Risk perception interaction method and system for elderly users in smart home environment

Abstract

The invention discloses an elderly user risk perception interaction method and system in a smart home environment, and relates to the technical field of data processing, and the method comprises the steps: connecting smart homes, analyzing the data features of the connected smart homes, and constructing a perception database; obtaining risk perception demands of the elderly users, constructing a risk perception demand library, carrying out risk demand analysis, and obtaining a risk analysis link according to an identification causal chain of the risk demands; and performing correlation analysis on the risk analysis link and the perception database, identifying a mapping relationship between the perception data and the risk analysis link, projecting the mapping relationship to the risk analysis link for risk identification, and obtaining a risk identification result. The technical problem that the prior art lacks accurate and real-time monitoring and personalized early warning of the health risk of the elderly user is solved, and the technical effects of real-time accurate recognition and personalized early warning of the health risk of the elderly user are achieved through multi-device data analysis and association analysis.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a risk perception interaction method and system for elderly users in a smart home environment. Background Art

[0002] With the advent of an aging society, the elderly are facing health problems such as physical decline and chronic diseases, which make them more vulnerable to health risks such as falls and heart disease. However, existing smart home systems focus more on device control and environmental monitoring, lacking accurate perception and real-time analysis of the health status of the elderly, especially in terms of multi-device data integration and dynamic health risk warning. Although traditional health monitoring methods can monitor physiological data, they are usually unable to correlate and analyze multiple data sources, and it is difficult to provide personalized and real-time risk warnings. Summary of the invention

[0003] The present application provides a risk perception interaction method and system for elderly users in a smart home environment, which is used to solve the technical problem that the existing technology lacks accurate, real-time monitoring and personalized early warning of health risks for elderly users.

[0004] The first aspect of the present application provides a risk perception interaction method for elderly users in a smart home environment, the method comprising: connecting to smart homes, parsing data features of each connected smart home, and building a perception database; obtaining risk perception needs of elderly users and building a risk perception needs library; performing risk demand analysis based on the risk perception needs library, and obtaining a risk analysis link according to the risk demand identification causal chain; performing correlation analysis based on the risk analysis link and the perception database, identifying a mapping relationship between perception data and the risk analysis link, projecting the perception data in the perception database according to the mapping relationship to the risk analysis link associated with it, and performing risk identification according to the perception data through the risk analysis link. The risk identification result is obtained; wherein, risk demand analysis is performed according to the risk perception demand library, and a risk analysis link is obtained according to the identification causal chain of risk demand, including: taking risk perception demand as the top event, analyzing the accident sub-event of the top event based on historical data, and fitting the causal function relationship between the accident sub-event and the top event; analyzing the accident factor of the accident sub-event based on historical data, and fitting the causal function relationship between the accident sub-event and the accident factor; finding the minimum cut set of accident factor combination, when the accident factor combination in the minimum cut set is established, the accident sub-event must occur, and the risk analysis link is obtained based on the accident factor combination-accident sub-event-top event and their causal function relationship.

[0005] The second aspect of the present application provides a risk perception interaction system for elderly users in a smart home environment, the system comprising: a data feature parsing module, the data feature parsing module is used to connect to smart homes, parse the data features of each connected smart home, and build a perception database; a risk perception demand library construction module, the risk perception demand library construction module is used to obtain the risk perception needs of elderly users and build a risk perception demand library; a risk demand parsing module, the risk demand parsing module is used to perform risk demand parsing according to the risk perception demand library, and obtain a risk parsing link according to the risk demand identification causal chain; a risk identification module, the risk identification module is used to perform correlation analysis with the perception database according to the risk parsing link, identify the mapping relationship between the perception data and the risk parsing link, and project the perception data in the perception database according to the mapping relationship to the risk parsing link. In the associated risk analysis link, risk identification is performed according to the perception data through the risk analysis link to obtain risk identification results; the risk demand analysis module is also used to execute the following steps: risk demand analysis is performed according to the risk perception demand library, and a risk analysis link is obtained according to the identification causal chain of risk demand, including: taking risk perception demand as the top event, analyzing the accident sub-event of the top event based on historical data, and fitting the causal function relationship between the accident sub-event and the top event; analyzing the accident factor of the accident sub-event based on historical data, and fitting the causal function relationship between the accident sub-event and the accident factor; finding the minimum cut set of accident factor combination, when the accident factor combination in the minimum cut set is established, the accident sub-event must occur, and the risk analysis link is obtained based on the accident factor combination-accident sub-event-top event and their causal function relationship.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: The risk perception interaction method and system for elderly users in a smart home environment provided by the present application relate to the field of data processing technology. By connecting smart home devices and parsing data features, a perception database is constructed, the risk needs of elderly users are obtained and a demand library is established, risks are parsed based on the demand library, a risk analysis link is constructed, and the perception data is mapped to the link through correlation analysis for risk identification, and finally an accurate risk identification result is obtained. The technical problem that the existing technology lacks accurate, real-time monitoring and personalized early warning of health risks of elderly users is solved, and the technical effect of realizing real-time accurate identification and personalized early warning of health risks of elderly users through multi-device data parsing and correlation analysis is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0008] Figure 1 A schematic diagram of a flow chart of a risk perception interaction method for elderly users in a smart home environment provided by an embodiment of the present application; Figure 2 Schematic diagram of the structure of the risk perception interaction system for elderly users in a smart home environment provided in an embodiment of the present application.

[0009] Explanation of the accompanying drawings: data feature analysis module 11, risk perception demand library construction module 12, risk demand analysis module 13, risk identification module 14. DETAILED DESCRIPTION

[0010] The present application provides a risk perception interaction method and system for elderly users in a smart home environment, which is used to solve the technical problem that the existing technology lacks accurate, real-time monitoring and personalized early warning of health risks for elderly users.

[0011] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0012] It should be noted that the terms "first", "second", etc. in the specification 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 sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices.

[0013] Embodiment 1, as Figure 1 As shown, the present application provides a risk perception interaction method for elderly users in a smart home environment, the method comprising: P10: Connect smart homes, analyze the data characteristics of each connected smart home, and build a perception database.

[0014] Furthermore, step P10 of the embodiment of the present application further includes: P11: Analyze the type, content and frequency of the collected data of the connected smart homes to obtain the data characteristics of each smart home; P12: Construct the perception database according to the mapping relationship between the connected smart homes and the data characteristics.

