Alzheimer's disease risk assessment and intervention system and method

Abstract

The invention relates to the technical field of intelligent medical treatment, health internet of things and artificial intelligence, in particular to an Alzheimer's disease risk assessment and intervention system and method. Comprising a data acquisition layer, a data transmission and network layer, a cloud intelligent analysis and management platform, a personalized intervention strategy library and generation module, a user interaction and application layer and an old man interaction interface. The data acquisition layer comprises a household non-inductive perception layer; the family non-inductive perception layer is composed of a plurality of Internet of Things sensor nodes deployed in the home of a user and is used for automatically collecting amnesia behaviors, life rhythm and activity modes and active interaction and cognitive stimulation multi-dimensional behavior data; a cloud artificial intelligence algorithm is utilized to analyze the long-term change trend, an individual cognitive function decline risk model is constructed, personalized cognitive training and life intervention suggestions are provided accordingly, and a digital management closed loop of'evaluation-early warning-intervention-re-evaluation 'is formed.

Description

Technical Field

[0001] This invention relates to the technical fields of smart healthcare, health IoT and artificial intelligence, and in particular to an Alzheimer's disease risk assessment and intervention system and method. Background Technology

[0002] Alzheimer's disease (AD) is a neurodegenerative disease with insidious onset and progressive development. Its early symptoms (such as memory loss, executive function decline, and changes in lifestyle) are extremely subtle and individualized, and are often mistaken for normal aging, leading to delayed diagnosis and missing the best window of intervention.

[0003] Current clinical diagnosis relies heavily on neuropsychological scales (such as MMSE and MoCA) and imaging examinations, but these methods have the following limitations: 1. Randomness and subjectivity: The scale assessment is a discrete time point evaluation, which is greatly affected by the test subject's emotions, environment and tester, and cannot reflect continuous changes in real daily life. 2. Poor accessibility: Professional examinations are expensive and require visits to medical institutions, making them unsuitable for large-scale, routine early screening and disease monitoring; 3. Delay: When scales or imaging studies reveal obvious abnormalities, the pathological changes have often already progressed to the middle or late stages.

[0004] Some existing smart home or wearable devices attempt to manage health by monitoring single dimensions such as activity level and heart rate, but they cannot effectively capture complex behavioral patterns that are highly correlated with cognitive decline (such as the frequency of forgetfulness, disruption of daily rhythms, and changes in communication ability). At the same time, existing technologies mostly focus on "abnormal event alarms" (such as fall alarms) and lack the ability to quantitatively model and assess the risks of slow, long-term "behavioral trend deterioration". In addition, the assessment and intervention links are usually separated, failing to form a closed loop of personalized intervention based on data feedback.

[0005] Therefore, there is an urgent need in this field for a comprehensive solution that can collect behavioral data in a seamless, continuous, and multi-dimensional manner in natural life scenarios, mine early digital biomarkers of cognitive decline through long-term trend analysis, and provide personalized and actionable intervention measures to achieve early detection, early warning, and early intervention of Alzheimer's disease (AD). Summary of the Invention

[0006] To address the aforementioned technical issues, this invention provides an Alzheimer's disease risk assessment and intervention system and method that automatically collects multi-dimensional daily behavioral data related to cognitive function by deploying a non-invasive sensor network in the home, analyzes the long-term trends of these data using cloud-based artificial intelligence algorithms, constructs an individual cognitive function decline risk model, and provides personalized cognitive training and lifestyle intervention suggestions accordingly, forming a digital management closed loop of "assessment-early warning-intervention-reassessment".

