Intelligent old-age care supervision system based on AI intelligent agent

Abstract

The invention discloses an intelligent old-age care supervision system based on an AI intelligent agent, and the system comprises a wearable biological information collection intelligent agent which uses a GD32 chip as a core processing module and is used for collecting and receiving the basic information of a supervised person; the fixed environment variable acquisition agent uses an ESP8266 chip as a core processing module and is used for acquiring and receiving environment variable information; the fixed video acquisition and analysis intelligent agent is used for acquiring posture information of people in a public area and judging abnormal states of the people through AI identification; and the information processing center is used for summarizing information acquired by the three intelligent agents and sending the information to the friendly interaction intelligent agent for a user to call and check. The method is high in applicability, low in misjudgment rate of fall detection of the old people and accurate in position identification.

Description

Technical Field

[0001] This invention relates to the field of smart elderly care technology, and in particular to a smart elderly care monitoring system based on AI intelligent agents. Background Technology

[0002] Population aging and the continuous advancement of information technology are important development trends in the international community. In addressing population aging, traditional elderly care models have revealed numerous shortcomings, such as a shortage of care resources, uneven distribution of medical services, and low quality of life for the elderly. Existing research indicates that technological progress can help overcome the challenges of elderly care. The development and innovation of technologies such as the Internet, big data, and artificial intelligence have brought new opportunities to the field of elderly care. Smart elderly care, supported by intelligence, digitalization, and information technology, is an important path to actively address population aging. Technology-enabled smart elderly care has become an international development trend.

[0003] Currently, smart elderly care relies heavily on STM32 chips in the field of mobile wearable devices. However, due to the weaker operational skills of the elderly, the disadvantages of STM32 as a core chip will be amplified. At the same time, due to the updating of domestically produced equipment, the applicability of STM32 chips will be further reduced.

[0004] Current methods for fall detection in the elderly typically rely on ordinary smart wearable devices or image recognition alone, which are difficult to use accurately.

[0005] Chinese invention patent CN118411797B discloses a method and system for early warning processing in a smart elderly care platform. Step S1 involves pre-classifying early warning information into three levels: a first, a second, and a third. Step S2 involves real-time detection of early warning information using sensors and cameras. Step S3 involves converting the camera feed into corresponding background images and stick figure skeleton coordinates, storing the early warning information, background images, and stick figure skeleton coordinates in a cloud platform, and retrieving an index table of early warning information corresponding to each elderly person's information based on a time index. Step S4 involves transmitting the early warning information and stick figure skeleton coordinates to a monitoring platform. Step S5 involves the monitoring platform automatically determining the current early warning level and performing corresponding early warning processing. This invention simply categorizes early warning levels, leading to a high probability of misjudgment.

[0006] Chinese invention patent CN113113145B describes an integrated intelligent platform for smart home-based elderly care service management based on remote monitoring and video processing. It includes modules for acquiring basic information on elderly residents in the community, classifying and screening elderly individuals, acquiring standard medication parameters for elderly individuals with underlying medical conditions, using smart wearable terminals for data entry, collecting elderly individuals' health parameters, establishing a health parameter database, modeling and analyzing health coefficients, monitoring emergency video surveillance, a management server, and a remote service center. However, this invention only assesses data collected directly using simple methods and cannot effectively utilize video data resources, making it difficult to identify the most significant risk points for falls among the elderly. Summary of the Invention

[0007] This invention provides an AI-based intelligent elderly care monitoring system to address the technical problems of existing intelligent elderly care monitoring systems that rely excessively on STM32 chips, have low applicability, high false positive rates in fall detection, and inaccurate location recognition.

[0008] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a smart elderly care monitoring system based on AI intelligent agents, comprising: Wearable bio-information collection agent, fixed environmental variable collection agent, fixed video collection and analysis agent, information processing center, and user-friendly interactive agent; The wearable bio-information collection smart body uses the GD32 chip as the core processing module to collect and receive basic information of the monitored personnel, including heart rate, blood pressure, location, and status. The fixed environmental variable acquisition agent uses the ESP8266 chip as the core processing module to collect and receive environmental variable information, including smoke information and temperature and humidity information. The fixed video acquisition and analysis intelligent agent is used to collect physical information of people in public areas and to identify and judge abnormal states of people through AI. The information processing center is used to summarize the information collected by the above three intelligent agents and send it to the user-friendly interactive intelligent agent for users to call and view.