[0015] Optionally, the connection and data collection of smart home devices are the basis for realizing risk perception of elderly users. In order to build an effective perception database, in-depth analysis and data processing of connected smart home devices are required. For example, when a smart audio device is started for the first time, it needs to be bound through the network configuration and device number (such as downloading the corresponding APP through the mobile phone, entering the device number or scanning the QR code for binding). Once the device is connected successfully, the audio system can capture the user's instructions through the voice recognition function, and analyze and identify keywords such as "turn on AI assistant" or "turn off AI assistant". The system recognizes these voice commands through voice keyword extraction technology, and uploads the device usage data (such as the user's voice commands, device operation logs) to the perception database, further enhancing the collection ability of user behavior data. In this way, the system can continuously collect users' voice commands, activity information, environmental data, etc. to form a comprehensive perception database.

[0016] First, the data collected by all connected smart home devices are analyzed in terms of type, content, and frequency. Each device generates different types of data. Common data types include health sign data, motion location data, voice data, and environmental monitoring data. Health sign data comes from smart health devices, such as blood pressure monitors and heart rate monitors. The content of these data includes physiological indicators such as blood pressure, heart rate, and body temperature, which directly affect the health status of elderly users. Motion location data is generated by smart sensors or wearable devices. This type of data can monitor the user's activity status, location changes, and events such as falls in real time. Voice data usually comes from smart speakers or voice assistants, which can detect whether the user has made an emergency call or has an abnormal conversation. Environmental monitoring data comes from temperature and humidity sensors and air quality monitoring equipment. These data reflect the impact of the indoor environment on elderly users.

[0017] After analyzing the data type, the next focus is to analyze the content of each type of data, that is, the specific indicators and units of the data. For example, the data content collected by a blood pressure monitor may be "systolic pressure" and "diastolic pressure" values, while an environmental monitor may collect "temperature" and "humidity" information. Through the analysis of these contents, the system can effectively understand the data characteristics of each device. At the same time, the frequency of the data must also be analyzed. Different devices have different data collection frequencies. Health devices may collect data once a minute, while environmental monitoring devices may update data once an hour. These frequency information helps determine the real-time and comprehensiveness of data collection, ensuring that the system can identify potential risks in a timely manner.

[0018] By adopting the Internet of Things (IoT), wireless sensor networks (WSN), and Bluetooth, smart devices can efficiently transmit data to data centers or local processing units, realizing real-time data collection and transmission.

[0019] Next, we build a perception database by establishing mapping relationships. This database will be used to store and manage the data collected from various smart home devices. During the construction process, we first need to establish mapping relationships based on the data types and characteristics collected by each device to ensure that each type of data can be correctly identified and classified. The establishment of mapping relationships is a crucial process that helps match devices and data. For example, the data collected by blood pressure monitoring devices will be mapped to the "Health Signs" category, while the data collected by motion sensors will be mapped to the "Action Positioning" category. These mapping relationships ensure that the source of each data item can be clearly identified and can be used correctly in subsequent analysis.

[0020] In terms of technical implementation, the perception database needs to support the storage and efficient query of large-scale data. Relational databases (such as MySQL, PostgreSQL) or non-relational databases (such as MongoDB, InfluxDB) can be used for data management. For the needs of real-time data collection and rapid processing, streaming database technology (such as Apache Kafka, Apache Flink) can be used to achieve real-time storage and processing of data. At the same time, the use of data labeling technology can help identify and classify various types of data in the database for subsequent query and analysis.

[0021] By building a perception database, the collected data from various devices can be centrally stored, and the source of each data can be accurately identified according to the mapping relationship. The construction of the perception database not only provides a data warehouse for the system, but also provides support for subsequent risk identification and analysis, ensuring that the system monitors potential risks in the living environment of elderly users in real time and conducts effective risk identification and response.

[0022] Furthermore, step P12 of the embodiment of the present application also includes: P12-1: Taking the data type of the collected data as the first classification feature, perform primary classification on the smart home and its collected data to obtain primary classification data; P12-2: Taking the data content of the collected data as the second classification feature, perform secondary classification on the primary classification data to obtain secondary classification data; P12-3: Taking the collection frequency of the collected data as the third classification feature, perform time-aligned classification on the secondary classification data to obtain tertiary classification data; P12-4: Classify, grade and integrate the smart home and its collected data according to the classification results and hierarchical relationship of the primary classification data, secondary classification data and tertiary classification data to construct the perception database.

[0023] It should be understood that in the further refinement process, the collected data can be classified and graded in more detail, so as to more effectively build the perception database.

[0024] First, all smart home devices and the data they collect need to be classified according to the type of data collected. Different types of smart home devices collect different types of data. Common data types include health sign data (such as blood pressure, heart rate), environmental monitoring data (such as temperature and humidity, air quality), action positioning data (such as user location, movement trajectory), voice collection data (such as voice commands, conversation records), etc. According to the type of data, it can be classified into different categories such as health, environment, positioning, voice, etc. The purpose of this first-level classification is to provide a clear structure for subsequent data analysis, ensuring that each type of data can be grouped by type for subsequent processing and analysis.

[0025] Technically, this classification process can be achieved through data labeling and classification algorithms. By giving each data item a clear label, the type of data can be quickly identified and classified into the corresponding category. For example, machine learning algorithms (such as decision trees, K-means clustering, etc.) are used to automatically identify data types and assign data to corresponding categories.

[0026] Then, after completing the primary classification, the primary classification data is classified into secondary categories based on the content of the collected data. For example, for health sign data, the content may include specific physiological parameters such as blood pressure, heart rate, and body temperature; for environmental monitoring data, the content may include indicators such as temperature, humidity, and air quality. By using the data content as the second classification feature, the data in the primary classification can be further refined to obtain more detailed secondary classification data.

[0027] This secondary classification helps to analyze the specific content of the data more accurately and helps the system identify and understand the correlation between data of different contents. For example, the content classification of health sign data can help the system distinguish between "blood pressure" and "heart rate" data, thereby providing more accurate risk assessment in subsequent health analysis. The implementation of secondary classification usually relies on natural language processing (NLP) technology or data grouping algorithms to ensure that each piece of data can be accurately classified by parsing and labeling the data content.

[0028] Furthermore, the frequency of data collection is used as the third classification feature to perform time-aligned classification on the secondary classification data. Different types of equipment collect data at different time intervals. For example, health monitoring equipment may collect data once a minute, while environmental monitoring equipment may update data once an hour. Classification based on collection frequency can group data according to its time granularity, ensuring that the system can identify the timeliness of each type of data.

