evaluation report with the behavior pattern benchmark model, the behavior pattern benchmark model is a benchmark model established by at least analyzing the activity time distribution, activity area range and activity frequency characteristics of the target person in the historical time period, and the behavior pattern benchmark model is one of the behavior pattern warning rules; when the deviation between the evaluation report and the behavior pattern benchmark model exceeds a first set threshold, issuing a first warning message, the first warning message is used to prompt the nursing staff that the behavior of the current target person is abnormal; comparing the environmental conditions in the evaluation report with the environmental parameter dynamic warning rules, the environmental parameter dynamic warning rules are dynamically adjusted according to the sensitivity of the health records and disease information of the target person to indoor environmental factors, and the indoor environmental factors include at least indoor temperature, indoor humidity and indoor dust concentration; when the environmental conditions are greater than the set values ​​of the indoor environmental factors in the environmental parameter dynamic warning rules, issuing a second warning message, the second warning message is used to prompt the nursing staff that there is a health risk in the environment where the target person is currently located.

[0016] Optionally, when the deviation between the assessment report and the behavior pattern benchmark model exceeds a first set threshold, after issuing a first warning message, the method further includes: responding to the remote operation of the caregiver so that the caregiver can view the real-time status of the target person, and the remote operation includes at least one of the following: starting a three-dimensional virtual scene of the target person's environment through VR / AR technology, adjusting the camera angle and starting a voice call; or automatically linking multiple intelligent devices for collaborative monitoring and processing to record the real-time status of the target person.

[0017] Optionally, multiple intelligent devices are automatically linked to perform collaborative monitoring and processing to record the real-time status of the target person, including: in the event of a fall of the target person, controlling the intelligent voice assistant to call the caregiver and the platform customer service personnel, and simultaneously pushing detailed alarm information to the caregiver and the platform customer service personnel, wherein the alarm information includes the time and location of the fall and a real-time image screenshot obtained from the intelligent camera; automatically aiming the intelligent camera at the position where the target person falls to take high-definition photos, and using a deep learning image analysis algorithm to identify the body posture of the target person and predict the injured part; automatically communicating with the intelligent bracelet worn by the target person to obtain the heart rate and blood pressure data of the target person ; When the heart rate and the blood pressure data are within the normal range, control the smart speaker to issue a voice inquiry to the target person; when a voice response from the target person is received within the set time, voice prompt the target person to get up slowly, and inform the nursing staff and the platform customer service staff about the response of the target person, and continue to monitor the subsequent actions of the target person; when no voice response from the target person is received within the set time or the heart rate and the blood pressure data are not within the normal range, automatically dial an emergency number, and send the health record of the target person and real-time monitoring data to the emergency center, and the monitoring data at least includes the fall location, fall posture, heart rate and blood pressure data.

[0018] According to another aspect of the present application, a device for monitoring the safety of people at home is provided, and the device includes: an acquisition unit, which is used to acquire effective behavior characteristic data of a target person, wherein the target person is a person at home who needs to be monitored for safety, and the effective behavior characteristic data is a feature vector that characterizes the behavior of the target person; an analysis unit, which is used to perform data analysis on the effective behavior characteristic data using data analysis technology to obtain key behavior information of the target person, wherein the key behavior information includes at least behavior patterns, activity patterns, and correlations between behaviors; an evaluation unit, which is used to evaluate the key behavior information and generate a corresponding evaluation report, wherein the evaluation report includes at least the physical condition, mobility, and environmental conditions of the target person; and an early warning unit, which is used to perform monitoring and coordinated early warning with intelligent equipment based on the evaluation report in combination with set early warning rules.

[0019] According to another aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute any one of the methods described.

[0020] Applying the technical solution of this application, in the method for monitoring the safety of people at home, first, obtain the effective behavioral characteristic data of the target person, the target person is the person at home who needs to be monitored safely, and the effective behavioral characteristic data is the characteristic vector that characterizes the behavior of the target person; then, use data analysis technology to analyze the effective behavioral characteristic data to obtain the key behavioral information of the target person, the key behavioral information at least includes the behavior pattern, activity pattern and the correlation between behaviors; then, evaluate the key behavioral information to generate a corresponding evaluation report, the evaluation report at least includes the physical condition, mobility and environmental conditions of the target person; finally, monitor and coordinate warnings with intelligent equipment according to the evaluation report combined with the set warning rules. This application is based on real-time monitoring of multi-source data by monitoring equipment, and analyzes its activity level, activity preference, sleep pattern and other activities to generate corresponding evaluation reports, comprehensively capture the life patterns of elderly people in the home environment, summarize the unique situation of the elderly, and form a monitoring plan that can accurately analyze the abnormal state of the elderly, compare the current evaluation report with the set warning rules, and determine whether an abnormal situation occurs and issue a warning. Real-time and accurate monitoring and analysis of the behavior data of the elderly at home, providing timely and effective information support for family members and caregivers, will help improve the quality and safety of home-based elderly care services and enhance the quality of life of the elderly. At the same time, through personalized service push, it can better meet the personalized needs of the elderly and promote the precision and intelligent development of home-based elderly care services. This application solves the problem of low efficiency of intelligent monitoring and early warning of home safety for target personnel such as the elderly in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A hardware structure block diagram of a mobile terminal for executing a method for monitoring home safety of personnel provided in an embodiment of the present application is shown;

[0022] Figure 2 A schematic diagram of a process of a method for monitoring safety of a person at home provided according to an embodiment of the present application is shown;

[0023] Figure 3 A logic diagram of a person's home safety monitoring system provided according to an embodiment of the present application is shown;

[0024] Figure 4 A structural block diagram of a home safety monitoring system for personnel provided according to an embodiment of the present application is shown;

[0025] Figure 5 A structural block diagram of a device for monitoring home safety of personnel provided according to an embodiment of the present application is shown.

[0026] The above drawings include the following reference numerals:

[0027] 102, processor; 104, memory; 106, transmission device; 108, input and output devices. DETAILED DESCRIPTION

[0028] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0029] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0030] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0031] As introduced in the background technology, existing smart elderly care products or services in the prior art can only passively receive user instructions or abnormal information, cannot actively identify abnormal behavior to prevent it in advance, cannot provide personalized service push for the elderly, cannot make intelligent decisions in emergency situations, cannot effectively link and respond to equipment, and cannot achieve remote centralized management of smart elderly care equipment. In order to solve the problem of low efficiency of intelligent monitoring and early warning of home safety of target personnel such as the elderly in the prior art, the embodiments of the present application provide a person's home safety monitoring method, a person's home safety monitoring device and a computer-readable storage medium.

[0032] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0033] The method embodiments provided in the embodiments of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Taking running on a mobile terminal as an example, Figure 1 FIG. 1 is a hardware structure block diagram of a mobile terminal of a method for monitoring home safety of a person according to an embodiment of the present invention. Figure 1 As shown, the mobile terminal may include one or more ( Figure 1 Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, wherein the mobile terminal may also include a transmission device 106 and an input / output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the mobile terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations shown.

[0034] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the personnel home safety monitoring method in the embodiment of the present invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, the above method is implemented. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories can be connected to the mobile terminal via a network. Examples of the above-mentioned network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof. The transmission device 106 is used to receive or send data via a network. The above-mentioned specific examples of the network may include a wireless network provided by a communication provider of the mobile terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0035] In this embodiment, a method for monitoring the safety of people at home that runs on a mobile terminal, a computer terminal or a similar computing device is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0036] Figure 2 is a flow chart of a method for monitoring home safety of personnel according to an embodiment of the present application. Figure 2 As shown, the method comprises the following steps:

[0037] Step S201, obtaining effective behavior characteristic data of a target person, the target person being a person at home who needs to be monitored for safety, and the effective behavior characteristic data being a characteristic vector that characterizes the behavior of the target person.

[0038] Specifically, Figure 3 As shown in the figure, the data collection module is used to collect the behavioral characteristic data of the target person. Various intelligent devices are installed in the target person's living environment, including but not limited to motion sensors, environmental sensors (temperature, humidity), and biometric sensors (heart rate, blood pressure, etc.). These devices automatically collect the target person's daily behavioral data, such as activity time, activity location, indoor environmental parameters, vital signs, etc. Figure 4 As shown, after data collection, the collected raw data is cleaned and denoised to exclude outliers and invalid data, and then feature extraction is performed to convert complex behavioral data into quantifiable feature vectors.

[0039] Step S202, using data analysis technology to analyze the above-mentioned effective behavior feature data to obtain the above-mentioned key behavior information of the target person, and the above-mentioned key behavior information at least includes behavior patterns, activity rules and correlation relationships between behaviors.

[0040] Specifically, Figure 3 The data analysis module in clusters the above-mentioned valid behavioral feature data and identifies different behavioral patterns, such as living patterns, leisure patterns, and dining patterns. Analyze the time series characteristics of behavioral data, such as daily wake-up time, peak activity period, and sleep patterns. Use the Apriori algorithm, etc. to discover the correlation between behavioral data, such as the relationship between sleep quality and indoor environmental parameters. The identification of behavioral patterns helps the system understand the living habits of target personnel and provide more personalized services. The analysis of time patterns can predict the daily activities of target personnel and provide a time frame for early warning and response. Mining the correlation between behaviors helps to discover potential health risks, such as the impact of environmental factors on physical condition.

[0041] Step S203, evaluating the above-mentioned key behavior information to generate a corresponding evaluation report, wherein the above-mentioned evaluation report at least includes the physical condition, mobility and environmental conditions of the above-mentioned target person.

[0042] Specifically, the key information of the behavior is input into the behavior assessment model, which may be trained based on a machine learning algorithm and is used to assess the physical condition, mobility and environmental conditions of the target person. Based on the input information, the assessment model uses an algorithm to calculate the values ​​of various assessment indicators, such as activity level, sleep quality, and comfort of the indoor environment. A detailed assessment report is generated based on the assessment results, including current status, trend analysis, and potential risks. The assessment report provides family members and caregivers with a comprehensive view of the health status of the target person, facilitating timely care decisions. It can identify abnormal changes in behavioral patterns and provide early warning of health risks, such as the risk of falling and the risk of worsening chronic diseases.