[0009] Furthermore, the GD32 chip used in the wearable bio-information collection smart device is specifically the GD32F103C8T6 chip; The wearable bio-information collection smart device also includes a heart rate and blood pressure module, a Beidou positioning module, a gyroscope module, and an RFID tag; The heart rate and blood pressure module uses the MKB0908 module to collect the heart rate and blood pressure of the monitored personnel, and transmits the heart rate and blood pressure of the monitored personnel to the GD32F103C8T6 chip through a Universal Synchronous Asynchronous Transceiver (USART). The Beidou positioning module uses the SKG09A model positioning module and transmits the location information of the monitored personnel to the GD32F103C8T6 chip through the full-duplex asynchronous serial communication interface UART. The gyroscope module uses an MPU6050 gyroscope and transmits the status information of the monitored personnel to the GD32F103C8T6 chip via the integrated circuit bus (IIC). The status information of the monitored personnel includes pitch angle, roll angle, and yaw angle; The RFID tag is used to identify monitored personnel and is suitable for positioning in indoor environments.

[0010] Furthermore, the fixed environmental variable acquisition agent and the fixed video acquisition and analysis agent are deployed in the living area of ​​the supervised personnel, so that all non-privacy areas are fully covered; The information processing center is deployed on a cloud server; The user-friendly interactive agent is deployed on web pages and mobile devices.

[0011] Furthermore, the fixed environmental variable acquisition agent includes a smoke sensor and a temperature and humidity sensor; The smoke sensor uses the MQ-2 model and transmits smoke information to the ESP8266 chip via an ADC; The temperature and humidity sensor uses a DHT11 signal temperature and humidity sensor, which transmits the temperature and humidity signals to the ESP8266 chip via a full-duplex asynchronous serial communication interface (UART).

[0012] Furthermore, the fixed video acquisition and analysis intelligent agent includes a camera, an Nginx server, a local database, and an RFID reader; The camera collects video information of the monitored personnel, load balances the data through the Nginx server, stores it in a local database, and sends it to the information processing center for AI analysis. The RFID reader is used to read surrounding RFID tags to achieve indoor positioning of the monitored personnel.

[0013] Furthermore, the information processing center receives and classifies the monitored information collected by the wearable biometric information collection agent, the fixed environmental variable collection agent, and the fixed video collection and analysis agent; The classification analysis involves directly outputting temperature and humidity, ambient temperature and humidity, RFID reader information, and BeiDou positioning information to the user-friendly interactive intelligent agent. For heart rate and blood pressure, the data is compared with the normal range and output, and it is determined whether the data exceeds the normal range. The result is then output to the user-friendly interactive intelligent agent. For pitch angle, roll angle, yaw angle, and video information of the monitored personnel, the YOLOv11 Pose model is optimized to determine whether the situation is dangerous, and the determination result is sent to the user-friendly interactive intelligent agent.

[0014] Furthermore, the step of determining whether a dangerous state is in effect by optimizing the YOLOv11 Pose model is as follows: S1, using pitch angle, roll angle, and yaw angle as digital attitude judgment standards, calculate the angular velocity vector magnitude per unit time, where the unit time is denoted as 0.5s, and determine the range. The normal range for pitch angle is -30° to +30°, and the fall characteristic range is ±80° or more. The normal range for roll angle is -45° to +45°, and the fall characteristic range is ±70° or more. The normal range for yaw angle is <20° / 0.5 seconds, and the fall characteristic range is >45° / 0.5 seconds. When pitch angle, roll angle, and yaw angle are all within the fall characteristic range, determine the probability that the monitored person is in a suspected fall state, SD1. S2 uses the YOLOv11 Pose model to analyze the video information of the monitored person, extracts 17 key points of the human body as a skeleton model to simulate human movements, and continuously processes 20 key sequences through a gated loop unit to determine whether there is a fall action and outputs the probability of a suspected fall state SD2. S3, when SD1 and SD2 are both in a suspected fall state, it is determined that the monitored user is in a suspected fall state, which is used to prevent false alarms of other normal actions. The probability of the suspected fall state is set to 80%.