[0029] The construction of three-level classification data is crucial to ensure data consistency and comparability. Time alignment ensures that data of different frequencies can be compared within the same time window, facilitating data fusion and comprehensive analysis across devices. Technically, this step can be accomplished through time series analysis and timestamp matching technology to ensure that each data item can be accurately aligned with other data on the timeline.

[0030] Finally, the classification results of the first-level classification, second-level classification, and third-level classification are integrated to build a complete perception database according to the hierarchical relationship. In this process, the perception database will not only be a simple storage structure, but also reflect the hierarchical relationship of data classification. For example, in the database, health data will be subdivided into subcategories according to their content (such as blood pressure, heart rate), and the data in these subcategories will be sorted according to time to form a complete data hierarchy. This hierarchical classification method helps to organize, store and access data efficiently, and also facilitates subsequent data retrieval and analysis.

[0031] In terms of technical implementation, a hierarchical database structure can be used to implement classified storage of data through hierarchical indexes and relational database design. Each level of classification will form an independent table or data set, and the levels are connected through foreign keys or reference relationships. In addition, data warehouse technology can also be used to support the storage and analysis of large-scale data.

[0032] This classification, grading and integration method ensures the integrity and real-time nature of the data, while improving the efficiency of data query and processing, providing a reliable foundation for subsequent risk identification and intelligent decision-making.

[0033] P20: Obtain the risk perception needs of elderly users and build a risk perception needs database.

[0034] Furthermore, step P20 of the embodiments of the present application further includes: P21: Enter the identity information and health data of the elderly target user; P22: Match the identity information and health data of the elderly target user with a preset risk perception library to obtain recommended risk perception requirements; P23: Select and set or customize according to the recommended risk perception requirements to obtain the risk perception requirements of the elderly user.

[0035] Specifically, by obtaining the risk perception requirements of the elderly user and constructing a risk perception requirement library, customized risk perception services can be provided for the elderly user.

[0036] First of all, it is necessary to collect and enter the identity information and health data of the elderly user. The identity information includes the basic personal information of the user, such as name, age, gender, contact information, etc. The health data is the physiological state information of the elderly user, including but not limited to blood pressure, heart rate, body temperature, weight, exercise ability, etc. Exemplarily, the elderly user downloads and installs the intelligent APP, registers and binds the intelligent audio device. During the registration process, the user needs to fill in personal health information (such as data of blood pressure, heart rate, weight, etc.), and further verify the health status through intelligent health monitoring devices (such as sphygmomanometer, heart rate monitor, etc.). Technically, the entry of identity information and health data can be interactively input through a mobile application, a web page or a voice assistant, while using data encryption technology to ensure the security of user privacy. During the collection process of health data, the Internet of Things technology is used to synchronize data with intelligent devices, and cloud storage is used to ensure data backup and sharing.

[0037] Furthermore, by analyzing the identity information and health data of the elderly target user and combining with the preset risk perception library, corresponding risk perception requirements are recommended for the user. The risk perception library contains various common types of risk perception requirements, such as fall risk, health monitoring, environmental change, etc. The system will match these requirements according to the specific health status of the elderly user (such as hypertension, heart disease, diabetes, etc.), and automatically recommend relevant risk perception requirements. For example, the system intelligently recommends risk perception requirements based on this data and the user-provided living habits (such as exercise frequency, diet preferences, etc.). For example, if the user's blood pressure is high, the system will automatically recommend hypertension monitoring; if the user's exercise volume is small, the system may recommend fall monitoring and location tracking functions. The elderly can also customize the health risks to be monitored in the APP according to their own needs, such as "health monitoring", "fall monitoring", etc. All requirements and data will be stored in the risk perception requirement library for subsequent use.

[0038] For example, machine learning algorithms (such as decision trees, support vector machines, etc.) and big data analysis can be used to analyze the user's health data, and the health status of elderly users can be matched with preset risk perception needs through matching algorithms. For example, if the user's health data shows that he / she has a higher risk of falling, the system will automatically recommend "fall warning" and "location tracking" and other needs. At the same time, it is also possible to make predictions based on the user's historical health data and living habits (such as whether they have exercise habits, whether they have long-term chronic diseases, etc.) to provide more accurate risk identification.

[0039] Furthermore, users can select and set risk perception requirements based on the system's recommended risk perception requirements, or customize them based on their own specific needs. The system will provide a selection interface where users can activate or modify risk perception requirements through simple operations (such as button selection, slider adjustment, etc.). For example, users can choose to turn on the "fall monitoring" or "health monitoring" functions, or adjust the monitoring frequency, risk warning threshold, etc. as needed.

[0040] In addition, users can also customize designs for certain specific needs, such as setting specific health indicator monitoring, environmental change warnings and other functions. For elderly users with special health conditions, one-click activation of the "Emergency Help" or "Smart Call" function can also be provided to ensure timely help in an emergency. In order to achieve customized settings, a graphical interface and human-computer interaction design can be used to enable users to intuitively configure various functions. Through voice input or gesture operation, users can easily customize without complicated operations.

[0041] Through this customized service, the smart home system can effectively improve the life safety of elderly users and provide users with more suitable health management solutions.

[0042] P30: Perform risk demand analysis based on the risk perception demand library, and obtain a risk analysis link according to the risk demand identification causal chain.

[0043] Furthermore, step P30 of the embodiment of the present application further includes: P31: Taking risk perception demand as the top event, analyze the accident sub-events of the top event based on historical data, and fit the causal function relationship between the accident sub-event and the top event; P32: Analyze the accident factors of the accident sub-events based on historical data, and fit the causal function relationship between the accident sub-event and the accident factors; P33: Find the minimum cut set of accident factor combinations. When the accident factor combination in the minimum cut set is established, the accident sub-event is bound to occur. Based on the accident factor combination-accident sub-event-top event and their causal function relationship, the risk analysis link is obtained.

[0044] It should be understood that the risk perception demand library that has been constructed is analyzed in detail, and a specific risk analysis link is constructed based on the identification causal chain of these risk demands. For example, assuming that the system receives a "fall monitoring" demand, the system will first analyze "fall" as the main risk event, and analyze potential triggering factors based on the user's health data (such as whether there is a balance disorder, mobility) and environmental data (such as whether the ground is slippery). The system will analyze the causal relationship of the fall risk based on the causal chain analysis. For example, a slippery ground may be a key factor leading to a fall, and limited activity of the elderly (such as unstable walking) is also an important risk factor. Through these analyses, the system establishes a fall risk analysis link, which contains multiple causal chains, such as "slippery environment → fall", "weak motor ability of the elderly → fall", etc.