[0007] The present invention provides an Alzheimer's disease risk assessment and intervention system, comprising a data acquisition layer, a data transmission and network layer, a cloud-based intelligent analysis and management platform, a personalized intervention strategy library and generation module, a user interaction and application layer, and an elderly interaction interface. The data acquisition layer includes a home-based seamless sensing layer; The home-based seamless sensing layer consists of multiple IoT sensor nodes deployed in the user's home, used to automatically collect multi-dimensional behavioral data such as forgetfulness, daily rhythms and activity patterns, and proactive interactions and cognitive stimuli. Data transmission and network layer: Used to transmit the collected data from IoT sensor nodes to a cloud-based intelligent analysis and management platform via home Wi-Fi and / or low-power wide area network; Cloud-based intelligent analysis and management platform: including a multimodal data fusion and storage module, a behavioral digital biomarker extraction engine, and a long-term trend analysis and risk assessment module; ² Multimodal data fusion and storage module: Receives and cleans heterogeneous data from different sensors, aligns and stores it according to time series, and builds a "behavioral digital twin" database for each user; Behavioral digital biomarker extraction engine: Extracts quantitative features related to cognitive function from raw data; Long-term trend analysis and risk assessment module: This module performs long-term time-series analysis on the extracted digital biomarkers, establishes an individual baseline for each feature, and uses statistical models and machine learning algorithms to identify its slow but continuous trend of deterioration. By integrating the deterioration trends from multiple dimensions, it generates a comprehensive cognitive function decline risk score and low, medium, and high risk levels. Personalized intervention strategy library and generation module: used to store various digital intervention solutions, and automatically generate and distribute personalized intervention plans based on the user's risk level and specific weaknesses; User interaction and application layer: Use a client APP for family members and caregivers to view risk assessment reports, trend charts, receive early warning notifications, and monitor the implementation of intervention plans; The elderly user interface allows them to receive and execute intervention tasks via a voice assistant, engage in daily cognitive training games, receive life reminders, and participate in social conversations.

[0008] Preferably, the extraction of quantitative features related to cognitive function from the raw data includes: The weekly average incidence of forgetful events, the complexity of daytime activity paths, the weekly coefficient of variation of nighttime waking-up-to-the-toilet frequency, and the average response latency when conversing with a voice assistant.

[0009] Preferably, the IoT sensor nodes include a forgetfulness behavior monitoring node, a daily rhythm and activity pattern monitoring node, an active interaction and cognitive stimulation node, a water immersion sensor, a human infrared sensor, a smart socket, a door magnetic sensor, a sleep monitoring strip, a toilet flushing sensor, a home wireless LAN, a voice assistant, a smart pillbox, a thermometer and hygrometer, a smart night light, and an SOS press alarm. Forgetfulness monitoring nodes: By utilizing door magnetic sensors on the front door and refrigerator door, smart sockets, water immersion sensors, and smart pillboxes, the frequency of forgetfulness events related to short-term memory and executive function is quantified and recorded. Circadian rhythm and activity pattern monitoring nodes: Using human infrared sensors deployed in each room, sleep monitoring belts in bedrooms, and toilet flushing sensors in bathrooms, the system collects users' activity trajectories, spatial transfer patterns, toilet frequency and duration, and daytime bed rest time to analyze the regularity of circadian rhythms and subtle changes in activity levels. Proactive interaction and cognitive stimulation nodes: Utilize the voice assistant to initiate structured dialogue interactions periodically, and record the response time, accuracy, and fluency data of the interaction process.

[0010] Preferably, the data transmission and network layer further includes an encryption unit; Encryption unit: Used to encrypt the collected data before transmitting it to the cloud-based intelligent analysis and management platform.

[0011] Preferably, the data transmission and network layer further includes an IoT gateway; The IoT gateway connects to the home's wireless LAN and low-power WAN, enabling devices within the home's seamless sensing layer to access the cloud-based intelligent analysis and management platform.

[0012] Preferably, it also includes an infrared sensor for the entrance door; An infrared sensor is installed at the entrance door to monitor when people leave and return home.

[0013] Preferably, a thermometer and hygrometer are also included; Thermometers and hygrometers are installed in various spaces in a user's home to monitor indoor temperature and humidity.

[0014] Preferred options also include smart night lights; Smart nightlights are installed in various spaces and hallways throughout the home to automatically sense nighttime activity and provide illumination.