[0015] Furthermore, the friendly interactive intelligent agent is used to receive data from the information processing center via HTTPS and display it to the regulated personnel and their families.

[0016] The beneficial effects of the technical solution provided by this invention include at least the following: This invention comprises a wearable biometric data collection agent, a fixed environmental variable data collection agent, a fixed video acquisition and analysis agent, an information processing center, and a user-friendly interactive agent. The wearable biometric data collection agent uses a GD32 chip as its core processing module to collect and receive basic information about monitored individuals, including heart rate, blood pressure, location, and status. The fixed environmental variable data collection agent uses an ESP8266 chip as its core processing module to collect and receive environmental variable information, including smoke information and temperature and humidity information. The fixed video acquisition and analysis agent collects physical information of individuals in public areas and uses AI to identify and determine abnormal states. The information processing center aggregates the information collected by the three agents and sends it to the user-friendly interactive agent for user access and viewing. This collaborative approach between different agents enables highly efficient monitoring of monitored individuals.

[0017] In this invention, the GD32F103C8T6 chip is used in combination with a heart rate and blood pressure module, a Beidou positioning module, a gyroscope module, and an RFID tag to replace the traditional STM32 chip in a wearable bio-information collection smart body. This method has wide applicability and good compatibility with domestic equipment.

[0018] This invention uses three methods—a combination of BeiDou positioning modules, RFID tags and RFID readers, and video capture—to locate monitored personnel with high accuracy.

[0019] This invention reduces the probability of false positives by optimizing the YOLOv11 Pose model to determine whether a state is dangerous. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of an AI-based smart elderly care supervision system provided in an embodiment of the present invention; Figure 2 A schematic diagram of the pinout of the GD32F103C8T6 chip provided in an embodiment of the present invention; Figure 3 This is a standing test image from the YOLOv11 Pose model. Figure 4 This is a fall test diagram from the YOLOv11 Pose model. Figure 5 This is a schematic diagram of a user-friendly interactive intelligent agent on a web page. Figure 6 This is a schematic diagram of a user-friendly interactive smart agent mobile app. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0023] This embodiment provides a smart elderly care monitoring system based on AI intelligent agents, including: Wearable bio-information collection agent, fixed environmental variable collection agent, fixed video collection and analysis agent, information processing center, and user-friendly interactive agent.

[0024] Please refer to Figure 1 A schematic diagram of an AI-based intelligent elderly care supervision system provided in this embodiment of the invention.

[0025] The wearable bio-information collection smart body uses the GD32 chip as the core processing module to collect and receive basic information of the monitored personnel, including heart rate, blood pressure, location, and status. The fixed environmental variable acquisition agent uses the ESP8266 chip as the core processing module to collect and receive environmental variable information, including smoke information and temperature and humidity information. The fixed video acquisition and analysis intelligent agent is used to collect physical information of people in public areas and to identify and judge abnormal states of people through AI. The information processing center is used to summarize the information collected by the above three intelligent agents and send it to the user-friendly interactive intelligent agent for users to call and view.

[0026] 1. The specific model of the GD32 chip used in the wearable biometric data collection device is GD32F103C8T6. Please refer to... Figure 2 A schematic diagram of the pinout of the GD32F103C8T6 chip provided in this embodiment of the invention.

[0027] The wearable bio-information collection smart device also includes a heart rate and blood pressure module, a Beidou positioning module, a gyroscope module, and RFID tags.