[0045] Specifically, first, the risk perception demand is regarded as a top event (i.e., a main event). This top event can be a certain risk faced by elderly users, such as "fall risk" or "health problem warning". Then, based on historical data, the accident sub-events related to the top event are analyzed. These sub-events are potential factors that lead to the occurrence of the top event.

[0046] By analyzing historical data, the system can identify a series of sub-events related to the top event. For example, for "fall risk", sub-events may include "unsteady walking", "slipping on the ground" or "fatigue". These sub-events will be fitted with causal function relationships based on historical data, that is, through statistical analysis and machine learning models, the causal function between these sub-events and the top event is established. The methods used in this process may include regression analysis, Bayesian networks, or causal inference models to analyze how sub-events affect the probability of occurrence of top events. Exemplarily, data mining techniques and machine learning algorithms, such as random forests, support vector machines (SVM) and other models are used to learn the causal relationship between top events and sub-events. At the same time, Bayesian networks can provide a clear way to express and reason about causal relationships.

[0047] Next, analyze the specific accident factors of each accident sub-event. Accident factors refer to those factors that directly lead to accidents. For example, for the accident sub-event of "unsteady walking", accident factors may include "lack of light", "uneven ground", "wearing inappropriate shoes", etc. By analyzing historical data, these accident factors are identified, and causal function relationships are further established, that is, how these factors affect the occurrence of accident sub-events. For example, through the "unsteady walking" event in historical data, the system can identify that "wearing inappropriate shoes" or "slippery ground" has a direct impact on the occurrence of this event. For example, causal inference models (such as structural equation models SEM, causal diagrams, etc.) and data regression analysis are used to establish causal relationship models between accident factors and accident sub-events. These models can reveal the relationship between potential factors and provide a basis for subsequent risk analysis.

[0048] Furthermore, based on the causal function relationship, the minimum cut set of the accident factor combination is solved. The minimum cut set refers to a set of minimum accident factors. When these factor combinations are established, the accident sub-event will inevitably occur. By identifying these minimum cut sets and solving the minimum cut sets, the most direct and core factor combination that leads to the occurrence of the accident sub-event can be clearly identified. For example: First, based on historical data and expert knowledge, list all possible accident factors that may affect the occurrence of accident sub-events. For example, for the accident sub-event of "falling while walking in the bathroom", possible accident factors include "slippery floor", "dim lighting", "user mobility difficulties", "too high bathroom threshold", etc. Combine these accident factors to form all possible accident factor combinations. For example, combination 1 can be "slippery floor + dim lighting", combination 2 can be "slippery floor + user mobility difficulties", combination 3 can be "dim lighting + user mobility difficulties", and combination 4 can be "slippery floor + bathroom threshold is too high".

[0049] Next, through historical data analysis, expert evaluation or simulation experiments, each combination of accident factors is evaluated to determine whether the combination will inevitably lead to the occurrence of accident sub-events. For example, through the analysis of historical data, it is found that when "slippery ground" and "user mobility difficulties" occur at the same time, the probability of falling increases significantly, while the probability of falling is lower when "dim lighting" occurs alone. Through evaluation, those accident factor combinations that, once established, have a probability of occurrence of accident sub-events greater than the probability threshold (such as 80%) are screened out. These combinations are the minimum cut sets.

[0050] Furthermore, according to the probability of the event, determine which accident factor combinations are the minimum combinations that will inevitably cause the accident sub-events to occur. For example, for the accident sub-event "falling down while walking in the bathroom", after evaluation, it was found that the probability of falling down in combination 1 (slippery ground + dim light) is 50%, the probability of falling down in combination 2 (slippery ground + user's mobility difficulties) is 80%, the probability of falling down in combination 3 (slippery ground + dim light + user's mobility difficulties) is 90%, and the probability of falling down in combination 4 (slippery ground + high toilet threshold) is 40%. After evaluation, it is determined that the user's fall risk has reached the risk threshold (such as 80%) when "slippery ground + user's mobility difficulties", then (slippery ground + user's mobility difficulties) is considered to be the minimum cut set. If these two factors appear at the same time, the system can infer that the sub-event "falling down" is a high probability risk that is about to occur. Then, the causal function relationship between this factor combination and the accident sub-event and the top event (such as "falling down risk") can be integrated to construct a complete risk analysis link, that is: slippery ground + user's mobility difficulties → falling down risk. This link not only records the logical progressive relationship of risk events, but also structurally retains the causal dependency between triggering conditions, reasoning paths and final impacts.

[0051] Based on this risk analysis link, the system can match the environmental data and health data collected in the perception database with the minimum cut set defined in the link in real time during operation. If the cut set factors are detected to be established at the same time, the risk identification result can be directly output, and the top event in the analysis link can be used as the current warning item to trigger the subsequent feedback mechanism. Through this link modeling method based on the minimum factor combination and with logical closed-loop characteristics, it not only has good prediction capabilities, but also supports multi-risk cross-cutting and multi-level causal reasoning, providing high-confidence and high-timeliness risk identification and warning services for elderly users.

[0052] P40: Perform correlation analysis on the risk analysis link and the perception database, identify the mapping relationship between the perception data and the risk analysis link, project the perception data in the perception database to the risk analysis link associated with it according to the mapping relationship, perform risk identification according to the perception data through the risk analysis link, and obtain a risk identification result.

[0053] Furthermore, step P40 of the embodiment of the present application also includes: P41: Identify the cross-correlation links of each risk analysis link; P42: Perform risk association identification based on the cross-correlation links to obtain associated risk identification results; P43: Perform time-series identification risk fusion on the associated risk identification results to obtain the risk identification results, which can reflect the sequential relationship of risk occurrence.

[0054] Optionally, the mapping relationship between the perception data and the risk analysis link is identified by performing correlation analysis on the risk analysis link and the perception database, which stores perception data from various smart home devices, such as heart rate data collected by smart bracelets and motion trajectory data captured by smart cameras. The risk analysis link clarifies the factors and logical relationships that lead to specific risks. For example, for the risk analysis link of "fall risk warning", the system will identify perception data related to falls, such as environmental monitoring data of slippery ground and abnormal changes in user motion trajectories. Through correlation analysis, the system can identify the mapping relationship between perception data and the risk analysis link, and project the perception data to the risk analysis link according to the mapping relationship for risk identification, thereby obtaining preliminary risk identification results.

