Safety state monitoring and early warning method and system for elderly people living alone

Abstract

The invention provides a safety state monitoring and early warning method and system for elderly people living alone, and relates to the technical field of smart pension, the system comprises a sensing layer, a transmission layer, a processing layer and an application layer, during implementation, a sensing unit is firstly installed, then relay equipment is installed, then adaptability training of the system is carried out, and finally alarm information is processed. Through dual-mode cooperation of environment perception and attitude monitoring, accurate analysis of home behaviors of the elderly living alone is realized, the probability of false alarm is greatly reduced, deep learning is combined to drive closed-loop optimization, through three-stage threshold iteration, the tumble recognition rate is greatly improved, the false alarm rate is greatly reduced, linkage response is triggered in an emergency scene, and the accuracy of tumble recognition is improved. The community disposal efficiency is greatly improved, the beacons are integrated with intelligent insoles / magnetic chest cards and other non-inductive wearing designs, and the house type self-adaptive base station deployment is combined, so that the disturbance of grid members to the old is reduced, and meanwhile, the monitoring efficiency and accuracy are improved.

Description

Technical Field

[0001] The present invention relates to the field of smart elderly care technology, and in particular to a safety status monitoring and early warning method and system for elderly people living alone. Background Art

[0002] Smart elderly care refers to an intelligent model that leverages technologies such as the Internet of Things (IoT), artificial intelligence (AI), wearable devices, and big data to provide health monitoring, safety assurance, and life support services to seniors living at home or in the community. Its core goal is to improve the quality of life of seniors, reduce the risk of accidents, and alleviate social pressure on elderly care through technological means.

[0003] However, in the existing technology, the existing safety status monitoring of elderly people living alone, in addition to monitoring the water, electricity, gas and other equipment in the elderly’s homes, also has major defects in the use of visual monitoring to increase the information obtained by monitoring and the use of smart bracelets to monitor the physical condition of the elderly. For example, visual monitoring requires mature image recognition technology to identify the posture of the elderly in the image, and this technology often has a large delay and is very prone to delays. Smart bracelets can only monitor basic data such as the elderly’s heart rate and blood oxygen. When the elderly fall, it is difficult to issue an early warning based on the monitoring data of the smart bracelet. Even the combination of visual monitoring and smart bracelets will have the problem of low data correlation. Therefore, there is still a lot of room for improvement in the early warning of the safety status monitoring of elderly people living alone. Summary of the Invention

[0004] The purpose of the present invention is to address the shortcomings of the existing technology. Through the dual-modal collaboration of environmental perception and posture monitoring, accurate analysis of the home behavior of elderly people living alone can be achieved, the probability of false alarms is greatly reduced, and combined with deep learning to drive closed-loop optimization, the fall recognition rate is greatly improved and the false alarm rate is greatly reduced after three-stage threshold iteration. Emergency scenarios trigger a linkage response, and the community handling efficiency is greatly improved. The beacon integrates smart insoles / magnetic badges and other non-contact wearable designs, combined with household adaptive base station deployment, which reduces the disturbance of grid workers to the elderly while also improving monitoring efficiency and accuracy.

[0005] To achieve the above objectives, the present invention adopts the following technical solutions: a method and system for monitoring and warning the safety status of elderly people living alone, comprising: A perception layer, comprising an environment perception unit and a posture perception unit; A transport layer comprising an edge gateway for transmitting data to an environment sensing unit and a beacon base station for transmitting data to a posture sensing unit; a processing layer comprising a posture analysis module for analyzing user data acquired by the posture perception unit, an environment analysis module for analyzing user data acquired by the environment perception unit, and a collaborative control module for mobilizing the operation of the perception layer based on the analysis results of the posture analysis module and the environment analysis module; The application layer includes a visualization platform for displaying the user's environmental status and posture status and an early warning classification module for classifying early warnings according to abnormal content and issuing early warnings.