[0028] It should be noted that the GD32F103C8T6 chip is a 32-bit microcontroller developed by GigaDevice based on the ARM Cortex-M3 core. It is a domestically produced microcontroller used to replace the STM32 chip. Its biggest advantage is that it is almost compatible with all the functions of the STM32 chip, and its main frequency is 108MHz, which is a significant performance improvement compared to the 72MHz of the STM32 chip. Moreover, as it is a domestically developed product, it has a more significant advantage when purchasing.

[0029] The wearable biometric data acquisition device integrates a heart rate and blood pressure module, a Beidou positioning module, a gyroscope module, and RFID tags.

[0030] Compared to traditional smart bracelets and other devices, the addition of RFID tags makes them particularly suitable for accurate identification of monitored individuals in nursing homes, communities, and other similar settings due to RFID's short range and high accuracy.

[0031] In this embodiment, the RFID tag uses an EM-18 RFID module, which is connected to the GD32F103C8T6 chip via a UART serial port at a baud rate of 9600. It directly receives the 12-byte card number, specifically:

[0032] The heart rate and blood pressure module uses the MKB0908 module to collect the heart rate and blood pressure of the monitored personnel, and transmits the heart rate and blood pressure of the monitored personnel to the GD32F103C8T6 chip via a Universal Synchronous Asynchronous Transceiver (USART). Specifically:

[0033] The Beidou positioning module uses the SKG09A model and transmits the location information of the monitored personnel to the GD32F103C8T6 chip via a full-duplex asynchronous serial communication interface (UART). Specifically:

[0034] The gyroscope module uses an MPU6050 model gyroscope and transmits the status information of the monitored personnel to the GD32F103C8T6 chip via the integrated circuit bus (IIC). Specifically:

[0035] The status information of the monitored personnel includes pitch angle, roll angle, and yaw angle.

[0036] It should be noted that pitch angle, roll angle, and yaw angle measurements are basic functions of a gyroscope.

[0037] The RFID tag is used to identify monitored personnel and is suitable for positioning in indoor environments.

[0038] It should be noted that RFID tags are used in conjunction with RFID readers.

[0039] 2. Fixed environmental variable acquisition agents and fixed video acquisition and analysis agents are deployed in the living areas of the monitored personnel, ensuring full coverage of non-privacy areas; The information processing center is deployed on a cloud server.

[0040] It should be noted that in this embodiment, the cloud server selected is a general-performance GPU server from Alibaba Cloud, which can meet the general needs of smart elderly care and also realize image analysis.

[0041] The user-friendly interactive agent is deployed on web pages and mobile devices.

[0042] It should be noted that the main function of the webpage and mobile app is to provide a visual presentation for family, friends, and supervisors, making it easy to view and manage.

[0043] 3. The fixed environmental variable acquisition agent includes a smoke sensor and a temperature and humidity sensor; The smoke sensor uses the MQ-2 model and transmits smoke information to the ESP8266 chip via an ADC. Specifically:

[0044] The temperature and humidity sensor used is a DHT11 model. It transmits the temperature and humidity signals to the ESP8266 chip via a full-duplex asynchronous serial communication interface (UART). Specifically:

[0045] 4. The fixed video acquisition and analysis intelligent agent includes a camera, an Nginx server, a local database, and an RFID reader; The camera collects video information of the monitored personnel, load balances the data through the Nginx server, stores it in a local database, and sends it to the information processing center for AI analysis. The RFID reader is used to read surrounding RFID tags to achieve indoor positioning of the monitored personnel.

[0046] 5. The information processing center receives the monitored information collected by the wearable biometric information collection agent, the fixed environmental variable collection agent, and the fixed video collection and analysis agent, and performs classification and analysis on the information. The classification analysis involves directly outputting temperature and humidity, ambient temperature and humidity, RFID reader information, and BeiDou positioning information to the user-friendly interactive intelligent agent. For heart rate and blood pressure, the data is compared with the normal range and output, and it is determined whether the data exceeds the normal range. The result is then output to the user-friendly interactive intelligent agent. For pitch angle, roll angle, yaw angle, and video information of the monitored personnel, the YOLOv11 Pose model is optimized to determine whether the situation is dangerous, and the determination result is sent to the user-friendly interactive intelligent agent.