[0015] Preferably, it also includes an SOS press alarm; SOS alarm buttons are installed in various spaces throughout the home, allowing residents to press a button to trigger the alarm.

[0016] A preferred method for Alzheimer's disease risk assessment and intervention includes the following steps: S1. Seamless and routine data collection: After system deployment, each sensor starts working continuously, the voice assistant initiates interactions as planned, and all data is automatically uploaded to the cloud. S2. Baseline Establishment and Feature Learning: In the initial stage, the system learns the user's behavioral patterns under normal conditions and establishes the individual baseline level of each digital biomarker. S3. Long-term trend analysis and risk assessment: The platform periodically analyzes the changing trends of each feature relative to the baseline. When the system detects that multiple key features show a statistically significant synergistic deterioration trend, it determines that the risk of cognitive decline has increased and generates a risk assessment report. S4. Early Warning Triggering and Personalized Intervention: a. Push risk warnings and detailed explanations to family members' mobile applications; b. At the same time, based on the risk profile, intervention modules are called and combined from the strategy library to generate personalized solutions; for users with obvious forgetfulness trends, timed reminders are strengthened through the voice assistant; for users with reduced activity, family members are advised to encourage specific activities through the APP; for users with declining communication skills, the difficulty and frequency of daily structured dialogues by the voice assistant are increased, and educational game tasks are pushed on the APP. S5. Intervention Feedback and Model Iteration: The system continuously monitors changes in behavioral data after the implementation of intervention measures, evaluates the effectiveness of the intervention, and dynamically adjusts the risk model and intervention strategies to form a management closed loop.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. Achieve true early risk warning: Shift from relying on discrete clinical examinations to assessments based on continuous behavioral trends in daily life, enabling the detection of extremely subtle and early functional changes, and significantly advancing the warning time. 2. Objective assessment with high ecological validity: The data comes from non-intrusive monitoring in the natural home environment, which is completely objective, avoids subjective scale bias, and truly reflects the user's daily cognitive function status. 3. Multi-dimensional integrated analysis for precise profiling: By integrating data from multiple dimensions such as forgetfulness, daily rhythms, activity patterns, and language abilities, a comprehensive digital profile of cognitive function is constructed, making risk assessment more accurate and reliable; 4. Forming an "Assessment-Intervention" Closed Loop: It is the first to directly link the risk assessment results based on the Internet of Things with digital intervention methods, realizing the leap from passive monitoring to proactive intervention and enhancing the practical value of the system; 5. Privacy-friendly and highly compliant: It uses non-video sensors to protect user privacy, and the seamless monitoring and voice interaction are natural, providing a good user experience and making it easy to maintain long-term use. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the system structure of the present invention; Figure 2 This is a schematic diagram of the installation of the infrared sensor on the entrance door; Figure 3 This is a schematic diagram of the installation of a human infrared sensor; Figure 4 This is a diagram illustrating the mapping relationship between multi-dimensional behavioral data and cognitive function; Figure 5 This is a schematic diagram illustrating the mapping relationship between multi-dimensional behavioral data and cognitive function; Figure 6 This is a flowchart of the closed-loop workflow for long-term trend analysis, early warning, and intervention.

[0019] The following are labeled in the attached diagram: 101, Human Infrared Sensor; 102, Smart Socket; 103, Door Magnetic Sensor; 104, Sleep Monitoring Strip; 105, Toilet Flushing Sensor; 106, Home Wireless LAN; 107, Entrance Door Infrared Sensor; 108, Voice Assistant; 109, Smart Pillbox; 110, Thermohygrometer; 111, Smart Night Light; 112, SOS Press Alarm. Detailed Implementation

[0020] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. The present invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete. Example