[0006] The environmental sensing unit includes a water flow sensor for obtaining the user's water consumption data, a power sensor for obtaining the user's electricity consumption data, a gas sensor for obtaining the user's gas consumption data, and a door magnetic sensor for obtaining the user's door switch data. The environmental sensing unit obtains the user's original water consumption data, electricity consumption data, gas consumption data, and door magnetic sensor data, which are processed by the edge gateway and uploaded to the environmental analysis module of the processing layer for analysis; The posture sensing unit includes a head beacon composed of UWB and BLE dual-mode beacons, a hand beacon composed of a flexible RFID bracelet, a torso beacon composed of three-axis acceleration and UWB, and a foot beacon composed of a pressure-sensing insole. The head beacon, hand beacon, torso beacon and foot beacon of the posture sensing unit are scanned by the beacon base station and the relative distance between each beacon is calculated, and then uploaded to the posture analysis module for posture analysis.

[0007] As a preferred embodiment, it also includes a collaborative optimization module constructed with the collaborative control module, the early warning classification module, the collaborative trigger module, the early warning disposal platform and the deep learning model. The collaborative optimization module obtains the abnormal water, electricity and gas usage, door magnetic opening and closing abnormality, and posture abnormality obtained by the environmental analysis module and the posture analysis module according to the preset water, electricity and gas abnormality thresholds, the door magnetic opening and closing time thresholds, and the beacon relative distance thresholds. The collaborative control module sends a verification request to the posture analysis module according to the abnormal situation, and activates the beacon base station to re-scan the posture perception unit through the collaborative trigger. At the same time, the collaborative control module generates specific alarm content according to the abnormal event, and transmits the alarm content, environmental perception data and posture perception data to the posture analysis module. The alarm is sent to the early warning and disposal platform. After processing the alarm content, the early warning and disposal platform uploads the alarm content, environmental perception data, posture perception data and alarm disposal feedback to the deep learning model, analyzes the relationship between the environmental perception data and posture perception data in the abnormal event, generates new perception data thresholds and sends them to the posture analysis module and the environmental analysis module respectively, updates the settings of the posture analysis module and the environmental analysis module, and at the same time, the collaborative control module activates the beacon base station, rescans to obtain beacon data and uploads it to the posture analysis module, uses the updated settings to scan and analyze the user's body posture, and feeds the updated settings of the posture analysis module and the environmental analysis module and the beacon scanning data back to the deep learning model via the collaborative control module to optimize the deep learning model.

[0008] As a preferred embodiment, the beacon base station scans the head beacon, hand beacon, torso beacon and foot beacon to obtain the scanning data of each beacon, and then uploads the relative position data of the beacon to the posture analysis module to calculate the relative positions between the head beacon, hand beacon, torso beacon and foot beacon, analyzes the human posture corresponding to the head beacon, hand beacon, torso beacon and foot beacon according to the human motion model, and analyzes the abnormal data of the beacon data according to the set beacon relative spacing threshold, and uploads the abnormal data to the collaborative control module, and the collaborative control module processes the abnormal data according to the operation steps of the collaborative optimization module.

[0009] As a preferred embodiment, when the original environmental data obtained by the environmental perception unit is uploaded to the transmission layer, the edge gateway processes the original environmental data, analyzes the abnormal data in the original environmental data according to the pre-set water, electricity, gas abnormality thresholds and door magnetic opening and closing time thresholds, and uploads the abnormal data to the collaborative control module. The collaborative control module processes the abnormal data according to the operation steps of the collaborative optimization module.

[0010] As a preferred embodiment, the application layer obtains the alarm information obtained by the processing layer, displays it through the visualization platform, and classifies and warns the alarm information through the warning classification module. The visualization platform is installed in the community service center.

[0011] A safety status monitoring and early warning method for elderly people living alone, comprising the following steps: S1. Installation of the sensing unit: Install the environment sensing unit in the user's home, hand over the posture sensing unit to the user, and guide the user to learn how to wear the posture sensing unit; S2. Installation of relay equipment: Install the edge gateway and beacon base station in the user's home and connect them to the early warning system; S3. System Adaptive Training: Observe the processing layer's processing of environmental and posture perception data, perform manual error correction, and collaboratively optimize the module's operational processes. Repeatedly train the deep learning model until the deep learning model's processing error for posture perception data, environmental perception data, and abnormal responses is less than 5%. S4. Processing of alarm information: Based on the alarm information displayed on the visualization platform and the graded warning obtained by the early warning analysis module, grid workers are dispatched to the user's home to process the alarm information.