[0043] Step S204, monitoring and intelligent equipment collaborative warning are performed based on the above assessment report and set warning rules.

[0044] Specifically, based on the information in the assessment report, combined with the health records and personalized needs of the elderly, warning thresholds and conditions are set, such as long periods of inactivity and abnormally short sleep time. The system monitors the key behavioral information of the target personnel in real time, and immediately triggers the warning mechanism once the data deviates from the normal range set by the warning rules. According to the severity of the warning, the system sends an alert to the preset contacts (such as family members and medical staff), and provides detailed behavioral data and environmental information to quickly assess the situation and take action. Real-time monitoring and early warning of health risks of the elderly are achieved, and the response speed and efficiency in emergency situations are improved. Improve the accuracy and pertinence of early warnings to avoid unnecessary worries and waste of resources. Promote the continuous attention of families and caregivers to the safety of the elderly, and provide a safer living environment and timely health intervention for the elderly.

[0045] In summary, through the implementation of the above steps, the system can continuously monitor and analyze the behavioral characteristics of the elderly, generate evaluation reports and early warning information, significantly improve the intelligent level of home-based elderly care services, help to timely discover and respond to health risks, and ensure the safety and health of the elderly. The present invention can monitor and analyze the behavioral data of the elderly at home in real time and accurately, provide timely and effective information support for family members and caregivers, help improve the service quality and safety of home-based elderly care, and improve the quality of life of the elderly. At the same time, through personalized service push, it can better meet the personalized needs of the elderly and promote the precision and intelligent development of home-based elderly care services.

[0046] In this embodiment, first, obtain the effective behavior characteristic data of the target person, the target person is a person at home who needs to be monitored safely, and the effective behavior characteristic data is a characteristic vector that characterizes the behavior of the target person; then, use data analysis technology to analyze the effective behavior characteristic data to obtain the key behavior information of the target person, the key behavior information at least includes behavior patterns, activity rules and correlations between behaviors; then, evaluate the key behavior information to generate a corresponding evaluation report, the evaluation report at least includes the physical condition, mobility and environmental conditions of the target person; finally, monitor and coordinate warnings with intelligent devices based on the evaluation report and the set warning rules. This application is based on real-time monitoring of multi-source data by monitoring equipment, and analyzes its activity level, activity preference, sleep pattern and other activities to generate corresponding evaluation reports, comprehensively capture the life patterns of elderly people in the home environment, summarize the unique situation of the elderly, and form a monitoring plan that can accurately analyze the abnormal state of the elderly, compare the current evaluation report with the set warning rules, and determine whether an abnormal situation occurs and issue a warning. Real-time and accurate monitoring and analysis of the behavior data of the elderly at home, providing timely and effective information support for family members and caregivers, will help improve the quality and safety of home-based elderly care services and enhance the quality of life of the elderly. At the same time, through personalized service push, it can better meet the personalized needs of the elderly and promote the precision and intelligent development of home-based elderly care services. This application solves the problem of low efficiency of intelligent monitoring and early warning of home safety for target personnel such as the elderly in the prior art.

[0047] In order to enable those skilled in the art to more clearly understand the technical solution of the present application, the implementation process of the method for monitoring the safety of persons at home of the present application will be described in detail below in combination with specific embodiments.

[0048] In order to ensure the accuracy and reliability of the monitoring data, in an optional implementation manner, the above step S201 includes:

[0049] Step S2011, obtaining real-time behavioral characteristic data of the target person through intelligent devices, wherein the intelligent devices at least include intelligent voice assistants, intelligent cameras, smart home devices, and health monitoring sensor devices, and the health monitoring sensor devices at least include infrared sensors and millimeter wave monitoring sensors;

[0050] Step S2012, performing data cleaning, data denoising and format conversion on the above-mentioned real-time behavior characteristic data in sequence to obtain the above-mentioned effective behavior characteristic data.

[0051] In the above embodiments, it should be noted that, based on the advanced NB-IoT technology, unified standard data docking of devices is achieved to create a comprehensive and intelligent home-based elderly care safety monitoring and early warning system. The system conducts in-depth analysis of daily monitoring data collected by various home safety protection and health monitoring devices to build a data model of the elderly's home-based elderly care living environment, aiming to accurately and comprehensively analyze the elderly's home behavior patterns, so as to effectively predict the severity of the situation corresponding to the alarm information and provide strong support for rapid rescue. A series of advanced sensor devices are used, including infrared, millimeter wave and comprehensive health monitoring devices that can accurately detect changes in movement patterns and effectively identify potential fall risks. These sensors can not only monitor the physical movement status of the elderly in real time, but also collect key health indicator data such as heart rate and blood pressure. For the home environment, it is equipped with complete indoor environment detection equipment, covering gas detection, temperature detection, humidity detection and other functions, to fully ensure the safety and comfort of the home environment. At the same time, security alarm equipment is set up, such as a boundary alarm device that can monitor abnormal intrusions around the house in real time, and an SOS alarm device provides a convenient channel for the elderly to seek help in an emergency. All sensor devices use NB-IoT technology to establish efficient and stable data connections with the system, ensuring real-time data transmission and unified standard docking. The device sends the collected raw data to the cloud server, which first performs preliminary screening and cleaning of the data to remove outliers and noise interference to ensure the accuracy and reliability of the data. Based on the real-time monitoring data of the device, a database of elderly behavior characteristics is established. This database not only stores the daily activity data of the elderly, but also records and analyzes their activity levels, activity preferences, sleep patterns, etc. in detail, comprehensively capturing the details of the life patterns of the elderly in the home environment, and providing rich data support for subsequent behavior analysis and early warning judgments. The main processing logic is as follows: Figure 3 shown.

[0052] Receive and integrate real-time behavior feature data from the sensor network, and perform preliminary preprocessing and formatting. That is, clean, denoise and convert the original real-time behavior feature data to obtain effective behavior feature data so that it can be effectively processed by subsequent analysis modules.

[0053] In addition, data preprocessing technology: data processing is an important part of the monitoring process, including data cleaning, quality control, data analysis and result reporting. Through effective data processing, the accuracy and reliability of monitoring data can be ensured. In data mining, data preprocessing is to perform a series of processing work such as necessary cleaning, integration, conversion, discretization and reduction on the original data before data mining, so that it can meet the minimum specifications and standards required by the mining algorithm for knowledge acquisition research. Data cleaning is to eliminate errors or duplicate data in the data, remove noise or irrelevant data, and process missing data fields in the data; data integration combines data from multiple data sources in a unified data storage, resolves semantic ambiguity and integrates it into a consistent data storage; data selection is based on the understanding of the discovery task and the content of the data itself, looking for useful features of the expression data that depend on the discovery target, reducing the size of the data, and maximizing the amount of data while keeping the original data as much as possible. Data selection can make the regularity and potential characteristics of the data more obvious. Through data preprocessing, incomplete data can be made complete, wrong data can be corrected, redundant data can be eliminated, the required data can be selected for data integration, and the inappropriate data format can be converted to the required format, so as to achieve data type homogeneity, data format consistency, data information refinement and data storage centralization. In short, after preprocessing, not only can the data set required by the mining system be obtained, but also the cost of the mining system can be minimized and the effectiveness and comprehensibility of the mined knowledge can be improved.

[0054] In order to ensure that data is not lost and improve data query efficiency, in an optional implementation manner, after the above step S2011, the method further includes:

[0055] Step S301, classifying the above-mentioned real-time behavior characteristic data from the time dimension and the space dimension to obtain multi-dimensional behavior characteristic data;

[0056] Step S302, using data compression and encoding technology to compress and optimize the multi-dimensional behavior feature data to obtain compressed and optimized behavior feature data;

[0057] Step S303, using a deep neural network to abstractly describe the above-mentioned compressed optimized behavior feature data, generate a corresponding feature vector and store the above-mentioned feature vector in a behavior feature database, wherein the above-mentioned feature vector represents the above-mentioned behavior feature data in a corresponding time period.

[0058] In the above embodiment, for massive, continuous and time-space related data, a distributed data storage model based on time-space indexing is adopted. By dividing and indexing the time and space dimensions, the target person's behavior data is classified and stored according to the time series and spatial position. For example, the target person's behavior data in different rooms and different time periods are stored in corresponding time-space partitions, and each partition corresponds to a specific time range and spatial area. In this way, when querying, the relevant data partitions can be quickly located to improve the query efficiency.

[0059] Using data sharding and replication technology, large-scale data is divided into smaller fragments and multiple copies are stored on different nodes. When querying data, the system can read data fragments from multiple nodes in parallel to speed up data reading. At the same time, the replication mechanism ensures data availability and fault tolerance. When a node fails, other replicas can continue to provide services to ensure that data is not lost.

[0060] In addition, data compression and encoding technologies are introduced to compress the stored data and reduce the storage space occupied by the data. For example, lossless compression algorithms (such as LZ4, etc.) are used to compress the raw data collected by the sensor to reduce storage costs without losing data accuracy. At the same time, data encoding is optimized, such as using columnar storage formats (such as Parquet, etc.) to improve data query performance, especially when querying specific columns, data can be quickly filtered and read.

[0061] When storing the target person's behavior data, deep learning technology is used for abstract description and optimized storage. A deep neural network is used to model and extract features of the target person's behavior over a period of time. The target person's behavior sequence is input into the neural network, and the intrinsic pattern and feature representation of the behavior are learned through a multi-layer perceptron or a recurrent neural network (such as LSTM, GRU, etc.). Then, the learned feature vector is stored as an abstract representation of the behavior in that time period, replacing the traditional detailed record of the state at each time point.

[0062] For example, for the target person's one-day activity behavior data, the deep learning model extracts the main activity patterns (such as rest, activity, and dining) and related features (such as activity duration, activity frequency, and activity intensity), and combines these features into a feature vector for storage. This not only greatly reduces the update frequency and storage space occupied by the database, but also retains the key information of the behavior data, which is convenient for subsequent analysis and processing.

[0063] In addition, the storage process can also include: In the scenario of intelligent device linkage, to cope with the rapid growth of data volume over time and the high requirements for memory space and query efficiency, the innovative technology of integrating distributed storage and cloud computing is introduced. Distributed file systems (such as Ceph, etc.) are used to store data in multiple nodes to achieve horizontal expansion of data and improve storage capacity and reliability. At the same time, combined with the elastic computing resources of the cloud computing platform, the storage and computing resource configuration is dynamically adjusted according to the real-time growth and access requirements of data.