[0021] like Figures 1 to 6 As shown, the Alzheimer's disease risk assessment and intervention system of the present invention includes a data acquisition layer, a data transmission and network layer, a cloud-based intelligent analysis and management platform, a personalized intervention strategy library and generation module, a user interaction and application layer, and an elderly interaction interface. The data acquisition layer includes a home-based seamless sensing layer; The home-based seamless sensing layer consists of multiple IoT sensor nodes deployed in the user's home, used to automatically collect multi-dimensional behavioral data such as forgetfulness, daily rhythms and activity patterns, and proactive interactions and cognitive stimuli. Data transmission and network layer: Used to transmit the collected data from IoT sensor nodes to a cloud-based intelligent analysis and management platform via home Wi-Fi and / or low-power wide area network; Cloud-based intelligent analysis and management platform: including a multimodal data fusion and storage module, a behavioral digital biomarker extraction engine, and a long-term trend analysis and risk assessment module; ² Multimodal data fusion and storage module: Receives and cleans heterogeneous data from different sensors, aligns and stores it according to time series, and builds a "behavioral digital twin" database for each user; Behavioral digital biomarker extraction engine: Extracts quantitative features related to cognitive function from raw data; Long-term trend analysis and risk assessment module: This module performs long-term time-series analysis on the extracted digital biomarkers, establishes an individual baseline for each feature, and uses statistical models and machine learning algorithms to identify its slow but continuous trend of deterioration. By integrating the deterioration trends from multiple dimensions, it generates a comprehensive cognitive function decline risk score and low, medium, and high risk levels. Personalized intervention strategy library and generation module: used to store various digital intervention solutions, and automatically generate and distribute personalized intervention plans based on the user's risk level and specific weaknesses; User interaction and application layer: Use a client APP for family members and caregivers to view risk assessment reports, trend charts, receive early warning notifications, and monitor the implementation of intervention plans; The elderly user interface allows them to receive and execute intervention tasks through the voice assistant 108, engage in daily cognitive training games, receive life reminders, and participate in social conversations. The extraction of quantitative features related to cognitive function from raw data includes: The weekly average incidence of forgetful events, the complexity of daytime activity paths, the weekly coefficient of variation of nighttime waking-up-to-the-toilet frequency, and the average response latency when conversing with the voice assistant 108. Example

[0022] Based on Example 1, the present invention provides an Alzheimer's disease risk assessment and intervention system, wherein the Internet of Things sensor nodes include a forgetfulness behavior monitoring node, a daily rhythm and activity pattern monitoring node, an active interaction and cognitive stimulation node, a water immersion sensor, a human infrared sensor 101, a smart socket 102, a door magnetic sensor 103, a sleep monitoring belt 104, a toilet flushing sensor 105, a home wireless LAN 106, a voice assistant 108, a smart pillbox 109, a thermometer and hygrometer 110, a smart night light 111, and an SOS press alarm 112; Forgetfulness monitoring node: By utilizing the door magnetic sensor 103 on the front door and refrigerator door, smart socket 102, water immersion sensor and smart pillbox 109, the frequency of forgetfulness events related to short-term memory and executive function is quantitatively recorded; Circadian rhythm and activity pattern monitoring nodes: Using human infrared sensors 101 deployed in each room, sleep monitoring belts 104 in the bedroom, and toilet flushing sensors 105 in the bathroom, the activity trajectory, spatial transfer patterns, toilet frequency and duration, and daytime bed rest duration are collected to analyze the regularity of circadian rhythm and subtle changes in activity capacity. Active interaction and cognitive stimulation nodes: Use the voice assistant 108 to initiate structured dialogue interactions periodically and record the response time, accuracy, and fluency data of the interaction process; The data transmission and network layer also includes an encryption unit; Encryption unit: Used to encrypt the collected data before transmitting it to the cloud-based intelligent analysis and management platform; The data transmission and network layer also includes an IoT gateway; The IoT gateway connects to the home wireless LAN and low-power WAN, enabling devices within the home's seamless sensing layer to access the cloud-based intelligent analysis and management platform. It also includes the entrance door infrared sensor 107; An infrared sensor 107 is installed at the entrance door to monitor when people leave and return home. It also includes a thermometer and hygrometer 110; The temperature and humidity meter 110 is installed in various spaces in the user's home to detect indoor temperature and humidity; It also includes the Smart Night Light 111; The Smart Night Light 111 is installed in various spaces and corridors throughout the home to automatically sense nighttime activity and provide lighting. It also includes the SOS press alarm 112; The SOS press alarm 112 is installed in various spaces in the home for residents to press to trigger the alarm. In this embodiment, true early risk warning is achieved: the approach shifts from relying on discrete clinical examinations to assessments based on continuous behavioral trends in daily life, enabling the detection of extremely subtle and early functional changes, and significantly advancing the warning time. The assessment is objective and has high ecological validity: the data comes from non-observable monitoring in the natural home environment, which is completely objective, avoids subjective scale bias, and truly reflects the user's daily cognitive function status. Multi-dimensional integrated analysis for precise profiling: By integrating data from multiple dimensions such as forgetfulness, daily rhythms, activity patterns, and language abilities, a comprehensive digital profile of cognitive function is constructed, making risk assessment more accurate and reliable; Forming an "assessment-intervention" closed loop: It is the first to directly link the risk assessment results based on the Internet of Things with digital intervention methods such as voice interaction, cognitive games and intelligent reminders, realizing the leap from passive monitoring to active intervention and enhancing the practical value of the system; Privacy-friendly and highly compliant: It uses non-video sensors to protect user privacy, and the non-intrusive monitoring and voice interaction are natural, providing a good user experience and making it easy to maintain long-term use. Example