[0012] As a preferred embodiment, in step S2, since the house where the user lives may hinder the scanning of the beacon base station, multiple beacon base stations are installed in the user's home according to the user's house type and the scanning surface shape of the beacon base station.

[0013] As a preferred implementation, in step S3, the coverage area of ​​a single early warning system is planned based on the community. Due to differences in household types, the installation locations of the beacon base stations in each user's home are different, and the beacon base stations need to be arranged in the manner of S2. Therefore, when the posture analysis module analyzes the beacon data obtained by scanning the beacon base stations, the data with the clearest data and the least noise among multiple beacon base stations is used as the main data, and the other data is used as fuzzy verification data.

[0014] As a preferred embodiment, in step S3, obvious errors in the environment analysis module and the posture analysis module are manually eliminated at the initial stage of deep learning model training of the collaborative optimization module. When the new perception data threshold obtained by the deep learning model makes the analysis result error of the posture analysis module and the environment analysis module within the range of 20% to 15%, the first round of verification is entered, and the collaborative control center triggers the re-acquisition of perception data according to the abnormal items, and trains the deep learning model according to the re-acquired perception data. The new perception data threshold obtained by the deep learning model is used to adjust the environment analysis module and the posture analysis module. The adjusted environment analysis module and posture analysis module then obtain new perception data, and the analysis results are input into the collaborative control center. The collaborative control center enters the second round of verification according to the abnormal items until the new perception data threshold obtained by the deep learning model makes the analysis result error of the posture analysis module and the environment analysis module less than 5%.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are: The present invention achieves accurate analysis of the home behavior of elderly people living alone through the dual-modal collaboration of environmental perception and posture monitoring, greatly reducing the probability of false alarms. In addition, combined with deep learning to drive closed-loop optimization, the fall recognition rate is greatly improved and the false alarm rate is greatly reduced after three-stage threshold iteration. Emergency scenarios trigger a linkage response, and the community handling efficiency is greatly improved. In addition, the beacon integrates non-contact wearable designs such as smart insoles / magnetic badges, combined with household adaptive base station deployment, which reduces the disturbance of grid workers to the elderly while also improving monitoring efficiency and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 The present invention proposes a module diagram of a safety status monitoring and early warning system for elderly people living alone; Figure 2 The present invention proposes a flow chart of a method for monitoring and early warning the safety status of elderly people living alone. DETAILED DESCRIPTION

[0017] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0018] Example like Figure 1-2 As shown, the present invention provides a technical solution: a safety status monitoring and early warning system for elderly people living alone, comprising: Perception layer: The perception layer includes an environment perception unit and a posture perception unit; The environmental sensing unit includes a water flow sensor for obtaining the user's water consumption data, a power sensor for obtaining the user's electricity consumption data, a gas sensor for obtaining the user's gas consumption data, and a door magnetic sensor for obtaining the user's door switch data. The water flow sensor can be a pipe clamp piezoelectric sensor, the power sensor can be a non-intrusive current transformer, the gas sensor can be implemented by combining catalytic combustion and semiconductor dual-mode, and the door magnetic sensor can be implemented by a Hall effect combined with a reed switch dual-redundancy design. In addition, the posture sensing unit includes a head beacon composed of UWB and BLE dual-mode beacons, a hand beacon composed of a flexible RFID bracelet, a torso beacon composed of three-axis acceleration and UWB, and a foot beacon composed of a pressure-sensing insole.