[0047] 6. The optimized steps for determining whether a YOLOv11 Pose model is in a dangerous state are as follows: S1, using pitch angle, roll angle, and yaw angle as digital attitude judgment standards, calculate the angular velocity vector magnitude per unit time, where the unit time is denoted as 0.5s, and determine the range. The normal range for pitch angle is -30° to +30°, and the fall characteristic range is ±80° or more. The normal range for roll angle is -45° to +45°, and the fall characteristic range is ±70° or more. The normal range for yaw angle is <20° / 0.5 seconds, and the fall characteristic range is >45° / 0.5 seconds. When pitch angle, roll angle, and yaw angle are all within the fall characteristic range, determine the probability that the monitored person is in a suspected fall state, SD1. S2 uses the YOLOv11 Pose model to analyze the video information of the monitored personnel. First, 17 human key points are extracted from each frame of the image. 20 consecutive frames are taken to form a sequence. After missing points are filled by KNN, the sequence is fed into a two-layer 64-unit GRU. Finally, the Sigmoid outputs the probability of falling and outputs the probability of suspected fall SD2. Please refer to Figure 3 Standing test diagram in YOLOv11 Pose model and Figure 4 Fall test diagram in YOLOv11 Pose model.

[0048] S3, when SD1 and SD2 are both in a suspected fall state, it is determined that the monitored user is in a suspected fall state, which is used to prevent false alarms of other normal actions. The probability of the suspected fall state is set to 80%.

[0049] A user-friendly interactive intelligent agent is used to receive data from the information processing center via HTTPS and display it to the monitored individuals and their families. This information is accessed through a web page and a mobile app. Please refer to [link / reference]. Figure 5 A diagram of a user-friendly interactive intelligent agent on a web page. Figure 6 A schematic diagram of the user-friendly interactive intelligent agent mobile app.

[0050] Furthermore, it should be noted that the present invention can be provided as a method, apparatus, or computer program product. Therefore, embodiments of the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, embodiments of the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code.

[0051] Embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0052] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing terminal equipment to cause a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0053] It should also be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0054] Finally, it should be noted that the above description represents a preferred embodiment of the present invention. It should be pointed out that although preferred embodiments have been described, those skilled in the art, once they understand the basic inventive concept of the present invention, can make various improvements and modifications without departing from the principles described herein. These improvements and modifications should also be considered within the scope of protection of the present invention. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the embodiments of the present invention.

Claims

1. A smart elderly care monitoring system based on AI intelligent agents, characterized in that, include: Wearable bio-information collection agent, fixed environmental variable collection agent, fixed video collection and analysis agent, information processing center, and user-friendly interactive agent; The wearable bio-information collection smart body uses the GD32 chip as the core processing module to collect and receive basic information of the monitored personnel, including heart rate, blood pressure, location, and status. The fixed environmental variable acquisition agent uses the ESP8266 chip as the core processing module to collect and receive environmental variable information, including smoke information and temperature and humidity information. The fixed video acquisition and analysis intelligent agent is used to collect physical information of people in public areas and to identify and judge abnormal states of people through AI. The information processing center is used to summarize the information collected by the above three intelligent agents and send it to the user-friendly interactive intelligent agent for users to call and view.

2. The smart elderly care supervision system based on AI intelligent agents as described in claim 1, characterized in that: The specific model of the GD32 chip used in the wearable bio-information collection intelligent agent is GD32F103C8T6 chip. The wearable bio-information collection smart device also includes a heart rate and blood pressure module, a Beidou positioning module, a gyroscope module, and an RFID tag; The heart rate and blood pressure module uses the MKB0908 module to collect the heart rate and blood pressure of the monitored personnel, and transmits the heart rate and blood pressure of the monitored personnel to the GD32F103C8T6 chip through a Universal Synchronous Asynchronous Transceiver (USART). The Beidou positioning module uses the SKG09A model positioning module and transmits the location information of the monitored personnel to the GD32F103C8T6 chip through the full-duplex asynchronous serial communication interface UART. The gyroscope module uses an MPU6050 gyroscope and transmits the status information of the monitored personnel to the GD32F103C8T6 chip via the integrated circuit bus (IIC). The status information of the monitored personnel includes pitch angle, roll angle, and yaw angle; The RFID tag is used to identify monitored personnel and is suitable for positioning in indoor environments.