[0064] By carefully defining the attributes and methods of the class, object-oriented behavior data storage is achieved. The target person's behavior data is encapsulated into a class. The attributes of the class contain the characteristic parameters of the behavior, and the methods are used to operate and process the behavior data, such as writing, reading, updating and querying data. At the same time, the class is mapped to the database table using the object-relational mapping (ORM) framework (such as Hibernate) to achieve automatic storage and management of data. This can improve the flexibility and maintainability of data storage and facilitate developers to operate and manage data, that is, the class is written into the feature vector as a feature in the feature vector for storage.

[0065] During the storage process, data version control technology is combined to record and manage changes in the target person's behavior data. Each time the behavior data is updated, a new version is created, and the differences and change history between versions are recorded. This makes it easy to trace the data change process and restore to a specific historical version when necessary, providing strong support for data analysis and troubleshooting. For example, when the behavior data of a certain period of time is found to be abnormal, version control can be used to trace back to the previous normal version, compare and analyze the data changes, and find out the cause of the abnormality.

[0066] In order to accurately identify and understand the behavior patterns, activity patterns and correlations between different behaviors of the target person, in an optional implementation, the above step S202 includes:

[0067] Step S2021, clustering the effective behavior feature data using a clustering analysis algorithm to identify the behavior pattern of the target person, where the behavior pattern at least includes a living pattern, a leisure pattern, and a dining pattern;

[0068] Step S2022, using a time series analysis method to perform time processing on the effective behavior feature data to analyze the activity pattern of the target person, the activity pattern at least including the target person's wake-up time, bedtime and activity peak period;

[0069] Step S2023, using association rule mining technology to perform data mining on the above-mentioned effective behavior feature data to analyze the association relationship between the above-mentioned behaviors of the above-mentioned target person, and the above-mentioned association relationship between behaviors at least includes the relationship between sleep quality and indoor environment.

[0070] In the above embodiment, Figure 3 The data analysis module uses data mining, machine learning, deep learning and other technologies to conduct in-depth analysis of the collected data and extract key information such as the target personnel's behavior patterns, activity patterns, etc.

[0071] Using cluster analysis algorithms, the daily behavior data of target personnel are classified to identify different behavior patterns, such as living patterns, leisure patterns, dining patterns, etc. Using time series analysis methods, the time patterns of target personnel's behaviors are analyzed, such as daily wake-up time, bedtime, peak activity period, etc. Using association rule mining technology, the correlation between different behaviors is discovered, such as the relationship between sleep quality and indoor temperature and humidity.

[0072] In addition, a behavioral data set for the target personnel is constructed. The demand analysis revolves around data acquisition; data processing, including preprocessing and storage; and data application, i.e., for behavior identification and prediction. At the same time, it is clear that the functions provided in this work phase need to select the basis for relevant technologies and methods. The collected and preprocessed data is used for behavioral analysis. It is necessary to consider which behaviors can reflect the daily living habits of the target personnel; which indicators of these behaviors are predicted and analyzed, such as the start time, duration, and other behaviors that occurred during the process, whether it is necessary to further analyze the work and rest habits of the target personnel or which health indicators are related to them, etc.; the efficiency and accuracy of the behavioral prediction method used.

[0073] Since each target person may perform the same behavior multiple times a day, or multiple behaviors of multiple target persons are intertwined, a large amount of data will be generated, and the data will change in real time at different time points in the morning and evening every day.

[0074] The association rule mining algorithm based on the Apriori algorithm is used. First, the processed events are serialized, that is, the transactions are represented by numbers. Since the same event may occur multiple times during the day, the same event on the same day can be distinguished by different uppercase and lowercase numbers. The same event may occur multiple times on the same day. The first time it occurs, it is represented by lowercase numbers, and when it occurs again, it can be represented by uppercase numbers, such as going to the toilet and taking a nap. The target person's daily behavior may be getting out of bed, going to the toilet (toilet stay time), living room activities (living room sedentary time), and so on. Then, the minimum support and minimum confidence association rule data mining algorithm program based on the Apriori algorithm is implemented in Python language. The input is the event sequence with minimum support and minimum confidence, and the output is the association rule.

[0075] The frequent item sets are found from the behavior list data set through the Apriori algorithm. Frequent item sets are the frequently associated monitoring data of the target person in the multi-linkage monitoring environment of smart devices. They can reflect the daily behavior habits of the target person and can mine the daily life pattern of the target person. The frequent item sets will be in the form of {1, 2} or {1, 2, 3}, that is, the behavior group of {getting out of bed, going to the toilet} or {getting out of bed, going to the toilet, living room activities}. Then generate association rules. The association rules are the correlation between the behaviors of the target person, which can explain whether there is a correlation between the behaviors of the target person. Each rule has only one item on the right side. Therefore, one behavior or a combination of several behaviors can be associated with another behavior to achieve the behavior prediction of the target person. The style of the association rule is as follows: (getting out of bed, going to the toilet) -> living room activities, to indicate the correlation between behaviors. After getting out of bed and going to the toilet, it is predicted that living room activities will occur. If the duration of going to the toilet deviates from the daily duration, a certain warning will be issued.

[0076] In order to improve the accuracy of behavior assessment, in an optional implementation, the above step S203 includes:

[0077] Step S2031, input the above-mentioned key behavior information into the behavior evaluation model, so as to analyze the above-mentioned key behavior information using the above-mentioned behavior evaluation model and output the behavior evaluation result of the above-mentioned target person, the above-mentioned behavior evaluation model is a convergence model obtained by inputting sample training data into a predetermined model for iterative training using a machine learning algorithm, the above-mentioned sample training data includes sample key behavior information and sample behavior evaluation results corresponding to the above-mentioned sample key behavior information, the above-mentioned behavior evaluation results at least include activity volume index, work and rest regularity index and environmental adaptation index, the above-mentioned activity volume index at least includes walking steps and activity time, the above-mentioned work and rest regularity index at least includes sleeping time and waking time, and the above-mentioned environmental adaptation index at least includes indoor temperature adaptation degree and indoor humidity adaptation degree;

[0078] Step S2032: Generate the above-mentioned evaluation report according to the above-mentioned behavior evaluation result.

[0079] In the above embodiment, a comprehensive behavior evaluation index system is established, including activity indicators (such as walking steps, activity time), work and rest regularity indicators (such as sleep duration, stability of waking time), environmental adaptation indicators (such as adaptation to indoor temperature and humidity), etc. A behavior evaluation model is constructed based on a machine learning algorithm, and key behavioral information of the target person is input to evaluate the behavior of the target person, such as daily activity, sleep quality, dietary regularity, etc., output the behavior evaluation results, and generate a corresponding evaluation report. The system can not only conduct real-time and accurate evaluation of the behavior of the target person, but also generate an evaluation report with practical guiding significance, so as to realize more refined home-based elderly care services and improve the quality of life and safety of the elderly.

[0080] In order to timely discover the abnormal situation of the target personnel and issue an early warning, in an optional implementation manner, the above step S204 includes:

[0081] Step S2041, comparing the action capability in the assessment report with a behavior pattern benchmark model, wherein the behavior pattern benchmark model is a benchmark model established by analyzing at least the activity time distribution, activity area range, and activity frequency characteristics of the target person in a historical time period, and the behavior pattern benchmark model is one of the behavior pattern warning rules;

[0082] Step S2042, when the deviation between the assessment report and the behavioral pattern benchmark model exceeds a first set threshold, a first warning message is issued, wherein the first warning message is used to remind the nursing staff that the behavior of the target person is abnormal;

[0083] Step S2043, comparing the environmental conditions in the assessment report with the environmental parameter dynamic warning rules, the environmental parameter dynamic warning rules being dynamically adjusted according to the sensitivity of the target person's health records and disease information to indoor environmental factors, the indoor environmental factors at least including indoor temperature, indoor humidity and indoor dust concentration;

[0084] Step S2044, when the above environmental conditions are greater than the set value of the above indoor environmental factors in the above environmental parameter dynamic warning rules, a second warning message is issued, and the above second warning message is used to remind the above caregiver that there is a health risk in the environment where the above target person is currently located.

[0085] In the above embodiment, if Figure 3As shown in the figure, the early warning module uses the latest artificial intelligence technology to achieve real-time monitoring and accurate early warning of the target personnel's behavior and environment. By building a multi-layer perception intelligent monitoring network and integrating multiple sensor data (such as motion sensors, environmental sensors, biosensors, etc.), comprehensive information on the target personnel's living status is obtained. Deep learning algorithms are used to analyze and process these multi-source data in real time, and a dynamic model of the elderly's behavior and environment is constructed to detect abnormal situations in a timely manner and issue early warnings.

[0086] The system automatically identifies and builds personalized behavior patterns by deeply learning the long-term behavior data of the target person. For example, the system analyzes the characteristics of the target person's daily activity time distribution, activity area range, activity frequency, etc., and establishes a unique behavior benchmark model for each target person. When the behavior of the elderly is significantly deviated from the benchmark model, an early warning is triggered. For example, if the target person usually has a certain activity pattern in the morning, but has not shown activities that conform to his habits for several consecutive hours, the system will automatically issue an early warning (i.e., the first early warning information mentioned above). Time series analysis and anomaly detection algorithms are introduced to monitor the time trend of the target person's behavior in real time. Not only does it focus on the absolute state of the target person's behavior, but it also analyzes whether the trend of behavior over time is abnormal. For example, by analyzing the daily activity change curve of the target person, if it is found that the activity on a certain day has dropped significantly in a short period of time and exceeds the normal fluctuation range calculated based on historical data, even if the current absolute value of the activity may not have reached the traditional early warning threshold, the system will pre-identify and warn in advance.