[0019] Transport layer: The transport layer includes an edge gateway for transmitting data to the environment sensing unit and a beacon base station for transmitting data to the posture sensing unit; When the raw environmental data acquired by the environmental perception unit is uploaded to the transmission layer, the edge gateway processes the raw environmental data, analyzes the abnormal data in the raw environmental data according to the pre-set water, electricity, gas abnormality thresholds and the door magnetic opening and closing time thresholds, and uploads the abnormal data to the collaborative control module. The collaborative control module processes the abnormal data according to the operation steps of the collaborative optimization module. Furthermore, the environmental perception unit obtains the user's original water consumption data, electricity consumption data, gas usage data and door magnetic sensor data, which are processed by the edge gateway and uploaded to the environmental analysis module of the processing layer for analysis. When the edge gateway obtains the original data of the water flow sensor, power sensor, gas sensor and door magnetic sensor, it first cleans the data. Then, the edge gateway analyzes the cleaned environmental perception data in real time and caches it locally. By judging the water, electricity and gas thresholds and analyzing the door magnetic behavior, it determines the specific abnormal items in the environmental perception data, and then generates a pre-processing event. After data compression and encryption, the edge gateway uploads the pre-processing event to the environmental analysis module for further analysis. In addition, environmental perception data can be combined with behavioral analysis, such as analyzing 24-hour water consumption data combined with toilet frequency, calculating the basic nighttime power consumption of power sensors to determine user sleep status, obtaining daily gas usage hours from gas sensors to determine kitchen usage, and analyzing data from door magnetic sensors to determine the duration of a user's single outing. Furthermore, the head beacon, hand beacon, torso beacon and foot beacon of the posture sensing unit are scanned by the beacon base station and the relative distances between the beacons are calculated, and then uploaded to the posture analysis module for posture analysis.

[0020] Processing layer: The processing layer includes a posture analysis module for analyzing user data obtained by the posture perception unit, an environment analysis module for analyzing user data obtained by the environment perception unit, and a collaborative control module for mobilizing the operation of the perception layer according to the analysis results of the posture analysis module and the environment analysis module; The beacon base station scans the head beacon, hand beacon, torso beacon and foot beacon to obtain the scanning data of each beacon, and then uploads the relative position data of the beacon to the posture analysis module to calculate the relative positions between the head beacon, hand beacon, torso beacon and foot beacon, analyzes the human posture corresponding to the head beacon, hand beacon, torso beacon and foot beacon according to the human motion model, and analyzes the abnormal data of the beacon data according to the set beacon relative spacing threshold, and uploads the abnormal data to the collaborative control module, and the collaborative control module processes the abnormal data according to the operation steps of the collaborative optimization module; Furthermore: the head beacon is used as a spatial reference point, and fall detection and balance assessment can be achieved by calculating the distance between the head beacon and the foot beacon; the hand beacon is used as a motion capture base point, and the forward leaning posture of the human body is identified by calculating the vector angle between the torso beacon and the hand beacon, and the state of the holding posture is identified by calculating the acceleration jitter of the hand beacon; the torso beacon is used as the core reference point, and the center of mass height of the human posture is identified by calculating the distance between the torso beacon and the foot beacon, and fall detection is performed by calculating the acceleration of the torso beacon; standing, sitting and lying postures are identified by calculating the pressure data of the foot beacon, and gait is analyzed by calculating the ratio of the support phase to the swing phase of the foot beacon; Specifically, when the posture analysis module performs posture analysis, it first obtains the original coordinate data of the beacon base station and performs spatial calibration in combination with multiple beacon base stations. Then, it constructs joint vectors based on the head-torso vector, left hand-torso vector, right hand-torso vector, left foot-torso vector, and right foot-torso vector. Then, it performs posture classification based on the data of the joint vectors, detects abnormal data in the data of the posture perception unit, and outputs the posture analysis results. In addition, the pre-processing events uploaded by the edge gateway include environmental perception data and abnormal items in the environmental perception data. After the pre-processing events are uploaded to the environmental analysis module, the environmental analysis module first performs multi-source data fusion, then makes specific judgments based on the threshold rules, and generates environmental abnormality events. The generated environmental abnormality events are transmitted to the collaborative control module. The collaborative control module mobilizes the posture perception unit and the posture analysis module to re-analyze the posture, and verifies and optimizes the analysis results of the environmental analysis module in combination with the posture perception data obtained by re-scanning.