[0055] For example, the system can use the "dizziness" recognized by the voice device as a trigger factor, combined with health data analysis (such as high blood pressure, hypoglycemia, etc.), to output a comprehensive fall risk identification result. This result will trigger the system's early warning mechanism, and remind users to take timely response measures through different feedback channels (such as voice prompts, APP push, family notifications, etc.).

[0056] However, in real life, the risks faced by elderly users often do not exist in isolation. For example, a fall of the elderly may cause changes in heart rate, and abnormal changes in heart rate may also be a signal of other health problems. Therefore, the system needs to further identify the cross-correlation links between the risk analysis links. This means that the system must not only analyze a single risk analysis link, but also identify the mutual influence between different risk links. For example, there may be a cross-correlation between the fall risk link and the health sign risk link, and a fall event may cause abnormal changes in health signs such as heart rate and blood pressure. For example, when an elderly person falls, not only the risk analysis link of falling is involved, but other links related to heart rate changes, blood pressure fluctuations, etc. may also be triggered. By identifying these cross-links, the system can more comprehensively grasp the health status and potential risks of elderly users.

[0057] Technically, identifying cross-correlation links can be achieved through multi-dimensional data analysis and network analysis methods (such as graph theory). Each risk analysis link can be regarded as a node in the graph, while the cross-correlation link is an edge established through the mutual influence between different data. By analyzing the relationship between different nodes and edges, the system can identify potential cross-links.

[0058] Next, risk association identification is required based on the identified cross-correlation links. By analyzing the cross-influences between different risk analysis links, the system can identify the relationship between multiple risk factors. For example, when an elderly person falls, their heart rate may increase abnormally, and this change in heart rate may be due to high blood pressure or long-term chronic diseases. Based on this cross-relationship, the system can comprehensively analyze multiple related links to obtain a more accurate correlation risk identification result. This result can help the system fully understand the causes of risks and how the risks influence and interact with each other, thereby providing a basis for subsequent intervention and treatment.

[0059] Technically, this process can be achieved through multimodal data fusion technology, combining data from different devices and sensors to conduct association rule mining and causal analysis. Deep learning and neural network models can also be used to analyze complex cross-risk relationships and help the system automatically identify the inherent connections between multiple risks.

[0060] Next, the identified associated risk identification results are subjected to time series identification and risk fusion. Since risk events usually have a time sequence, for example, a fall may occur first, followed by fluctuations in heart rate, which may eventually trigger a health alert, it is necessary to analyze the order in which these events occur. Through time series identification, the order in which risk events occur is determined. For example, the system can determine whether a fall first causes an abnormal heart rate, or an abnormal heart rate causes a fall, so as to better understand the risk propagation process and perform risk fusion based on time series data. The fused results can reflect the correlation and order of occurrence of different risk events, allowing the system to more accurately identify and predict the health status of elderly users.

[0061] Technically, time series recognition can be achieved through time series analysis methods (such as dynamic time warping (DTW), hidden Markov model (HMM), etc.). By analyzing the timestamps and event sequence in historical data, the system can build a time series model to predict and warn of possible risk events in the future.

[0062] For example, assume that the system detects the following sensory data: Smart camera captures elderly man falling in bathroom.

[0063] The smart bracelet recorded an abnormally high heart rate after the fall.

[0064] The smart mattress recorded that the elderly man had poor sleep quality and tossed and turned frequently the night before.

[0065] Through correlation analysis, the system identifies the cross-correlation between the fall risk link and the health sign risk link. Falling may cause an abnormal increase in heart rate, and poor sleep quality may indicate that the elderly have potential health problems, such as anxiety or heart disease. Through risk association identification, the system determines the association between falling and abnormal heart rate, as well as the potential connection between poor sleep quality and fall risk.

[0066] Finally, through time-series risk fusion, the sequence of events can be restored: the elderly person fell in the bathroom due to poor sleep quality, which made him unable to move, and his heart rate increased abnormally after the fall. This comprehensive risk identification result not only helps the system accurately assess the health status of the elderly, but also provides an important basis for subsequent treatment interventions, such as recommending cardiovascular examinations, improving the sleeping environment, and strengthening anti-slip measures for the elderly.

[0067] Through correlation analysis, data from different devices and sensors are integrated to identify the cross-relationships between different risks, and the risk identification results are further optimized through time series analysis. This process not only helps the system identify the root causes of risks, but also provides effective support for subsequent treatment interventions, health management and personalized services.

[0068] Furthermore, the embodiment of the present application further includes step P50, and step P50 further includes: P51: Obtain the risk feedback channel, and parse the feedback data features according to the risk feedback channel, wherein the feedback data features are used to characterize the display form and display content features of the feedback path; P52: Perform content matching and identification on the risk identification results according to the feedback data features, and obtain the feedback core content, feedback auxiliary content and channel feedback parameters; P53: Generate early warning feedback information according to the feedback core content, feedback auxiliary content and channel feedback parameters, and perform early warning feedback through the risk feedback channel.

[0069] In a possible embodiment of the present application, early warning feedback information can be generated and sent according to the risk identification results, and effective early warning can be carried out through different feedback channels (such as text, voice, vision, etc.). This process can help elderly users and their families respond to risks quickly and ensure timely intervention measures.

[0070] Specifically, we first need to determine the risk feedback channel, that is, the way to deliver warning information to elderly users and their families. These feedback channels may include text (such as text messages, APP push notifications), voice (such as voice prompts, smart speaker voice feedback), visual (such as graphics or text reminders on the display), and other methods. The system needs to understand how to display warning information by analyzing the data characteristics of these feedback channels. Feedback data characteristics include display form (such as text, images, sounds, etc.) and display content characteristics (such as the level of detail, tone, priority, etc. of warning information). For example, in the voice feedback channel, the warning information may need to be reminded through a voice assistant (such as Alexa or Siri), while in the visual feedback channel, the risk information may be displayed through a smart home display, or a push notification may be sent through the elderly's mobile phone APP. By analyzing these characteristics, the display content and display method are customized for each feedback channel to ensure the effective communication and understanding of the warning information.

[0071] Further, according to the characteristics of the parsed feedback data, the previously obtained risk identification results are matched and identified. That is, the risk identification results are matched with the characteristics of the feedback channel to determine which risk information is suitable for display through a specific feedback method. Exemplarily, the core content of the feedback is first identified, which is the most important information, such as whether the elderly are at risk of falling, abnormal heart rate, etc. Next, the auxiliary content of the feedback is identified, which can provide supplementary information, such as emergency measures to be taken, relevant health advice, etc. Finally, according to the characteristics of different channels, suitable channel feedback parameters are identified, such as the volume of the voice, the font size of the display, or the color of the visual icon. For example, when the risk of falling is identified, the system can choose to send a reminder to the elderly through the voice channel: "Please be careful, the ground is slippery, pay attention to safety!" At the same time, a push notification is sent to the family through the text channel: "The elderly are at risk of falling, the current environment is slippery, please pay attention in time."