[0023] The present invention provides a method for Alzheimer's disease risk assessment and intervention, comprising the following steps: S1. Seamless and routine data collection: After system deployment, each sensor starts working continuously, the voice assistant 108 initiates interaction as planned, and all data is automatically uploaded to the cloud. S2. Baseline Establishment and Feature Learning: In the initial stage, the system learns the user's behavioral patterns under normal conditions and establishes the individual baseline level of each digital biomarker. S3. Long-term trend analysis and risk assessment: The platform periodically analyzes the changing trends of each feature relative to the baseline. When the system detects that multiple key features show a statistically significant synergistic deterioration trend, it determines that the risk of cognitive decline has increased and generates a risk assessment report. S4. Early Warning Triggering and Personalized Intervention: a. Push risk warnings and detailed explanations to family members' mobile applications; b. At the same time, based on the risk profile, intervention modules are called and combined from the strategy library to generate personalized solutions; for users with obvious forgetfulness trends, timed reminders are strengthened through voice assistant 108; for users with reduced activity, family members are advised to encourage specific activities through the APP; for users with declining communication skills, the difficulty and frequency of daily structured dialogues of voice assistant 108 are increased, and educational game tasks are pushed on the APP. S5. Intervention Feedback and Model Iteration: The system continuously monitors changes in behavioral data after the implementation of intervention measures, evaluates the effectiveness of the intervention, and dynamically adjusts the risk model and intervention strategies to form a management closed loop. Example