[0021] Application layer: The application layer includes a visualization platform for displaying the user's environmental status and posture status and an early warning classification module for classifying and issuing early warnings based on abnormal content; The application layer obtains the alarm information obtained by the processing layer, displays it through the visualization platform, and rates and issues warnings for the alarm information through the warning classification module. The visualization platform is installed in the community service center. In addition, the visualization platform can add a data parsing engine, an environmental status module and a posture display module, and display them on the screen. After the data of the processing layer is uploaded to the application layer, the data of the processing layer is parsed by the data parsing engine, and the corresponding environment-related data and posture-related data are processed respectively by the environmental status module and the posture display module, and the processed data are displayed on the screen in the form of a water, electricity, gas instrument panel and a three-dimensional three-body model. At the same time, the early warning classification module performs early warning classification according to the data of the processing layer and displays it on the screen.

[0022] Collaborative Optimization Module: The collaborative optimization module is constructed with the collaborative control module, early warning classification module, collaborative trigger module, early warning disposal platform and deep learning model. The collaborative optimization module obtains the abnormal water, electricity and gas usage, door magnetic opening and closing abnormality, and posture abnormality analyzed by the environmental analysis module and the posture analysis module based on the preset water, electricity and gas abnormality thresholds, door magnetic opening and closing time thresholds, and beacon relative distance thresholds; Furthermore, the collaborative control module sends a verification request to the posture analysis module according to the abnormal situation, and activates the beacon base station through the collaborative trigger to re-scan the posture perception unit. At the same time, the collaborative control module generates specific alarm content according to the abnormal event, and transmits the alarm content, environmental perception data and posture perception data to the early warning and disposal platform. After processing the alarm content, the early warning and disposal platform uploads the alarm content, environmental perception data, posture perception data and alarm disposal feedback to the deep learning model, analyzes the correlation between the environmental perception data and the posture perception data in the abnormal event, generates a new perception data threshold and sends it to the posture analysis module and the environmental analysis module respectively, updates the settings of the posture analysis module and the environmental analysis module, and at the same time, the collaborative control module activates the beacon base station, re-scans to obtain beacon data and uploads it to the posture analysis module, uses the updated settings to scan and analyze the user's body posture, and feeds back the updated settings of the posture analysis module and the environmental analysis module and the beacon scanning data to the deep learning model via the collaborative control module to optimize the deep learning model; Specifically, in the initial stage of deep learning model training of the collaborative optimization module, obvious errors in the environment analysis module and the posture analysis module are manually eliminated. When the new perception data threshold derived by the deep learning model makes the analysis result error of the posture analysis module and the environment analysis module within the range of 20% to 15%, the first round of verification is entered. The collaborative control center triggers the re-acquisition of perception data based on the abnormal items, and trains the deep learning model based on the re-acquired perception data. The new perception data threshold derived by the deep learning model is used to adjust the environment analysis module and the posture analysis module. The adjusted environment analysis module and posture analysis module then acquire new perception data, and input the analysis results into the collaborative control center. The collaborative control center enters the second round of verification based on the abnormal items until the new perception data threshold derived by the deep learning model makes the analysis result error of the posture analysis module and the environment analysis module less than 5%.