[0087] Using intelligent environmental perception technology, various parameters of the indoor environment are monitored in real time, and the warning threshold is dynamically adjusted according to the real-time status and historical data of the target personnel. For example, by analyzing the health records and historical disease information of the target personnel, their sensitivity to environmental factors such as temperature and humidity can be understood. For target personnel with respiratory diseases, when the indoor humidity exceeds its suitable range and lasts for a certain period of time, the system will issue more sensitive warning information in advance according to their specific health status and historical disease patterns. Combined with the Internet of Things and big data analysis, the associated warning of environmental parameters is realized. Not only does it focus on the abnormality of a single environmental parameter, but it also analyzes the potential impact of the coordinated changes between multiple environmental parameters on the health of the target personnel. For example, when the indoor temperature suddenly rises and the air quality indicators decrease at the same time (such as an increase in the concentration of harmful gases or an excess of PM2.5), the system will comprehensively consider these factors and issue a higher level of warning (i.e., the second warning information mentioned above), indicating possible health risks, because this environmental combination may have a greater impact on the respiratory system, cardiovascular system, etc. of the target personnel.

[0088] In order to improve the service quality and safety of home-based elderly care, in an optional implementation manner, after the above step S2042, the method further includes:

[0089] Step S401, responding to the remote operation of the nursing staff so that the nursing staff can view the real-time status of the target person, wherein the remote operation includes at least one of the following: starting a three-dimensional virtual scene of the target person's environment through VR / AR technology, adjusting the camera angle and starting a voice call;

[0090] Step S402, or, automatically linking multiple intelligent devices to perform collaborative monitoring and processing, and record the above-mentioned real-time status of the above-mentioned target personnel.

[0091] In the above embodiment, when the behavior data is monitored to trigger the warning rule, the system uses a variety of intelligent methods to send warning information to the preset contacts. In addition to traditional SMS, phone calls and application push, intelligent voice assistants and virtual reality (VR) / augmented reality (AR) technology are also introduced. The intelligent voice assistant can directly call the contact's phone and broadcast detailed warning content and the current status information of the elderly through voice. For example, inform the family members that "your relative has been inactive for a long time at [specific time]. The current indoor temperature is [X] degrees and the humidity is [Y]%. We have observed that the elderly is at [specific location] through the smart camera. Please pay attention to it in time." VR / AR technology is used to provide family members and caregivers with a more intuitive display of the real-time status of the elderly. After receiving the warning information, family members and caregivers can view the three-dimensional virtual scene of the elderly's environment and the real-time status of the elderly (such as a virtual model generated in real time by images captured by the camera and sensor data) through VR / AR applications on mobile phones or other devices. At the same time, they can interact with the system through gestures or voice commands to obtain more detailed information or perform remote operations, such as adjusting the camera angle and viewing historical data.

[0092] After the warning is triggered, the system realizes intelligent linkage response of multiple devices. For example, when the fall detection device finds that the elderly have fallen and triggers the fall alarm, the system will not only immediately make a phone call and send an alarm message to relatives and platform customer service personnel, but also automatically link multiple devices for collaborative processing. The smart camera will quickly adjust the angle to take a high-definition picture of the elderly's fall position, and use image recognition and analysis technology to evaluate the elderly's posture and possible injuries in real time. At the same time, through the smart lighting system installed indoors, the lights around the fall position are brightened to observe the elderly's condition more clearly.

[0093] The system will make intelligent decisions based on the data collected by the linkage devices. For example, by interacting with the data of smart bracelets or other wearable devices, the vital signs information of the elderly (such as heart rate, blood pressure, blood oxygen saturation, etc.) is obtained. If these vital signs data show normal, the system will automatically start the voice interaction module and communicate with the elderly through smart speakers or other voice devices installed nearby to confirm whether they can get up on their own or what kind of help they need. If the elderly respond that there is nothing wrong, the system will perform corresponding operations according to their instructions, such as temporarily stopping further alarm notifications, but continue to monitor the status of the elderly. If the vital signs data is abnormal, the system will automatically call the emergency number according to the preset urgency level, and send the elderly’s location information, health records and real-time monitoring data to the emergency center with one click, and notify the community medical staff to go to the scene as soon as possible to assist.

[0094] In order to improve the response speed and processing accuracy of the fall event, in an optional implementation manner, the above step S402 includes:

[0095] Step S4021, when the target person falls, the intelligent voice assistant is controlled to call the caregiver and the platform customer service personnel, and push detailed alarm information to the caregiver and the platform customer service personnel, the alarm information includes the time and place of the fall and the real-time image screenshot obtained from the intelligent camera;

[0096] Step S4022, automatically aiming the smart camera at the position where the target person falls to take a high-definition photo, and using a deep learning image analysis algorithm to identify the body posture of the target person and predict the injured part;

[0097] Step S4023, automatically communicating with the smart bracelet worn by the target person to obtain the heart rate and blood pressure data of the target person;

[0098] Step S4024, when the heart rate and the blood pressure data are within a normal range, controlling the smart speaker to send a voice inquiry to the target person;

[0099] Step S4025, when a voice response from the target person is received within a set time, a voice prompt is given to the target person to stand up slowly, and the nursing staff and the platform customer service staff are informed of the response of the target person, while the subsequent actions of the target person are continuously monitored;

[0100] Step S4026, if no voice response is received from the target person within the set time or the heart rate and blood pressure data are not within the normal range, automatically dial the emergency number and send the health record of the target person and the real-time monitoring data to the emergency center. The monitoring data at least includes the fall location, fall posture, heart rate and blood pressure data.

[0101] In the above embodiment, an application example is given: an elderly person accidentally falls while walking in the living room. At this time, the fall detection device installed in the living room quickly detects this situation and triggers a fall alarm.

[0102] The system immediately responded according to the preset process. First, the intelligent voice assistant called the family members and the platform customer service staff, and pushed detailed alarm information to their mobile applications, including the time and location of the fall and the real-time image screenshots obtained from the camera. Then, the smart camera in the living room was automatically linked to the position where the elderly fell to take high-definition photos, and the deep learning image analysis algorithm was used to quickly identify the elderly’s body posture and possible injured parts. For example, by analyzing the elderly’s body posture and facial expressions, it was initially determined whether there were signs of head injuries or fractures.

[0103] At the same time, the system communicates with the smart bracelet worn by the elderly to obtain their real-time heart rate and blood pressure data. If the heart rate and blood pressure data are normal, the system will send a voice inquiry through the smart speaker: "Hello, do you need help? Can you get up by yourself?" If the elderly answer that they can get up by themselves within the set time, the system will voice prompt the elderly to get up slowly, and inform the family members and platform customer service staff of the elderly's response, while continuing to monitor the elderly's follow-up actions. If the elderly do not respond within the set time or the vital signs data are abnormal, the system will immediately call the emergency number and send the pre-prepared health records of the elderly (including medical history, allergy history, etc.) and real-time monitoring data (such as fall location, posture, heart rate, blood pressure, etc.) to the emergency center. In addition, the system will notify community medical staff to bring necessary first aid equipment to the scene as soon as possible to provide initial assistance before the emergency personnel arrive.

[0104] This early warning mechanism of multi-device linkage and intelligent decision-making has greatly improved the response speed and accuracy of handling falls. By communicating with the elderly in a timely manner to confirm their condition, unnecessary panic and false alarms are avoided, and ineffective intervention by guardian family members and platform employees due to uncertain situations is also reduced. In cases where emergency rescue is really needed, the system can quickly and accurately provide comprehensive information to emergency centers and community medical staff, buying precious treatment time for the elderly and effectively reducing the serious consequences of falls.

[0105] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0106] The embodiment of the present application also provides a device for monitoring the safety of a person at home. It should be noted that the device for monitoring the safety of a person at home in the embodiment of the present application can be used to execute the method for monitoring the safety of a person at home provided in the embodiment of the present application. The device is used to implement the above-mentioned embodiments and preferred implementation modes, and those that have been described will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware for a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceived.

[0107] The following is an introduction to the personnel home safety monitoring device provided in the embodiments of the present application.

[0108] Figure 5 is a structural block diagram of a device for monitoring home safety of personnel according to an embodiment of the present application. Figure 5 As shown, the device comprises:

[0109] The acquisition unit 10 is used to acquire effective behavior characteristic data of a target person, where the target person is a person at home who needs to be monitored for safety, and the effective behavior characteristic data is a characteristic vector that characterizes the behavior of the target person.

[0110] Specifically, Figure 3 As shown in the figure, the data collection module is used to collect the behavioral characteristic data of the target person. Various intelligent devices are installed in the target person's living environment, including but not limited to motion sensors, environmental sensors (temperature, humidity), and biometric sensors (heart rate, blood pressure, etc.). These devices automatically collect the target person's daily behavioral data, such as activity time, activity location, indoor environmental parameters, vital signs, etc. Figure 4 As shown, after data collection, the collected raw data is cleaned and denoised to exclude outliers and invalid data, and then feature extraction is performed to convert complex behavioral data into quantifiable feature vectors.

[0111] The analysis unit 20 is used to perform data analysis on the above-mentioned effective behavior feature data using data analysis technology to obtain the above-mentioned key behavior information of the target person, and the above-mentioned key behavior information at least includes behavior patterns, activity rules and correlation relationships between behaviors.

[0112] Specifically, Figure 3The data analysis module in clusters the above-mentioned valid behavioral feature data and identifies different behavioral patterns, such as living patterns, leisure patterns, and dining patterns. Analyze the time series characteristics of behavioral data, such as daily wake-up time, peak activity period, and sleep patterns. Use the Apriori algorithm, etc. to discover the correlation between behavioral data, such as the relationship between sleep quality and indoor environmental parameters. The identification of behavioral patterns helps the system understand the living habits of target personnel and provide more personalized services. The analysis of time patterns can predict the daily activities of target personnel and provide a time frame for early warning and response. Mining the correlation between behaviors helps to discover potential health risks, such as the impact of environmental factors on physical condition.

[0113] The evaluation unit 30 is used to evaluate the above-mentioned key behavior information and generate a corresponding evaluation report, and the above-mentioned evaluation report at least includes the physical condition, mobility and environmental conditions of the above-mentioned target person.