[0024] 1. System Deployment: Install this system in Grandpa Zhang's home. Install door magnets on the front door and refrigerator. Connect the electric kettle to the smart socket 102. Install a water immersion sensor in the kitchen. Place a sleep monitoring belt 104 under the bed in the bedroom. Install infrared sensors in the living room and bathroom. Install a flushing sensor on the toilet in the bathroom. Place a smart voice assistant 108 in the living room. Connect all devices to the network. Install a monitoring APP on the children's mobile phones. 2. Baseline learning: During the initial system run, Grandpa Zhang's normal behavior patterns were recorded. The baseline showed an average of 0.5 forgetful events per week, relatively fixed daily activity paths, and an average response time of 2.1 seconds when conversing with voice assistant 108. 3. Trend Analysis and Risk Warning: Cloud platform analysis of data trends over the past three months revealed the following: Forgetfulness frequency: gradually increased from 2 times per month to 8 times per month, showing a significant linear growth trend; Activity pattern: The proportion of time spent in the living room and bedroom increases daily, while activity in the kitchen and bathroom decreases, and the complexity of activity paths decreases by 15%; Voice interaction: The average response time has increased to 3.5 seconds, and the avoidance rate for complex questions has increased; The comprehensive assessment module determined that multiple key digital biomarkers showed synergistic deterioration, and the overall risk score reached "intermediate". 4. Early warning and intervention: Warning: The system will immediately push a warning report to the child's APP, which includes the above trend charts and risk interpretation. It is recommended to pay attention and consider a professional assessment. Intervention: The system simultaneously launches a personalized intervention plan: Enhanced reminders: Voice Assistant 108 proactively reminds Grandpa Zhang to "check doors, windows, water, and electricity" before he leaves home and before he goes to bed. Cognitive training: The app pushes suggestions to the children, encouraging them to have a 10-minute "memory recollection" conversation with their father every day through the voice assistant 108. At the same time, the voice assistant 108 actively invites Grandpa Zhang to play a word classification game once a day. Lifestyle suggestions: The app recommends that children take their fathers out more often on weekends and consider setting up more prominent reminder labels at home; 5. Closed-loop tracking: After the intervention is implemented, the system continues to monitor. Data shows that after the reminders are strengthened, the frequency of forgetful events tends to stabilize. After two months of consistent cognitive training, the response delay of voice interaction has slightly improved. The system feeds the effect back to the model, and the entire "assessment-intervention" closed loop continues to operate.

[0025] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. An Alzheimer's disease risk assessment and intervention system, characterized in that, It includes a data acquisition layer, a data transmission and network layer, a cloud-based intelligent analysis and management platform, a personalized intervention strategy library and generation module, a user interaction and application layer, and an interface for elderly users. The data acquisition layer includes a home-based seamless sensing layer; The home-based seamless sensing layer consists of multiple IoT sensor nodes deployed in the user's home, used to automatically collect multi-dimensional behavioral data such as forgetfulness, daily rhythms and activity patterns, and proactive interactions and cognitive stimuli. Data transmission and network layer: Used to transmit the collected data from IoT sensor nodes to a cloud-based intelligent analysis and management platform via home Wi-Fi and / or low-power wide area network; Cloud-based intelligent analysis and management platform: including a multimodal data fusion and storage module, a behavioral digital biomarker extraction engine, and a long-term trend analysis and risk assessment module; ² Multimodal data fusion and storage module: Receives and cleans heterogeneous data from different sensors, aligns and stores it according to time series, and builds a "behavioral digital twin" database for each user; Behavioral digital biomarker extraction engine: Extracts quantitative features related to cognitive function from raw data; Long-term trend analysis and risk assessment module: This module performs long-term time-series analysis on the extracted digital biomarkers, establishes an individual baseline for each feature, and uses statistical models and machine learning algorithms to identify its slow but continuous trend of deterioration. By integrating the deterioration trends from multiple dimensions, it generates a comprehensive cognitive function decline risk score and low, medium, and high risk levels. Personalized intervention strategy library and generation module: used to store various digital intervention solutions, and automatically generate and distribute personalized intervention plans based on the user's risk level and specific weaknesses; User interaction and application layer: Use a client APP for family members and caregivers to view risk assessment reports, trend charts, receive early warning notifications, and monitor the implementation of intervention plans; The elderly user interface: They receive and execute intervention tasks through the voice assistant (108), play daily cognitive training games, receive life reminders, and participate in social dialogues.

2. The Alzheimer's disease risk assessment and intervention system as described in claim 1, characterized in that, The extraction of quantitative features related to cognitive function from raw data includes: The weekly average incidence of forgetful events, the complexity of daytime activity paths, the weekly coefficient of variation of nighttime waking-up-to-the-toilet frequency, and the average response delay when conversing with a voice assistant (108).