[0023] Based on the above content, a safety status monitoring and early warning method for elderly people living alone is also proposed, which includes the following steps: S1. Installation of the sensing unit: Install the environment sensing unit in the user's home, hand over the posture sensing unit to the user, and guide the user to learn how to wear the posture sensing unit; Among them, the installation of the environmental sensing unit can be targeted according to the device types of water flow sensor, power sensor, gas sensor and door magnetic sensor. Since the posture sensing unit is wearable, the head beacon can be made into a headdress or hat type according to the user's preference, the torso beacon can be made into a chest beacon, button, belt type, and the hand beacon can be made into a bracelet type. Since the foot beacon needs to detect the pressure of the sole of the foot, the foot beacon needs to be integrated into the insole for wear. Shoes need to be replaced, so multiple pairs of foot beacons can be prepared for a single user, and a pressure switch can be set in the foot beacon. When the user continues to step on the foot beacon, the pressure switch is turned on. When the user does not apply pressure to the foot beacon for a period of time, the pressure switch is turned off and turned on again when the scanning signal of the beacon base station is received; S2. Installation of relay equipment: Install the edge gateway and beacon base station in the user's home and connect them to the early warning system; Among them, since the house where the user lives will hinder the scanning of the beacon base station, multiple beacon base stations are installed in the user's home according to the user's house type and the scanning area shape of the beacon base station; Specifically, taking a two-bedroom, one-living room, one-kitchen, and one-bathroom apartment as an example: install a beacon base station at each of the four corners of the living room ceiling, a beacon base station on the top of the bathroom door and the innermost top, and a beacon base station on each end of the top of the wall directly above and opposite the master bedroom door. The beacon base stations in the second bedroom are installed in the same way as the master bedroom. Moreover, the installation of the edge gateway only requires connecting the water flow sensor, power sensor, gas sensor, and door magnetic sensor via wireless or wired connections to maintain data connectivity; Finally, the data from the edge gateway and beacon base station can be uploaded to the server where the processing layer is located; S3. System Adaptive Training: Observe the processing layer's processing of environmental and posture perception data, perform manual error correction, and collaboratively optimize the module's operational processes. Repeatedly train the deep learning model until the deep learning model's processing error for posture perception data, environmental perception data, and abnormal responses is less than 5%. Specifically, the system's adaptive training is implemented with the collaborative optimization module as the core. When obvious errors occur in the processing of environmental perception data and posture perception data, manual corrections are performed to ensure that the perception data anomaly threshold calculated by the deep learning model is less than 30%. The system then enters a cyclic optimization process until the perception data anomaly threshold calculated by the deep learning model is less than 5%. Among them, the coverage area of ​​a single early warning system is planned on a community basis. Due to differences in household types, the installation locations of beacon base stations in each user's home are different. Beacon base stations need to be arranged in the manner of S2. Therefore, when the posture analysis module analyzes the beacon data obtained by scanning the beacon base stations, the data with the clearest data and the least noise among multiple beacon base stations is used as the main data, and the other data is used as fuzzy verification data; S4. Processing of alarm information: Based on the alarm information displayed on the visualization platform and the graded alarm obtained by the early warning analysis module, grid workers are dispatched to the user's home to process the alarm information; Among them, the grid worker arrives at the scene to verify the user's identity and then conducts a health assessment, and then takes targeted measures based on the user's health assessment results. If a false alarm occurs, the environmental perception unit and posture perception unit will be checked, and the false alarm information and related data will be extracted to optimize the system optimization module.

[0024] In the above content, the present invention realizes all-weather precise monitoring through multimodal perception and intelligent collaborative analysis: The sensing layer deploys pipe-clamp water flow sensors, non-intrusive current transformers, dual-mode gas sensors, and redundant door magnetic sensors to collect water, electricity, and gas door magnetic data in real time. UWB / BLE dual-mode beacons are also used to build a posture monitoring network based on head positioning, hand motion capture, torso posture, and foot pressure sensing to accurately capture the elderly's home activity status. The edge gateway cleans and analyzes environmental data in real time, identifying anomalies based on preset thresholds. The beacon base station calculates the relative distance between joints through collaborative scanning by multiple base stations, achieving millimeter-level spatial positioning. The processing layer innovatively establishes an "environment-posture" dual-engine collaborative mechanism: When the environmental analysis module detects hydropower anomalies, it immediately triggers the collaborative control module to activate posture scanning in specific areas. Risk scenarios are verified through joint vector modeling, forming a cross-validation closed loop. The application layer deploys a 3D visualization platform in the community center, transforming environmental data into a dynamic dashboard, rendering posture data into a real-time human model, and implementing intelligent responses through an early warning classification module; The collaborative optimization module drives the continuous evolution of the system through deep learning: after initial manual correction of obvious errors, a verification cycle is initiated when the analysis error enters the range of 15% to 20%. The collaborative control center re-samples data to train the model, generates new thresholds, and updates the analysis module. After multiple rounds of iteration, the error is reduced to <5%. This progressive optimization has greatly reduced the system's false alarm rate in actual measurements and improved the fall recognition rate.