[0114] Specifically, the key information of the behavior is input into the behavior assessment model, which may be trained based on a machine learning algorithm and is used to assess the physical condition, mobility and environmental conditions of the target person. Based on the input information, the assessment model uses an algorithm to calculate the values ​​of various assessment indicators, such as activity level, sleep quality, and comfort of the indoor environment. A detailed assessment report is generated based on the assessment results, including current status, trend analysis, and potential risks. The assessment report provides family members and caregivers with a comprehensive view of the health status of the target person, facilitating timely care decisions. It can identify abnormal changes in behavioral patterns and provide early warning of health risks, such as the risk of falling and the risk of worsening chronic diseases.

[0115] The early warning unit 40 is used to perform monitoring and intelligent equipment collaborative early warning according to the above-mentioned evaluation report combined with set early warning rules.

[0116] Specifically, based on the information in the assessment report, combined with the health records and personalized needs of the elderly, warning thresholds and conditions are set, such as long periods of inactivity and abnormally short sleep time. The system monitors the key behavioral information of the target personnel in real time, and immediately triggers the warning mechanism once the data deviates from the normal range set by the warning rules. According to the severity of the warning, the system sends an alert to the preset contacts (such as family members and medical staff), and provides detailed behavioral data and environmental information to quickly assess the situation and take action. Real-time monitoring and early warning of health risks of the elderly are achieved, and the response speed and efficiency in emergency situations are improved. Improve the accuracy and pertinence of early warnings to avoid unnecessary worries and waste of resources. Promote the continuous attention of families and caregivers to the safety of the elderly, and provide a safer living environment and timely health intervention for the elderly.

[0117] In summary, through the implementation of the above steps, the system can continuously monitor and analyze the behavioral characteristics of the elderly, generate evaluation reports and early warning information, significantly improve the intelligent level of home-based elderly care services, help to timely discover and respond to health risks, and ensure the safety and health of the elderly. The present invention can monitor and analyze the behavioral data of the elderly at home in real time and accurately, provide timely and effective information support for family members and caregivers, help improve the service quality and safety of home-based elderly care, and improve the quality of life of the elderly. At the same time, through personalized service push, it can better meet the personalized needs of the elderly and promote the precision and intelligent development of home-based elderly care services.

[0118] In this embodiment, the acquisition unit is used to obtain the effective behavior characteristic data of the target person, the target person is a person at home who needs to be monitored safely, and the effective behavior characteristic data is a characteristic vector that characterizes the behavior of the target person; the analysis unit is used to use data analysis technology to analyze the effective behavior characteristic data to obtain the key behavior information of the target person, and the key behavior information at least includes behavior patterns, activity rules and correlations between behaviors; the evaluation unit is used to evaluate the key behavior information and generate a corresponding evaluation report, and the evaluation report at least includes the physical condition, mobility and environmental conditions of the target person; the early warning unit is used to monitor and coordinate early warnings with intelligent devices based on the evaluation report combined with the set early warning rules. This application is based on real-time monitoring of multi-source data by monitoring equipment, and analyzes its activity level, activity preference, sleep pattern and other activities to generate corresponding evaluation reports, comprehensively capture the life patterns of elderly people in the home environment, summarize the unique situation of the elderly, and form a monitoring plan that can accurately analyze the abnormal state of the elderly, compare the current evaluation report with the set early warning rules, and determine whether an abnormal situation occurs and issue an early warning. Real-time and accurate monitoring and analysis of the behavior data of the elderly at home, providing timely and effective information support for family members and caregivers, will help improve the quality and safety of home-based elderly care services and enhance the quality of life of the elderly. At the same time, through personalized service push, it can better meet the personalized needs of the elderly and promote the precision and intelligent development of home-based elderly care services. This application solves the problem of low efficiency of intelligent monitoring and early warning of home safety for target personnel such as the elderly in the prior art.

[0119] In order to ensure the accuracy and reliability of the monitoring data, in an optional implementation manner, the acquisition unit includes:

[0120] An acquisition module, used to acquire the real-time behavioral characteristic data of the target person through intelligent devices, wherein the intelligent devices at least include intelligent voice assistants, intelligent cameras, smart home devices, and health monitoring sensor devices, and the health monitoring sensor devices at least include infrared sensors and millimeter wave monitoring sensors;

[0121] The preprocessing module is used to perform data cleaning, data denoising and format conversion on the above-mentioned real-time behavior feature data in sequence to obtain the above-mentioned effective behavior feature data.

[0122] In the above embodiments, it should be noted that, based on the advanced NB-IoT technology, unified standard data docking of devices is achieved to create a comprehensive and intelligent home-based elderly care safety monitoring and early warning system. The system conducts in-depth analysis of daily monitoring data collected by various home safety protection and health monitoring devices to build a data model of the elderly's home-based elderly care living environment, aiming to accurately and comprehensively analyze the elderly's home behavior patterns, so as to effectively predict the severity of the situation corresponding to the alarm information and provide strong support for rapid rescue. A series of advanced sensor devices are used, including infrared, millimeter wave and comprehensive health monitoring devices that can accurately detect changes in movement patterns and effectively identify potential fall risks. These sensors can not only monitor the physical movement status of the elderly in real time, but also collect key health indicator data such as heart rate and blood pressure. For the home environment, it is equipped with complete indoor environment detection equipment, covering gas detection, temperature detection, humidity detection and other functions, to fully ensure the safety and comfort of the home environment. At the same time, security alarm equipment is set up, such as a boundary alarm device that can monitor abnormal intrusions around the house in real time, and an SOS alarm device provides a convenient channel for the elderly to seek help in an emergency. All sensor devices use NB-IoT technology to establish efficient and stable data connections with the system, ensuring real-time data transmission and unified standard docking. The device sends the collected raw data to the cloud server, which first performs preliminary screening and cleaning of the data to remove outliers and noise interference to ensure the accuracy and reliability of the data. Based on the real-time monitoring data of the device, a database of elderly behavior characteristics is established. This database not only stores the daily activity data of the elderly, but also records and analyzes their activity levels, activity preferences, sleep patterns, etc. in detail, comprehensively capturing the details of the life patterns of the elderly in the home environment, and providing rich data support for subsequent behavior analysis and early warning judgments. The main processing logic is as follows: Figure 3 shown.

[0123] Receive and integrate real-time behavior feature data from the sensor network, and perform preliminary preprocessing and formatting. That is, clean, denoise and convert the original real-time behavior feature data to obtain effective behavior feature data so that it can be effectively processed by subsequent analysis modules.

[0124] In addition, data preprocessing technology: data processing is an important part of the monitoring process, including data cleaning, quality control, data analysis and result reporting. Through effective data processing, the accuracy and reliability of monitoring data can be ensured. In data mining, data preprocessing is to perform a series of processing work such as necessary cleaning, integration, conversion, discretization and reduction on the original data before data mining, so that it can meet the minimum specifications and standards required by the mining algorithm for knowledge acquisition research. Data cleaning is to eliminate errors or duplicate data in the data, remove noise or irrelevant data, and process missing data fields in the data; data integration combines data from multiple data sources in a unified data storage, resolves semantic ambiguity and integrates it into a consistent data storage; data selection is based on the understanding of the discovery task and the content of the data itself, looking for useful features of the expression data that depend on the discovery target, reducing the size of the data, and maximizing the amount of data while keeping the original data as much as possible. Data selection can make the regularity and potential characteristics of the data more obvious. Through data preprocessing, incomplete data can be made complete, wrong data can be corrected, redundant data can be eliminated, the required data can be selected for data integration, and the inappropriate data format can be converted to the required format, so as to achieve data type homogeneity, data format consistency, data information refinement and data storage centralization. In short, after preprocessing, not only can the data set required by the mining system be obtained, but also the cost of the mining system can be minimized and the effectiveness and comprehensibility of the mined knowledge can be improved.

[0125] In order to ensure that data is not lost and improve data query efficiency, in an optional implementation manner, the device further includes:

[0126] A classification unit is used to classify the real-time behavior characteristic data of the target person from the time dimension and the space dimension after obtaining the real-time behavior characteristic data to obtain multi-dimensional behavior characteristic data;

[0127] A compression coding unit, used to compress and optimize the coding of the multi-dimensional behavior characteristic data using data compression and coding technology to obtain compressed optimized behavior characteristic data;

[0128] The abstract storage unit is used to use a deep neural network to abstractly describe the above-mentioned compressed and optimized behavior feature data, generate corresponding feature vectors and store the above-mentioned feature vectors in a behavior feature database, and the above-mentioned feature vectors represent the above-mentioned behavior feature data in a corresponding time period.

[0129] In the above embodiment, for massive, continuous and time-space related data, a distributed data storage model based on time-space indexing is adopted. By dividing and indexing the time and space dimensions, the target person's behavior data is classified and stored according to the time series and spatial position. For example, the target person's behavior data in different rooms and different time periods are stored in corresponding time-space partitions, and each partition corresponds to a specific time range and spatial area. In this way, when querying, the relevant data partitions can be quickly located to improve the query efficiency.

[0130] Using data sharding and replication technology, large-scale data is divided into smaller fragments and multiple copies are stored on different nodes. When querying data, the system can read data fragments from multiple nodes in parallel to speed up data reading. At the same time, the replication mechanism ensures data availability and fault tolerance. When a node fails, other replicas can continue to provide services to ensure that data is not lost.

[0131] In addition, data compression and encoding technologies are introduced to compress the stored data and reduce the storage space occupied by the data. For example, lossless compression algorithms (such as LZ4, etc.) are used to compress the raw data collected by the sensor to reduce storage costs without losing data accuracy. At the same time, data encoding is optimized, such as using columnar storage formats (such as Parquet, etc.) to improve data query performance, especially when querying specific columns, data can be quickly filtered and read.

[0132] When storing the target person's behavior data, deep learning technology is used for abstract description and optimized storage. A deep neural network is used to model and extract features of the target person's behavior over a period of time. The target person's behavior sequence is input into the neural network, and the intrinsic pattern and feature representation of the behavior are learned through a multi-layer perceptron or a recurrent neural network (such as LSTM, GRU, etc.). Then, the learned feature vector is stored as an abstract representation of the behavior in that time period, replacing the traditional detailed record of the state at each time point.

[0133] For example, for the target person's one-day activity behavior data, the deep learning model extracts the main activity patterns (such as rest, activity, and dining) and related features (such as activity duration, activity frequency, and activity intensity), and combines these features into a feature vector for storage. This not only greatly reduces the update frequency and storage space occupied by the database, but also retains the key information of the behavior data, which is convenient for subsequent analysis and processing.