[0072] Furthermore, early warning feedback information is generated based on the core content of feedback, auxiliary content of feedback, and channel feedback parameters. This means that the system must not only consider how to deliver information, but also ensure that the way the information is delivered can reach the needs and perception of the target users. For example, for the elderly, voice feedback may be more intuitive and easy to understand, while for family members, text push notifications can provide more detailed event information.

[0073] According to different channels and feedback methods, the system will generate appropriate warning information based on the characteristics of each feedback and send it through the feedback channel. For example, if voice feedback is selected, the system may use a voice assistant for voice broadcast; if visual feedback is selected, the system can display graphics or text through mobile phone apps, smart home screens, etc. In addition, the warning information will be adjusted according to the user's personal settings. For example, elderly users may set the volume and speed of voice feedback, and family members may choose whether to receive real-time push notifications of risk information.

[0074] Through cross-analysis of multiple data links and matching of feedback information, the system can provide accurate and targeted early warning information to ensure the timeliness and effectiveness of risk management and health intervention.

[0075] Furthermore, step P52 of the embodiment of the present application also includes: P52-1: Analyze the risk identification results, and extract the risk type, occurrence time, and response measures as risk analysis parameters; P52-2: According to the feedback data characteristics, obtain the feedback execution device and its early warning feedback content parameters; P52-3: Use the feedback execution device and its early warning feedback content parameters to match the risk type, occurrence time, and response measures in the risk analysis parameters to obtain the risk analysis parameters with the greatest matching degree as the feedback core content, and the remaining risk analysis parameters except the feedback core content as the feedback auxiliary content; P52-4: Configure the early warning feedback content parameters of the feedback execution device according to the feedback core content and feedback auxiliary content to obtain the channel feedback parameters.

[0076] Specifically, in order to achieve accurate matching of risk identification results and personalized early warning feedback, to ensure that elderly users and their families can receive risk information efficiently and in a timely manner.

[0077] First, the system needs to parse the risk identification results and extract the risk type, occurrence time, and response measures as risk analysis parameters. These parameters help the system better understand the specific content of risk events, such as identifying that the smoke concentration in the elderly user’s home is too high, there may be a fire risk, the event occurred at 10:30 AM, and the response measures may be to evacuate the room immediately and call the fire alarm. Through these parsed key information, the system can clearly understand how to deal with each risk and prepare for subsequent feedback.

[0078] Next, the characteristics of the obtained feedback data are analyzed to determine the appropriate feedback execution device and the matching warning feedback content parameters. The feedback execution device can be a device such as a smart speaker, a smart phone or a display screen, and the warning feedback content parameters include voice volume, speech speed, or font size and color on the display screen. For example, text feedback is suitable for elderly users with good eyesight and can be displayed through a mobile phone application or a smart home control panel. Voice feedback is suitable for elderly users with poor eyesight or who need immediate attention and can be played through a smart speaker.

[0079] Furthermore, the feedback content is matched with the risk type, occurrence time, and response measures in the risk analysis parameters through the feedback execution device and its early warning feedback content parameters. The most relevant content will be selected as the core content of the feedback based on the matching results, and other relevant content will be used as auxiliary content of the feedback. For example, if a fire risk is identified, the system may send "Fire warning! Please evacuate the room immediately" as the core content of the feedback through a voice device, and send "If you need help, please press the emergency button" through a text channel.

[0080] Finally, based on the matched feedback core content, feedback auxiliary content, and channel feedback parameters, the system will generate the final warning feedback information and provide warning feedback through the selected feedback execution device. For example, the voice assistant will broadcast "Fire warning! Please evacuate the room immediately. If you need help, please press the emergency button." In this way, the system can ensure that the warning information is accurately delivered according to the user's needs and device characteristics, so that the elderly users and their families can respond in time.

[0081] Through this series of steps, the system can not only accurately match the feedback content with the risk type, but also ensure the efficient delivery of feedback information through personalized configuration. This precise risk identification and feedback mechanism provides a comprehensive and intelligent safety management system for elderly users, helping them deal with potential risks and ensure quality of life and safety.

[0082] Furthermore, the embodiment of the present application further includes step P60, and step P60 further includes: P61: Based on the risk identification result, continuous perception data is obtained; P62: Based on the continuous perception data, risk identification is performed by projecting it to the corresponding risk analysis link to obtain a continuous risk identification result; P63: According to the identification time of the risk identification result, the risk identification result and the continuous risk identification result are arranged in time series to obtain a risk identification change; P64: Risk warnings are continuously updated based on the risk identification change.

[0083] Optionally, the risk identification and feedback mechanism can be further expanded to maintain continuous monitoring and risk warning for elderly users in a constantly changing environment. The risk identification results are continuously updated and time-series analyzed through real-time perception data, thereby ensuring that users receive dynamic and real-time risk warnings throughout the process.

[0084] First, based on the risk identification results obtained, continuous perception data is obtained. Continuous perception data refers to real-time data continuously collected through smart home devices, health monitoring devices or other sensors. These data include not only the original environmental data and user activity data, but also changes in the user's health status, such as body temperature, heart rate, blood pressure, etc. These data will be used to further analyze the changing trends of risks and dynamically adjust the risk identification results based on new data. For example, if the smoke concentration risk in the environment has been detected before, the new perception data may further provide more details or changes about the smoke concentration in the environment.

[0085] Next, the continuous perception data is projected to the corresponding risk analysis link for risk identification to obtain updated continuous risk identification results. By combining new real-time data with known risk analysis links, its judgment of risks can be updated. For example, if abnormal heart rate data appears in the smoke concentration risk analysis link, the system will recognize that changes in heart rate may be a subsequent risk after the smoke concentration is too high, and then update the status of the smoke concentration risk. The projection of continuous perception data ensures that the system can continuously identify and classify risks based on the latest information, and promptly discover new potential risks.