[0025] During implementation, the S1-S4 standardized process is used: targeted installation of sensors, optimization of base station deployment according to household types, manual error correction training based on error thresholds, and on-site feedback from grid workers, ultimately forming a fully closed-loop intelligent monitoring system of "perception-analysis-warning-optimization".

[0026] The above description is merely a preferred embodiment of the present invention and does not constitute any other form of limitation to the present invention. Any person skilled in the art may utilize the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes for application in other fields. However, any simple modification, equivalent change, and modification of the above embodiments made in accordance with the technical essence of the present invention without departing from the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A method and system for monitoring and warning the safety status of elderly people living alone, characterized in that: include: A perception layer, comprising an environment perception unit and a posture perception unit; A transport layer comprising an edge gateway for transmitting data to an environment sensing unit and a beacon base station for transmitting data to a posture sensing unit; a processing layer comprising a posture analysis module for analyzing user data acquired by the posture perception unit, an environment analysis module for analyzing user data acquired by the environment perception unit, and a collaborative control module for mobilizing the operation of the perception layer based on the analysis results of the posture analysis module and the environment analysis module; An application layer, comprising a visualization platform for displaying the user's environmental status and posture status, and an early warning classification module for classifying and issuing early warnings based on abnormal content; The environmental sensing unit includes a water flow sensor for obtaining the user's water consumption data, a power sensor for obtaining the user's electricity consumption data, a gas sensor for obtaining the user's gas consumption data, and a door magnetic sensor for obtaining the user's door switch data. The environmental sensing unit obtains the user's original water consumption data, electricity consumption data, gas consumption data, and door magnetic sensor data, which are processed by the edge gateway and uploaded to the environmental analysis module of the processing layer for analysis; The posture sensing unit includes a head beacon composed of UWB and BLE dual-mode beacons, a hand beacon composed of a flexible RFID bracelet, a torso beacon composed of three-axis acceleration and UWB, and a foot beacon composed of a pressure-sensing insole. The head beacon, hand beacon, torso beacon and foot beacon of the posture sensing unit are scanned by the beacon base station and the relative distance between each beacon is calculated, and then uploaded to the posture analysis module for posture analysis.

2. A method and system for monitoring and warning the safety status of elderly people living alone according to claim 1, characterized in that: It also includes a collaborative optimization module constructed with the collaborative control module, early warning classification module, collaborative trigger module, early warning disposal platform and deep learning model. The collaborative optimization module obtains the abnormal water, electricity and gas usage, door magnetic opening and closing abnormality, and posture abnormality obtained by the environmental analysis module and the posture analysis module according to the preset water, electricity and gas abnormality thresholds, door magnetic opening and closing time thresholds, and beacon relative spacing thresholds. The collaborative control module sends a verification request to the posture analysis module according to the abnormal situation, and activates the beacon base station through the collaborative trigger to re-scan the posture perception unit. At the same time, the collaborative control module generates specific alarm content according to the abnormal event, and transmits the alarm content, environmental perception data and posture perception data to the early warning disposal After processing the alarm content, the early warning and disposal platform uploads the alarm content, environmental perception data, posture perception data and alarm disposal feedback to the deep learning model, analyzes the correlation between the environmental perception data and posture perception data in the abnormal event, generates new perception data thresholds and sends them to the posture analysis module and the environmental analysis module respectively, updates the settings of the posture analysis module and the environmental analysis module, and at the same time, the collaborative control module activates the beacon base station, rescans to obtain beacon data and uploads it to the posture analysis module, uses the updated settings to scan and analyze the user's body posture, and feeds back the updated settings of the posture analysis module and the environmental analysis module and the beacon scanning data to the deep learning model via the collaborative control module to optimize the deep learning model.