3. The Alzheimer's disease risk assessment and intervention system as described in claim 1, characterized in that, The IoT sensor nodes include a forgetfulness behavior monitoring node, a daily rhythm and activity pattern monitoring node, an active interaction and cognitive stimulation node, a water immersion sensor, a human infrared sensor (101), a smart socket (102), a door magnetic sensor (103), a sleep monitoring strip (104), a toilet flushing sensor (105), a home wireless LAN (106), a voice assistant (108), a smart pillbox (109), a thermometer and hygrometer (110), a smart night light (111), and an SOS press alarm (112). Forgetfulness monitoring node: By utilizing door magnetic sensors (103) on the front door and refrigerator door, smart sockets (102), water immersion sensors and smart pillboxes (109), the frequency of forgetfulness events related to short-term memory and executive function is quantified and recorded; Circadian rhythm and activity pattern monitoring nodes: Using human infrared sensors (101) deployed in each room, sleep monitoring belts (104) in the bedroom, and toilet flushing sensors (105) in the bathroom, the activity trajectory, spatial transfer patterns, toilet frequency and duration, and daytime bed rest duration of users are collected to analyze the regularity of circadian rhythm and subtle changes in activity capacity. Active interaction and cognitive stimulation nodes: Use a voice assistant (108) to initiate structured dialogue interactions periodically and record the response time, accuracy and fluency data of the interaction process.

4. The Alzheimer's disease risk assessment and intervention system as described in claim 1, characterized in that, The data transmission and network layer also includes an encryption unit; Encryption unit: Used to encrypt the collected data before transmitting it to the cloud-based intelligent analysis and management platform.

5. The Alzheimer's disease risk assessment and intervention system as described in claim 1, characterized in that, The data transmission and network layer also includes an IoT gateway; The IoT gateway connects to the home's wireless LAN and low-power WAN, enabling devices within the home's seamless sensing layer to access the cloud-based intelligent analysis and management platform.

6. The Alzheimer's disease risk assessment and intervention system as described in claim 1, characterized in that, It also includes an infrared sensor for the entrance door (107). An infrared sensor (107) is installed at the entrance door to monitor when people leave and return home.

7. The Alzheimer's disease risk assessment and intervention system as described in claim 1, characterized in that, It also includes a thermometer and hygrometer (110); The temperature and humidity meter (110) is installed in various spaces in the user's home to detect indoor temperature and humidity.

8. The Alzheimer's disease risk assessment and intervention system as described in claim 1, characterized in that, It also includes a smart night light (111); The smart night light (111) is installed in the hallway of various spaces in the home to automatically sense nighttime human activity and provide lighting.

9. The Alzheimer's disease risk assessment and intervention system as described in claim 1, characterized in that, It also includes an SOS press alarm (112); The SOS press alarm (112) is installed in various spaces in the home for residents to press to trigger the alarm.

10. A method for Alzheimer's disease risk assessment and intervention, characterized in that, Includes the following steps: S1. Data collection is carried out in a normalized and seamless manner: After the system is deployed, each sensor starts to work continuously, the voice assistant (108) initiates interaction as planned, and all data is automatically uploaded to the cloud. S2. Baseline Establishment and Feature Learning: In the initial stage, the system learns the user's behavioral patterns under normal conditions and establishes the individual baseline level of each digital biomarker. S3. Long-term trend analysis and risk assessment: The platform periodically analyzes the changing trends of each feature relative to the baseline. When the system detects that multiple key features show a statistically significant synergistic deterioration trend, it determines that the risk of cognitive decline has increased and generates a risk assessment report. S4. Early Warning Triggering and Personalized Intervention: a. Push risk warnings and detailed explanations to family members' mobile applications; b. At the same time, based on the risk profile, the intervention modules are called and combined from the strategy library to generate personalized solutions; for users with obvious forgetfulness, the voice assistant (108) is used to strengthen timed reminders; for users with reduced activity, the APP is used to suggest that family members encourage specific activities; for users with declining communication ability, the difficulty and frequency of daily structured dialogues of the voice assistant (108) are increased, and puzzle game tasks are pushed on the APP. S5. Intervention Feedback and Model Iteration: The system continuously monitors changes in behavioral data after the implementation of intervention measures, evaluates the effectiveness of the intervention, and dynamically adjusts the risk model and intervention strategies to form a management closed loop.

Images