3. The smart elderly care supervision system based on AI intelligent agents as described in claim 1, characterized in that: The fixed environmental variable acquisition agent and the fixed video acquisition and analysis agent are deployed in the living area of ​​the supervised personnel, so that all non-privacy areas are fully covered; The information processing center is deployed on a cloud server; The user-friendly interactive agent is deployed on web pages and mobile devices.

4. The smart elderly care supervision system based on AI intelligent agents as described in claim 1, characterized in that: The fixed environmental variable acquisition agent includes a smoke sensor and a temperature and humidity sensor; The smoke sensor uses the MQ-2 model and transmits smoke information to the ESP8266 chip via an ADC; The temperature and humidity sensor uses the DHT11 model and transmits the temperature and humidity signals to the ESP8266 chip via a full-duplex asynchronous serial communication interface (UART).

5. The intelligent elderly care supervision system based on AI intelligent agents as described in claim 1, characterized in that: The fixed video acquisition and analysis intelligent agent includes a camera, an Nginx server, a local database, and an RFID reader; The camera collects video information of the monitored personnel, load balances the data through the Nginx server, stores it in a local database, and sends it to the information processing center for AI analysis. The RFID reader is used to read surrounding RFID tags to achieve indoor positioning of the monitored personnel.

6. The smart elderly care supervision system based on AI intelligent agents as described in claim 1, characterized in that: The information processing center receives and classifies the monitored information collected by the wearable biometric information collection agent, the fixed environmental variable collection agent, and the fixed video collection and analysis agent. The classification analysis involves directly outputting temperature and humidity, ambient temperature and humidity, RFID reader information, and BeiDou positioning information to the user-friendly interactive intelligent agent. For heart rate and blood pressure, the data is compared with the normal range and output, and it is determined whether the data exceeds the normal range. The result is then output to the user-friendly interactive intelligent agent. For pitch angle, roll angle, yaw angle, and video information of the monitored personnel, the YOLOv11 Pose model is optimized to determine whether the situation is dangerous, and the determination result is sent to the user-friendly interactive intelligent agent.

7. The smart elderly care supervision system based on AI intelligent agents as described in claim 1, characterized in that, The step of determining whether a dangerous state is in effect by optimizing the YOLOv11 Pose model is as follows: S1, using pitch angle, roll angle, and yaw angle as digital attitude judgment standards, calculate the angular velocity vector magnitude per unit time, where the unit time is denoted as 0.5s, and determine the range. The normal range for pitch angle is -30° to +30°, and the fall characteristic range is ±80° or more. The normal range for roll angle is -45° to +45°, and the fall characteristic range is ±70° or more. The normal range for yaw angle is <20° / 0.5 seconds, and the fall characteristic range is >45° / 0.5 seconds. When pitch angle, roll angle, and yaw angle are all within the fall characteristic range, determine the probability that the monitored person is in a suspected fall state, SD1. S2 uses the YOLOv11 Pose model to analyze the video information of the monitored person, extracts 17 key points of the human body as a skeleton model to simulate human movements, and continuously processes 20 key sequences through a gated loop unit to determine whether there is a fall action and outputs the probability of a suspected fall state SD2. S3, when SD1 and SD2 are both in a suspected fall state, it is determined that the monitored user is in a suspected fall state, which is used to prevent false alarms of other normal actions. The probability of the suspected fall state is set to 80%.

8. The smart elderly care supervision system based on AI intelligent agents as described in claim 1, characterized in that: The user-friendly interactive intelligent agent is used to receive data from the information processing center via HTTPS and display it to the supervised personnel and their families. The information is delivered through a web page and a mobile app.

Images