[0086] Furthermore, according to the identification time of the risk identification results, the risk identification results and the continuous risk identification results are arranged in time series to obtain the risk identification change amount. At this time, the system not only cares about the accuracy of each risk identification result, but also needs to pay attention to the changing trend of the risk status. By comparing the risk identification results at different time points, the system can calculate the change amount of risk identification to help judge the trend of increasing or decreasing risks. For example, if the smoke concentration risk gradually increases over a period of time, the system will identify this change and prepare to make corresponding warning updates.

[0087] Finally, the risk warning is continuously updated based on the amount of risk identification changes. If the system detects a significant change in the risk identification results (such as a significant increase in smoke concentration risk in a short period of time), the warning mechanism will be triggered immediately, and the system will provide real-time feedback and adjust the warning information based on the updated risk identification results. For example, the system may adjust the urgency of the voice or visual warning, or provide more specific risk response measures by notifying users and their families. In this way, the risk warning is not only based on the initial identification results, but can also be dynamically updated according to changes in real-time data, ensuring that users and their families can always respond to potential risks in a timely manner.

[0088] Through the above steps, the system can achieve continuous risk identification and dynamic early warning feedback, ensuring that elderly users can receive timely, accurate, and personalized reminders and interventions when facing risks. This dynamic monitoring and update mechanism enhances the adaptability and real-time performance of the system, helping elderly users effectively avoid or reduce the risks brought by sudden problems.

[0089] In summary, the embodiments of the present application have at least the following technical effects: This application builds a perception database by connecting smart home devices and analyzing their data features; then obtains the risk perception needs of elderly users and establishes a risk perception needs library. Based on the needs library, the system performs risk needs analysis and builds a risk analysis link. Then, through correlation analysis, the system maps the data in the perception database to the risk analysis link, projects it to the link for risk identification, and finally obtains accurate risk identification results.

[0090] The technical effect of achieving real-time and accurate identification of health risks for elderly users and personalized early warning has been achieved through multi-device data parsing and correlation analysis.

[0091] Embodiment 2 is based on the same inventive concept as the risk perception interaction method for elderly users in a smart home environment in the aforementioned embodiment. Figure 2 As shown, the present application provides a risk perception interaction system for elderly users in a smart home environment. The system and method embodiments in the present application are based on the same inventive concept. The system includes: The data feature analysis module 11 is used to connect to smart homes, analyze the data features of each connected smart home, and build a perception database.

[0092] The risk perception demand library construction module 12 is used to obtain the risk perception needs of elderly users and construct a risk perception demand library.

[0093] The risk demand analysis module 13 is used to analyze risk demands according to the risk perception demand library, and obtain a risk analysis link according to the identification causal chain of risk demands.

[0094] The risk identification module 14 is used to perform correlation analysis based on the risk analysis link and the perception database, identify the mapping relationship between the perception data and the risk analysis link, project the perception data in the perception database to the risk analysis link associated with it according to the mapping relationship, perform risk identification according to the perception data through the risk analysis link, and obtain a risk identification result.

[0095] Furthermore, the data feature analysis module 11 is also used to perform the following steps: The collected data of the connected smart homes are analyzed in terms of type, content and frequency to obtain data features of each smart home; and the perception database is constructed according to the mapping relationship between the connected smart homes and the data features.

[0096] Furthermore, the data feature analysis module 11 is also used to perform the following steps: The data type of the collected data is used as the first classification feature, and the smart home and its collected data are classified into the first level to obtain the first level classification data; the data content of the collected data is used as the second classification feature, and the first level classification data is classified into the second level to obtain the second level classification data; the collection frequency of the collected data is used as the third classification feature, and the second level classification data is time-aligned and classified to obtain the third level classification data; the smart home and its collected data are classified, graded and integrated according to the classification results and hierarchical relationship of the first level classification data, the second level classification data, and the third level classification data, and the perception database is constructed.

[0097] Furthermore, the risk perception demand library construction module 12 is also used to perform the following steps: Enter the identity information and health data of the elderly target users; match the identity information and health data of the elderly target users with a preset risk perception library to obtain recommended risk perception needs; select and set or customize the settings based on the recommended risk perception needs to obtain the risk perception needs of the elderly users.

[0098] Furthermore, the risk requirement analysis module 13 is also used to perform the following steps: Taking the risk perception demand as the top event, the accident sub-event of the top event is analyzed based on historical data, and the causal function relationship between the accident sub-event and the top event is fitted; the accident factor of the accident sub-event is analyzed based on historical data, and the causal function relationship between the accident sub-event and the accident factor is fitted; the minimum cut set of the accident factor combination is calculated, and when the accident factor combination in the minimum cut set is established, the accident sub-event is bound to occur. Based on the accident factor combination-accident sub-event-top event and their causal function relationship, the risk analysis link is obtained.

[0099] Furthermore, the risk identification module 14 is further configured to perform the following steps: Identify the cross-correlation links of each risk analysis link; perform risk association identification based on the cross-correlation links to obtain an associated risk identification result; perform time-series identification risk fusion on the associated risk identification result to obtain the risk identification result, which can reflect the sequential relationship of risk occurrence.

[0100] Furthermore, the system also includes: A feedback data feature analysis module, the feedback data feature analysis module is used to obtain a risk feedback channel, analyze feedback data features according to the risk feedback channel, and the feedback data features are used to characterize the display form and display content features of the feedback channel; a content matching identification module, the content matching identification module is used to perform content matching identification on the risk identification result according to the feedback data features, and obtain feedback core content, feedback auxiliary content and channel feedback parameters; an early warning feedback module, the early warning feedback module is used to generate early warning feedback information according to the feedback core content, feedback auxiliary content and channel feedback parameters, and perform early warning feedback through the risk feedback channel.

[0101] Furthermore, the content matching and identification module is further configured to perform the following steps: Analyze the risk identification result, extract the risk type, occurrence time, and response measures as risk analysis parameters; obtain the feedback execution device and its early warning feedback content parameters according to the feedback data characteristics; use the feedback execution device and its early warning feedback content parameters to match the risk type, occurrence time, and response measures in the risk analysis parameters to obtain the risk analysis parameters with the greatest matching degree as the feedback core content, and use the remaining risk analysis parameters except the feedback core content as the feedback auxiliary content; configure the early warning feedback content parameters of the feedback execution device according to the feedback core content and the feedback auxiliary content to obtain the channel feedback parameters.

[0102] Furthermore, the system also includes: A cross-correlation link identification module, which is used to identify the cross-correlation links of each risk analysis link; a risk association identification module, which is used to perform risk association identification based on the cross-correlation links to obtain an associated risk identification result; a time sequence identification risk fusion module, which is used to perform time sequence identification risk fusion on the associated risk identification results to obtain the risk identification result, and the risk identification result can reflect the sequential relationship of risk occurrence.