[0134] In addition, the storage process can also include: In the scenario of intelligent device linkage, to cope with the rapid growth of data volume over time and the high requirements for memory space and query efficiency, the innovative technology of integrating distributed storage and cloud computing is introduced. Distributed file systems (such as Ceph, etc.) are used to store data in multiple nodes to achieve horizontal expansion of data and improve storage capacity and reliability. At the same time, combined with the elastic computing resources of the cloud computing platform, the storage and computing resource configuration is dynamically adjusted according to the real-time growth and access requirements of data.

[0135] By carefully defining the attributes and methods of the class, object-oriented behavior data storage is achieved. The target person's behavior data is encapsulated into a class. The attributes of the class contain the characteristic parameters of the behavior, and the methods are used to operate and process the behavior data, such as writing, reading, updating and querying data. At the same time, the class is mapped to the database table using the object-relational mapping (ORM) framework (such as Hibernate) to achieve automatic storage and management of data. This can improve the flexibility and maintainability of data storage and facilitate developers to operate and manage data, that is, the class is written into the feature vector as a feature in the feature vector for storage.

[0136] During the storage process, data version control technology is combined to record and manage changes in the target person's behavior data. Each time the behavior data is updated, a new version is created, and the differences and change history between versions are recorded. This makes it easy to trace the data change process and restore to a specific historical version when necessary, providing strong support for data analysis and troubleshooting. For example, when the behavior data of a certain period of time is found to be abnormal, version control can be used to trace back to the previous normal version, compare and analyze the data changes, and find out the cause of the abnormality.

[0137] In order to accurately identify and understand the behavior patterns, activity patterns and correlations between different behaviors of the target person, in an optional implementation, the analysis unit includes:

[0138] A cluster analysis module, used for clustering the effective behavior feature data using a cluster analysis algorithm to identify the behavior pattern of the target person, wherein the behavior pattern at least includes a living pattern, a leisure pattern and a dining pattern;

[0139] A time analysis module, used for performing time processing on the above-mentioned effective behavior characteristic data by using a time series analysis method, so as to analyze the above-mentioned activity pattern of the above-mentioned target person, wherein the above-mentioned activity pattern at least includes the waking time, sleeping time and activity peak period of the above-mentioned target person;

[0140] The mining and analysis module is used to perform data mining on the above-mentioned effective behavior feature data using association rule mining technology to analyze the association relationship between the above-mentioned behaviors of the above-mentioned target personnel, and the above-mentioned association relationship between behaviors at least includes the relationship between sleep quality and indoor environment.

[0141] In the above embodiment, Figure 3 The data analysis module uses data mining, machine learning, deep learning and other technologies to conduct in-depth analysis of the collected data and extract key information such as the target personnel's behavior patterns, activity patterns, etc.

[0142] Using cluster analysis algorithms, the daily behavior data of target personnel are classified to identify different behavior patterns, such as living patterns, leisure patterns, dining patterns, etc. Using time series analysis methods, the time patterns of target personnel's behaviors are analyzed, such as daily wake-up time, bedtime, peak activity period, etc. Using association rule mining technology, the correlation between different behaviors is discovered, such as the relationship between sleep quality and indoor temperature and humidity.

[0143] In addition, a behavioral data set for the target personnel is constructed. The demand analysis revolves around data acquisition; data processing, including preprocessing and storage; and data application, i.e., for behavior identification and prediction. At the same time, it is clear that the functions provided in this work phase need to select the basis for relevant technologies and methods. The collected and preprocessed data is used for behavioral analysis. It is necessary to consider which behaviors can reflect the daily living habits of the target personnel; which indicators of these behaviors are predicted and analyzed, such as the start time, duration, and other behaviors that occurred during the process, whether it is necessary to further analyze the work and rest habits of the target personnel or which health indicators are related to them, etc.; the efficiency and accuracy of the behavioral prediction method used.

[0144] Since each target person may perform the same behavior multiple times a day, or multiple behaviors of multiple target persons are intertwined, a large amount of data will be generated, and the data will change in real time at different time points in the morning and evening every day.

[0145] The association rule mining algorithm based on the Apriori algorithm is used. First, the processed events are serialized, that is, the transactions are represented by numbers. Since the same event may occur multiple times during the day, the same event on the same day can be distinguished by different uppercase and lowercase numbers. The same event may occur multiple times on the same day. The first time it occurs, it is represented by lowercase numbers, and when it occurs again, it can be represented by uppercase numbers, such as going to the toilet and taking a nap. The target person's daily behavior may be getting out of bed, going to the toilet (toilet stay time), living room activities (living room sedentary time), and so on. Then, the minimum support and minimum confidence association rule data mining algorithm program based on the Apriori algorithm is implemented in Python language. The input is the event sequence with minimum support and minimum confidence, and the output is the association rule.

[0146] The frequent item sets are found from the behavior list data set through the Apriori algorithm. Frequent item sets are the frequently associated monitoring data of the target person in the multi-linkage monitoring environment of smart devices. They can reflect the daily behavior habits of the target person and can mine the daily life pattern of the target person. The frequent item sets will be in the form of {1, 2} or {1, 2, 3}, that is, the behavior group of {getting out of bed, going to the toilet} or {getting out of bed, going to the toilet, living room activities}. Then generate association rules. The association rules are the correlation between the behaviors of the target person, which can explain whether there is a correlation between the behaviors of the target person. Each rule has only one item on the right side. Therefore, one behavior or a combination of several behaviors can be associated with another behavior to achieve the behavior prediction of the target person. The style of the association rule is as follows: (getting out of bed, going to the toilet) -> living room activities, to indicate the correlation between behaviors. After getting out of bed and going to the toilet, it is predicted that living room activities will occur. If the duration of going to the toilet deviates from the daily duration, a certain warning will be issued.

[0147] In order to improve the accuracy of behavior evaluation, in an optional implementation, the evaluation unit includes:

[0148] An input-output module, used for inputting the above-mentioned key behavior information into a behavior assessment model, so as to analyze the above-mentioned key behavior information using the above-mentioned behavior assessment model, and output the behavior assessment result of the above-mentioned target person, wherein the above-mentioned behavior assessment model is a convergence model obtained by inputting sample training data into a predetermined model for iterative training using a machine learning algorithm, wherein the above-mentioned sample training data includes sample key behavior information and sample behavior assessment results corresponding to the above-mentioned sample key behavior information, wherein the above-mentioned behavior assessment results at least include an activity index, a work and rest regularity index, and an environmental adaptation index, wherein the above-mentioned activity index at least includes the number of steps walked and the activity time, the above-mentioned work and rest regularity index at least includes the sleeping time and the waking time, and the above-mentioned environmental adaptation index at least includes the indoor temperature adaptation degree and the indoor humidity adaptation degree;

[0149] A generation module is used to generate the above evaluation report based on the above behavior evaluation results.

[0150] In the above embodiment, a comprehensive behavior evaluation index system is established, including activity indicators (such as walking steps, activity time), work and rest regularity indicators (such as sleep duration, stability of waking time), environmental adaptation indicators (such as adaptation to indoor temperature and humidity), etc. A behavior evaluation model is constructed based on a machine learning algorithm, and key behavioral information of the target person is input to evaluate the behavior of the target person, such as daily activity, sleep quality, dietary regularity, etc., output the behavior evaluation results, and generate a corresponding evaluation report. The system can not only conduct real-time and accurate evaluation of the behavior of the target person, but also generate an evaluation report with practical guiding significance, so as to realize more refined home-based elderly care services and improve the quality of life and safety of the elderly.

[0151] In order to timely detect abnormal conditions of target personnel and issue an early warning, in an optional implementation manner, the early warning unit includes:

[0152] A first comparison module is used to compare the action capability in the assessment report with a behavior pattern benchmark model, wherein the behavior pattern benchmark model is a benchmark model established by analyzing at least the activity time distribution, activity area range and activity frequency characteristics of the target person in a historical time period, and the behavior pattern benchmark model is one of the behavior pattern warning rules;

[0153] A first issuing module is used to issue a first warning message when the deviation between the evaluation report and the behavioral pattern benchmark model exceeds a first set threshold, wherein the first warning message is used to remind the nursing staff that the behavior of the target person is abnormal;

[0154] A second comparison module is used to compare the environmental conditions in the assessment report with the environmental parameter dynamic warning rules, wherein the environmental parameter dynamic warning rules are dynamically adjusted according to the sensitivity of the health records and disease information of the target person to indoor environmental factors, and the indoor environmental factors include at least indoor temperature, indoor humidity and indoor dust concentration;

[0155] The second issuing module is used to issue a second warning message when the above-mentioned environmental conditions are greater than the set value of the above-mentioned indoor environmental factors in the above-mentioned environmental parameter dynamic warning rules. The above-mentioned second warning message is used to remind the above-mentioned nursing staff that there is a health risk in the environment where the above-mentioned target person is currently located.

[0156] In the above embodiment, if Figure 3As shown in the figure, the early warning module uses the latest artificial intelligence technology to achieve real-time monitoring and accurate early warning of the target personnel's behavior and environment. By building a multi-layer perception intelligent monitoring network and integrating multiple sensor data (such as motion sensors, environmental sensors, biosensors, etc.), comprehensive information on the target personnel's living status is obtained. Deep learning algorithms are used to analyze and process these multi-source data in real time, and a dynamic model of the elderly's behavior and environment is constructed to detect abnormal situations in a timely manner and issue early warnings.

[0157] The system automatically identifies and builds personalized behavior patterns by deeply learning the long-term behavior data of the target person. For example, the system analyzes the characteristics of the target person's daily activity time distribution, activity area range, activity frequency, etc., and establishes a unique behavior benchmark model for each target person. When the behavior of the elderly is significantly deviated from the benchmark model, an early warning is triggered. For example, if the target person usually has a certain activity pattern in the morning, but has not shown activities that conform to his habits for several consecutive hours, the system will automatically issue an early warning (i.e., the first early warning information mentioned above). Time series analysis and anomaly detection algorithms are introduced to monitor the time trend of the target person's behavior in real time. Not only does it focus on the absolute state of the target person's behavior, but it also analyzes whether the trend of behavior over time is abnormal. For example, by analyzing the daily activity change curve of the target person, if it is found that the activity on a certain day has dropped significantly in a short period of time and exceeds the normal fluctuation range calculated based on historical data, even if the current absolute value of the activity may not have reached the traditional early warning threshold, the system will pre-identify and warn in advance.