3. The method and system for monitoring and warning the safety status of elderly people living alone according to claim 1, characterized in that: The beacon base station scans the head beacon, hand beacon, torso beacon and foot beacon to obtain the scanning data of each beacon, and then uploads the relative position data of the beacon to the posture analysis module to calculate the relative positions between the head beacon, hand beacon, torso beacon and foot beacon, analyzes the human posture corresponding to the head beacon, hand beacon, torso beacon and foot beacon according to the human body motion model, and analyzes the abnormal data of the beacon data according to the set beacon relative spacing threshold, and uploads the abnormal data to the collaborative control module, and the collaborative control module processes the abnormal data according to the operation steps of the collaborative optimization module.

4. The safety status monitoring and early warning system for elderly people living alone according to claim 1 is characterized by: When the raw environmental data acquired by the environmental perception unit is uploaded to the transmission layer, the edge gateway processes the raw environmental data, analyzes the abnormal data in the raw environmental data according to the pre-set water, electricity, gas abnormality thresholds and the door magnetic opening and closing time thresholds, and uploads the abnormal data to the collaborative control module. The collaborative control module processes the abnormal data according to the operation steps of the collaborative optimization module.

5. The method and system for monitoring and warning the safety status of elderly people living alone according to claim 1, characterized in that: The application layer obtains the alarm information obtained by the processing layer, displays it through the visualization platform, and classifies and issues warnings to the alarm information through the warning classification module. The visualization platform is installed in the community service center.

6. A method for monitoring and warning the safety status of elderly people living alone, according to a system for monitoring and warning the safety status of elderly people living alone according to any one of claims 1 to 5, characterized in that: The following steps are involved: S1. Installation of the sensing unit: Install the environment sensing unit in the user's home, hand over the posture sensing unit to the user, and guide the user to learn how to wear the posture sensing unit; S2. Installation of relay equipment: Install the edge gateway and beacon base station in the user's home and connect them to the early warning system; S3. System Adaptive Training: Observe the processing layer's processing of environmental and posture perception data, perform manual error correction, and collaboratively optimize the module's operational processes. Repeatedly train the deep learning model until the deep learning model's processing error for posture perception data, environmental perception data, and abnormal responses is less than 5%. S4. Processing of alarm information: Based on the alarm information displayed on the visualization platform and the graded warning obtained by the early warning analysis module, grid workers are dispatched to the user's home to process the alarm information.

7. The method for monitoring and warning the safety status of elderly people living alone according to claim 6, characterized in that: In step S2, since the house where the user lives may hinder the scanning of the beacon base station, multiple beacon base stations are installed in the user's home according to the user's house type and the scanning surface shape of the beacon base station.

8. The method for monitoring and warning the safety status of elderly people living alone according to claim 7, characterized in that: In step S3, the coverage area of ​​a single early warning system is planned based on the community. Due to differences in household types, the installation locations of the beacon base stations in each user's home are different, and the beacon base stations need to be arranged in the manner of S2. Therefore, when the posture analysis module analyzes the beacon data obtained by scanning the beacon base stations, the data with the clearest data and the least noise among multiple beacon base stations is used as the main data, and the other data is used as fuzzy verification data.

9. The method for monitoring and warning the safety status of elderly people living alone according to claim 8, characterized in that: In step S3, at the initial stage of deep learning model training of the collaborative optimization module, obvious errors in the environment analysis module and the posture analysis module are manually eliminated. When the new perception data threshold derived by the deep learning model makes the analysis result errors of the posture analysis module and the environment analysis module within the range of 20% to 15%, the first round of verification is entered. The collaborative control center triggers the re-acquisition of perception data based on the abnormal items, and trains the deep learning model based on the re-acquired perception data. The new perception data threshold derived by the deep learning model is used to adjust the environment analysis module and the posture analysis module. The adjusted environment analysis module and posture analysis module then acquire new perception data, and input the analysis results into the collaborative control center. The collaborative control center enters the second round of verification based on the abnormal items until the new perception data threshold derived by the deep learning model makes the analysis result errors of the posture analysis module and the environment analysis module less than 5%.

Images