[0103] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0104] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

[0105] This specification and the drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A risk perception interaction method for elderly users in a smart home environment, characterized in that: include: Connect smart homes, analyze the data characteristics of each connected smart home, and build a perception database; Obtain risk perception needs of elderly users and build a risk perception needs database; Perform risk demand analysis according to the risk perception demand library, and obtain a risk analysis link according to the risk demand identification causal chain; Performing a correlation analysis on the risk analysis link and the perception database, identifying a mapping relationship between the perception data and the risk analysis link, projecting the perception data in the perception database to the risk analysis link associated therewith according to the mapping relationship, performing risk identification according to the perception data through the risk analysis link, and obtaining a risk identification result; Among them, risk demand analysis is performed according to the risk perception demand library, and a risk analysis link is obtained according to the identification causal chain of risk demand, including: Taking risk perception demand as the top event, analyzing the accident sub-events of the top event based on historical data, and fitting the causal function relationship between the accident sub-event and the top event; Analyze the accident factors of the accident sub-events based on historical data, and fit the causal function relationship between the accident sub-events and the accident factors; The minimum cut set of the accident factor combination is obtained. When the accident factor combination in the minimum cut set is established, the accident sub-event must occur. Based on the accident factor combination-accident sub-event-top event and their causal function relationship, the risk analysis link is obtained.

2. The risk perception interaction method for elderly users in a smart home environment according to claim 1 is characterized in that: Obtain risk identification results, followed by: Acquire a risk feedback channel, and parse feedback data features according to the risk feedback channel, wherein the feedback data features are used to characterize display form and display content features of the feedback channel; Perform content matching and identification on the risk identification results according to the feedback data characteristics to obtain the feedback core content, feedback auxiliary content and channel feedback parameters; According to the feedback core content, feedback auxiliary content and channel feedback parameters, early warning feedback information is generated, and early warning feedback is performed through the risk feedback channel.

3. The risk perception interaction method for elderly users in a smart home environment according to claim 2 is characterized in that: The risk identification results are matched and identified according to the feedback data characteristics to obtain the feedback core content, feedback auxiliary content and channel feedback parameters, including: Analyze the risk identification results and extract risk type, occurrence time, and response measures as risk analysis parameters; According to the feedback data characteristics, obtaining the feedback execution device and its early warning feedback content parameters; The feedback execution device and its early warning feedback content parameters are used to match the risk type, occurrence time, and response measures in the risk analysis parameters, and the risk analysis parameters with the greatest matching degree are obtained as the feedback core content, and the remaining risk analysis parameters except the feedback core content are used as the feedback auxiliary content; The early warning feedback content parameters of the feedback execution device are configured according to the feedback core content and the feedback auxiliary content to obtain channel feedback parameters.

4. The risk perception interaction method for elderly users in a smart home environment according to claim 1 is characterized in that: The step of analyzing the data features of each connected smart home and constructing a perception database includes: Analyze the type, content, and frequency of the collected data connected to smart homes to obtain the data characteristics of each smart home; The perception database is constructed according to the mapping relationship between the smart home and the data features.

5. The risk perception interaction method for elderly users in a smart home environment according to claim 4 is characterized in that: Building a perception database also includes: Using the data type of the collected data as a first classification feature, performing primary classification on the smart home and its collected data to obtain primary classification data; Using the data content of the collected data as a second classification feature, performing secondary classification on the primary classification data to obtain secondary classification data; Using the collection frequency of the collected data as the third classification feature, performing time alignment classification on the secondary classification data to obtain tertiary classification data; According to the classification results and hierarchical relationships of the first-level classification data, the second-level classification data, and the third-level classification data, the smart home and its collected data are classified, graded and integrated to construct the perception database.

6. The risk perception interaction method for elderly users in a smart home environment according to claim 1 is characterized in that: Obtain the risk perception needs of elderly users, including: Enter the identity information and health data of elderly target users; According to the identity information and health data of the elderly target user, the needs are matched with the preset risk perception library to obtain the recommended risk perception needs; Select and set or customize the recommended risk perception needs to obtain the risk perception needs of the elderly user.

7. The risk perception interaction method for elderly users in a smart home environment according to claim 2 is characterized in that: Obtain risk identification results, followed by: Obtaining continuous perception data based on the risk identification results; Projecting the continuous perception data to the corresponding risk analysis link to perform risk identification and obtain continuous risk identification results; According to the identification time of the risk identification result, the risk identification result and the continuous risk identification result are arranged in time series to obtain the risk identification change amount; The risk warning is continuously updated according to the risk identification change amount.

8. The risk perception interaction method for elderly users in a smart home environment according to claim 1 is characterized in that: Obtain risk identification results, including: Identify the cross-correlation links of each risk analysis link; Perform risk association identification according to the cross-correlation link to obtain an association risk identification result; The associated risk identification results are subjected to time series identification risk fusion to obtain the risk identification results, and the risk identification results can reflect the sequential relationship of risk occurrence.

9. The elderly user risk perception interactive system in the smart home environment is characterized by: The system comprises: A data feature analysis module, which is used to connect to smart homes, analyze data features of each connected smart home, and build a perception database; A risk perception demand library construction module, wherein the risk perception demand library construction module is used to obtain the risk perception needs of elderly users and construct a risk perception demand library; A risk demand analysis module, the risk demand analysis module is used to analyze risk demands according to the risk perception demand library, and obtain a risk analysis link according to the identification causal chain of risk demands; A risk identification module, which is used to perform a correlation analysis based on the risk analysis link and the perception database, identify the mapping relationship between the perception data and the risk analysis link, project the perception data in the perception database to the risk analysis link associated therewith according to the mapping relationship, perform risk identification according to the perception data through the risk analysis link, and obtain a risk identification result; The risk demand analysis module is also used to perform the following steps: The risk demand analysis is performed according to the risk perception demand library, and the risk analysis link is obtained according to the risk demand identification causal chain, including: Taking risk perception demand as the top event, analyzing the accident sub-events of the top event based on historical data, and fitting the causal function relationship between the accident sub-event and the top event; Analyze the accident factors of the accident sub-events based on historical data, and fit the causal function relationship between the accident sub-events and the accident factors; The minimum cut set of the accident factor combination is obtained. When the accident factor combination in the minimum cut set is established, the accident sub-event must occur. Based on the accident factor combination-accident sub-event-top event and their causal function relationship, the risk analysis link is obtained.

Images