[0158] Using intelligent environmental perception technology, various parameters of the indoor environment are monitored in real time, and the warning threshold is dynamically adjusted according to the real-time status and historical data of the target personnel. For example, by analyzing the health records and historical disease information of the target personnel, their sensitivity to environmental factors such as temperature and humidity can be understood. For target personnel with respiratory diseases, when the indoor humidity exceeds its suitable range and lasts for a certain period of time, the system will issue more sensitive warning information in advance according to their specific health status and historical disease patterns. Combined with the Internet of Things and big data analysis, the associated warning of environmental parameters is realized. Not only does it focus on the abnormality of a single environmental parameter, but it also analyzes the potential impact of the coordinated changes between multiple environmental parameters on the health of the target personnel. For example, when the indoor temperature suddenly rises and the air quality indicators decrease at the same time (such as an increase in the concentration of harmful gases or an excess of PM2.5), the system will comprehensively consider these factors and issue a higher level of warning (i.e., the second warning information mentioned above), indicating possible health risks, because this environmental combination may have a greater impact on the respiratory system, cardiovascular system, etc. of the target personnel.

[0159] In order to improve the service quality and safety of home-based elderly care, in an optional implementation manner, the device further includes:

[0160] A response unit, for responding to the remote operation of the nursing staff after issuing a first warning message when the deviation between the assessment report and the behavioral pattern benchmark model exceeds a first set threshold, so that the nursing staff can view the real-time status of the target person, wherein the remote operation includes at least one of the following: starting a three-dimensional virtual scene of the environment of the target person through VR / AR technology, adjusting a camera angle, and starting a voice call;

[0161] Alternatively, the collaborative monitoring unit is used to automatically link multiple intelligent devices to perform collaborative monitoring and processing, and record the above-mentioned real-time status of the above-mentioned target personnel.

[0162] In the above embodiment, when the behavior data is monitored to trigger the warning rule, the system uses a variety of intelligent methods to send warning information to the preset contacts. In addition to traditional SMS, phone calls and application push, intelligent voice assistants and virtual reality (VR) / augmented reality (AR) technology are also introduced. The intelligent voice assistant can directly call the contact's phone and broadcast detailed warning content and the current status information of the elderly through voice. For example, inform the family members that "your relative has been inactive for a long time at [specific time]. The current indoor temperature is [X] degrees and the humidity is [Y]%. We have observed that the elderly is at [specific location] through the smart camera. Please pay attention to it in time." VR / AR technology is used to provide family members and caregivers with a more intuitive display of the real-time status of the elderly. After receiving the warning information, family members and caregivers can view the three-dimensional virtual scene of the elderly's environment and the real-time status of the elderly (such as a virtual model generated in real time by images captured by the camera and sensor data) through VR / AR applications on mobile phones or other devices. At the same time, they can interact with the system through gestures or voice commands to obtain more detailed information or perform remote operations, such as adjusting the camera angle and viewing historical data.

[0163] After the warning is triggered, the system realizes intelligent linkage response of multiple devices. For example, when the fall detection device finds that the elderly have fallen and triggers the fall alarm, the system will not only immediately make a phone call and send an alarm message to relatives and platform customer service personnel, but also automatically link multiple devices for collaborative processing. The smart camera will quickly adjust the angle to take a high-definition picture of the elderly's fall position, and use image recognition and analysis technology to evaluate the elderly's posture and possible injuries in real time. At the same time, through the smart lighting system installed indoors, the lights around the fall position are brightened to observe the elderly's condition more clearly.

[0164] The system will make intelligent decisions based on the data collected by the linkage devices. For example, by interacting with the data of smart bracelets or other wearable devices, the vital signs information of the elderly (such as heart rate, blood pressure, blood oxygen saturation, etc.) is obtained. If these vital signs data show normal, the system will automatically start the voice interaction module and communicate with the elderly through smart speakers or other voice devices installed nearby to confirm whether they can get up on their own or what kind of help they need. If the elderly respond that there is nothing wrong, the system will perform corresponding operations according to their instructions, such as temporarily stopping further alarm notifications, but continue to monitor the status of the elderly. If the vital signs data is abnormal, the system will automatically call the emergency number according to the preset urgency level, and send the elderly’s location information, health records and real-time monitoring data to the emergency center with one click, and notify the community medical staff to go to the scene as soon as possible to assist.

[0165] In order to improve the response speed and processing accuracy of the fall event, in an optional implementation manner, the above-mentioned collaborative monitoring unit includes:

[0166] The first control module is used to control the intelligent voice assistant to call the nursing staff and the platform customer service staff when the target person falls, and push detailed alarm information to the nursing staff and the platform customer service staff at the same time, wherein the alarm information includes the time and place of the fall and the real-time image screenshot obtained from the intelligent camera;

[0167] An image analysis module, used to automatically aim the smart camera at the position where the target person falls to the ground for high-definition photography, and use a deep learning image analysis algorithm to identify the body posture of the target person and predict the injured part;

[0168] A communication module is used to automatically communicate with the smart bracelet worn by the target person to obtain the heart rate and blood pressure data of the target person;

[0169] The second control module is used to control the smart speaker to send a voice inquiry to the target person when the heart rate and the blood pressure data are within a normal range;

[0170] A prompting module, which is used to, upon receiving a voice response from the target person within a set time, give a voice prompt to the target person to slowly stand up, and inform the nursing staff and the platform customer service staff about the response of the target person, while continuing to monitor the subsequent actions of the target person;

[0171] The first aid module is used to automatically dial the emergency number if no voice response is received from the target person within a set time or the heart rate and blood pressure data are not within the normal range, and send the health record of the target person and real-time monitoring data to the emergency center. The monitoring data at least includes the fall location, fall posture, heart rate and blood pressure data.

[0172] In the above embodiment, an application example is given: an elderly person accidentally falls while walking in the living room. At this time, the fall detection device installed in the living room quickly detects this situation and triggers a fall alarm.

[0173] The system immediately responded according to the preset process. First, the intelligent voice assistant called the family members and the platform customer service staff, and pushed detailed alarm information to their mobile applications, including the time and location of the fall and the real-time image screenshots obtained from the camera. Then, the smart camera in the living room was automatically linked to the position where the elderly fell to take high-definition photos, and the deep learning image analysis algorithm was used to quickly identify the elderly’s body posture and possible injured parts. For example, by analyzing the elderly’s body posture and facial expressions, it was initially determined whether there were signs of head injuries or fractures.

[0174] At the same time, the system communicates with the smart bracelet worn by the elderly to obtain their real-time heart rate and blood pressure data. If the heart rate and blood pressure data are normal, the system will send a voice inquiry through the smart speaker: "Hello, do you need help? Can you get up by yourself?" If the elderly answer that they can get up by themselves within the set time, the system will voice prompt the elderly to get up slowly, and inform the family members and platform customer service staff of the elderly's response, while continuing to monitor the elderly's follow-up actions. If the elderly do not respond within the set time or the vital signs data are abnormal, the system will immediately call the emergency number and send the pre-prepared health records of the elderly (including medical history, allergy history, etc.) and real-time monitoring data (such as fall location, posture, heart rate, blood pressure, etc.) to the emergency center. In addition, the system will notify community medical staff to bring necessary first aid equipment to the scene as soon as possible to provide initial assistance before the emergency personnel arrive.

[0175] This early warning mechanism of multi-device linkage and intelligent decision-making has greatly improved the response speed and accuracy of handling falls. By communicating with the elderly in a timely manner to confirm their condition, unnecessary panic and false alarms are avoided, and ineffective intervention by guardian family members and platform employees due to uncertain situations is also reduced. In cases where emergency rescue is really needed, the system can quickly and accurately provide comprehensive information to emergency centers and community medical staff, buying precious treatment time for the elderly and effectively reducing the serious consequences of falls.

[0176] The above-mentioned personnel home safety monitoring device includes a processor and a memory, and the above-mentioned units are stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize the corresponding functions. The above-mentioned modules are all located in the same processor; or, the above-mentioned modules are located in different processors in any combination.

[0177] The processor includes a kernel, which retrieves the corresponding program unit from the memory. One or more kernels can be set, and the problem of low efficiency of intelligent monitoring and early warning for home safety of target personnel such as the elderly in the prior art can be solved by adjusting kernel parameters.

[0178] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0179] An embodiment of the present invention provides a computer-readable storage medium, which includes a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute the above-mentioned person home safety monitoring method.

[0180] An embodiment of the present invention provides a processor, which is used to run a program, wherein the method for monitoring the safety of people at home is executed when the program is running.

[0181] An embodiment of the present invention provides a home security monitoring system, the home security monitoring system includes a processor, a memory, and a program stored in the memory and executable on the processor, and when the processor executes the program, at least the following steps are implemented:

[0182] Step S201, obtaining effective behavior characteristic data of a target person, the target person being a person at home who needs to be monitored for safety, and the effective behavior characteristic data being a characteristic vector that characterizes the behavior of the target person;

[0183] Step S202, using data analysis technology to analyze the above-mentioned effective behavior feature data to obtain the above-mentioned key behavior information of the target person, and the above-mentioned key behavior information at least includes behavior patterns, activity rules and correlation relationships between behaviors;

[0184] Step S203, evaluating the above-mentioned key behavior information to generate a corresponding evaluation report, wherein the above-mentioned evaluation report at least includes the physical condition, mobility and environment of the above-mentioned target person;

[0185] Step S204, monitoring and intelligent equipment collaborative warning are performed based on the above assessment report and set warning rules.

[0186] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program for initializing at least the following steps of a method for monitoring the safety of a person at home.

[0187] Obviously, those skilled in the art should understand that the above modules or steps of the present invention can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order than here, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present invention is not limited to any specific combination of hardware and software.

[0188] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0189] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0190] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0191] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0192] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0193] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0194] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0195] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0196] From the above description, it can be seen that the above embodiments of the present application achieve the following technical effects:

[0197] 1) The method for monitoring the safety of people at home in the present application, first, obtain the effective behavior characteristic data of the target person, the target person is the person at home who needs to be monitored safely, and the effective behavior characteristic data is the characteristic vector that characterizes the behavior of the target person; then, use data analysis technology to analyze the effective behavior characteristic data to obtain the key behavior information of the target person, the key behavior information at least includes the behavior pattern, activity pattern and the correlation between behaviors; then, evaluate the key behavior information to generate a corresponding evaluation report, the evaluation report at least includes the physical condition, mobility and environmental conditions of the target person; finally, monitor and coordinate warnings with intelligent equipment according to the evaluation report combined with the set warning rules. The present application is based on real-time monitoring of multi-source data by monitoring equipment, and analyzes its activity level, activity preference, sleep pattern and other activities to generate corresponding evaluation reports, comprehensively capture the life patterns of elderly people in the home environment, summarize the unique situation of the elderly, and form a monitoring plan that can accurately analyze the abnormal state of the elderly, compare the current evaluation report with the set warning rules, determine whether an abnormal situation occurs and issue a warning. Real-time and accurate monitoring and analysis of the behavior data of the elderly at home, providing timely and effective information support for family members and caregivers, will help improve the quality and safety of home-based elderly care services and enhance the quality of life of the elderly. At the same time, through personalized service push, it can better meet the personalized needs of the elderly and promote the precision and intelligent development of home-based elderly care services. This application solves the problem of low efficiency of intelligent monitoring and early warning of home safety for target personnel such as the elderly in the prior art.

[0198] 2) The home safety monitoring device of the present application includes an acquisition unit for acquiring effective behavior characteristic data of the target person, the target person is a home person who needs to be monitored safely, and the effective behavior characteristic data is a characteristic vector that characterizes the behavior of the target person; an analysis unit for analyzing the effective behavior characteristic data using data analysis technology to obtain the key behavior information of the target person, the key behavior information at least includes behavior patterns, activity rules and correlations between behaviors; an evaluation unit for evaluating the key behavior information to generate a corresponding evaluation report, the evaluation report at least includes the physical condition, mobility and environmental conditions of the target person; an early warning unit for monitoring and intelligent equipment collaborative early warning based on the evaluation report combined with the set early warning rules. The present application is based on real-time monitoring of multi-source data by monitoring equipment, and analyzing its activity level, activity preference, sleep pattern and other activities to generate a corresponding evaluation report, comprehensively capture the life pattern of the elderly in the home environment, summarize the unique situation of the elderly, and form a monitoring plan that can accurately analyze the abnormal state of the elderly, compare the current evaluation report with the set early warning rules, determine whether an abnormal situation occurs and issue an early warning. Real-time and accurate monitoring and analysis of the behavior data of the elderly at home, providing timely and effective information support for family members and caregivers, will help improve the quality and safety of home-based elderly care services and enhance the quality of life of the elderly. At the same time, through personalized service push, it can better meet the personalized needs of the elderly and promote the precision and intelligent development of home-based elderly care services. This application solves the problem of low efficiency of intelligent monitoring and early warning of home safety for target personnel such as the elderly in the prior art.

[0199] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for monitoring the safety of people at home, characterized in that: include: Acquire effective behavior characteristic data of a target person, the target person being a person at home who needs to be monitored for safety, the effective behavior characteristic data being a characteristic vector that characterizes the behavior of the target person; Using data analysis technology to analyze the effective behavior feature data to obtain key behavior information of the target person, the key behavior information at least includes behavior patterns, activity rules and correlations between behaviors; Evaluate the key behavior information to generate a corresponding evaluation report, the evaluation report at least including the physical condition, mobility and environment of the target person; Based on the assessment report and the set warning rules, monitoring and intelligent equipment collaborative warning are carried out.

2. The method according to claim 1, characterized in that Obtain effective behavioral characteristic data of the target person, including: Acquire the real-time behavior characteristic data of the target person through the intelligent device, wherein the intelligent device at least includes an intelligent voice assistant, an intelligent camera, a smart home device and a health monitoring sensor device, and the health monitoring sensor device at least includes an infrared sensor and a millimeter wave monitoring sensor; The real-time behavior feature data is sequentially cleaned, denoised and formatted to obtain the effective behavior feature data.

3. The method according to claim 1, characterized in that After acquiring the real-time behavior characteristic data of the target person, the method further includes: Classifying the real-time behavior characteristic data from a time dimension and a space dimension to obtain multi-dimensional behavior characteristic data; Using data compression and encoding technology to compress and optimize the multi-dimensional behavior characteristic data to obtain compressed and optimized behavior characteristic data; A deep neural network is used to abstractly describe the compressed and optimized behavior feature data, generate corresponding feature vectors, and store the feature vectors in a behavior feature database, wherein the feature vectors represent the behavior feature data within a corresponding time period.

4. The method according to claim 1, characterized in that The effective behavior characteristic data is analyzed using data analysis technology to obtain key behavior information of the target person, including: Using a cluster analysis algorithm to perform clustering processing on the effective behavior feature data to identify the behavior pattern of the target person, the behavior pattern at least including a living pattern, a leisure pattern and a dining pattern; Using a time series analysis method to perform time processing on the effective behavior feature data to analyze the activity pattern of the target person, the activity pattern at least includes the target person's waking time, sleeping time and activity peak period; The effective behavior feature data is mined using association rule mining technology to analyze the association relationship between the behaviors of the target person, and the association relationship between the behaviors at least includes the relationship between sleep quality and indoor environment.

5. The method according to claim 1, characterized in that Evaluate the key behavior information to generate a corresponding evaluation report, including: Input the behavior key information into a behavior assessment model, so as to analyze the behavior key information using the behavior assessment model and output the behavior assessment result of the target person, wherein the behavior assessment model is a convergence model obtained by inputting sample training data into a predetermined model for iterative training using a machine learning algorithm, wherein the sample training data includes sample behavior key information and sample behavior assessment results corresponding to the sample behavior key information, wherein the behavior assessment results include at least an activity index, a work and rest regularity index and an environmental adaptation index, wherein the activity index includes at least the number of steps walked and the activity time, the work and rest regularity index includes at least the sleeping time and the waking time, and the environmental adaptation index includes at least the indoor temperature adaptation degree and the indoor humidity adaptation degree; The evaluation report is generated according to the behavior evaluation result.

6. The method according to claim 1, characterized in that The set warning rules include behavior pattern warning rules and environmental parameter dynamic warning rules. Monitoring and intelligent equipment collaborative warning are performed according to the evaluation report combined with the set warning rules, including: Comparing the action capability in the assessment report with a behavior pattern benchmark model, wherein the behavior pattern benchmark model is a benchmark model established by analyzing at least the activity time distribution, activity area range, and activity frequency characteristics of the target person in a historical time period, and the behavior pattern benchmark model is one of the behavior pattern warning rules; When the deviation between the assessment report and the behavioral pattern benchmark model exceeds a first set threshold, a first warning message is issued, wherein the first warning message is used to remind the nursing staff that the behavior of the current target person is abnormal; Comparing the environmental conditions in the assessment report with the dynamic early warning rules for environmental parameters, the dynamic early warning rules for environmental parameters are dynamically adjusted according to the sensitivity of the health records and disease information of the target person to indoor environmental factors, the indoor environmental factors including at least indoor temperature, indoor humidity and indoor dust concentration; When the environmental condition is greater than the set value of the indoor environmental factor in the environmental parameter dynamic warning rule, a second warning message is issued, and the second warning message is used to remind the caregiver that there is a health risk in the environment where the target person is currently located.

7. The method according to claim 6, characterized in that When the deviation between the assessment report and the behavior pattern benchmark model exceeds a first set threshold, after issuing a first warning message, the method further includes: Responding to the remote operation of the caregiver so that the caregiver can view the real-time status of the target person, the remote operation comprising at least one of the following: starting a three-dimensional virtual scene of the target person's environment through VR / AR technology, adjusting the camera angle and starting a voice call; Alternatively, multiple intelligent devices are automatically linked to perform collaborative monitoring and processing to record the real-time status of the target person.

8. The method according to claim 7, characterized in that Automatically link multiple intelligent devices to perform collaborative monitoring and processing, and record the real-time status of the target personnel, including: In the event that the target person falls, the intelligent voice assistant is controlled to call the caregiver and the platform customer service personnel, and at the same time, detailed alarm information is pushed to the caregiver and the platform customer service personnel, and the alarm information includes the time and place of the fall and the real-time image screenshot obtained from the intelligent camera; Automatically aiming the smart camera at the position where the target person falls to take high-definition photos, and using a deep learning image analysis algorithm to identify the body posture of the target person and predict the injured part; Automatically communicate with the smart bracelet worn by the target person to obtain the heart rate and blood pressure data of the target person; When the heart rate and the blood pressure data are within a normal range, controlling the smart speaker to issue a voice inquiry to the target person; If a voice response from the target person is received within the set time, the target person is prompted by voice to stand up slowly, and the nursing staff and the platform customer service staff are informed of the target person's response, while continuing to monitor the target person's subsequent actions; If no voice response is received from the target person within the set time or the heart rate and the blood pressure data are not within the normal range, an emergency call will be automatically made, and the health record of the target person and the real-time monitoring data will be sent to the emergency center. The monitoring data includes at least the fall location, fall posture, the heart rate and the blood pressure data.

9. A device for monitoring the safety of people at home, characterized in that: The device comprises: An acquisition unit, used to acquire effective behavior characteristic data of a target person, wherein the target person is a person at home who needs to be monitored for safety, and the effective behavior characteristic data is a characteristic vector that characterizes the behavior of the target person; An analysis unit, configured to use data analysis technology to perform data analysis on the effective behavior feature data to obtain key behavior information of the target person, wherein the key behavior information at least includes behavior patterns, activity rules, and correlations between behaviors; An evaluation unit, configured to evaluate the key behavior information and generate a corresponding evaluation report, wherein the evaluation report at least includes the physical condition, mobility and environment of the target person; The early warning unit is used to monitor and coordinate early warning with intelligent equipment according to the evaluation report and set early warning rules.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the method according to any one of claims 1 to 8.

Images