LU602011B1 - An intelligent, interactive device and system based on a digitized home care service. - Google Patents

Abstract

The present invention discloses an intelligent interactive device and system based on digital home healthcare services, comprising a physiological parameter acquisition module, an embedded controller, a wireless communication module, a multimodal interactive interface, and an emergency response module. The physiological parameter acquisition module is connected to the embedded controller via an I2C bus. The health risk analysis algorithm is based on an LSTM neural network to achieve time-series modeling of somatic symptom data. The present invention discloses an intelligent interactive device and system based on digital home healthcare services. With regard to health monitoring, the multi-source sensor network can accurately collect user physiological data and environmental parameters in real time. Combined with advanced algorithms for in-depth analysis, this data can accurately and promptly assess health risks and provide a solid foundation for disease prevention and intervention. The multimodal interaction interface integrates multiple interaction modes, fully considers the operating habits and needs of elderly people, provides a natural and convenient interaction experience, and reduces difficulties in use.

Description

An intelligent interactive device and system based on a digitized home care service.

Technical area

The present invention relates to the technical field of domestic

Nursing care and in particular an intelligent interactive device and system on the

Basis for digital home care services.

Technology in the background

As the world's population ages, aging in place is becoming an important issue.

It has become a form of care. However, the field of aging currently faces many challenges. On the one hand, with the traditional approach to aging, it is difficult to comprehensively monitor the physiological health of older people in real time, and it is impossible to identify potential health risks in a timely manner. This leads to a lack of timely intervention and treatment of illnesses, which significantly impacts the quality of life, health, and safety of older people. On the other hand, the existing service system for older people lacks intelligence and

Personalization makes it difficult to offer accurate and effective services according to the actual needs and health condition of elderly people.

Regarding the interaction experience, most devices only have a single one.

Interaction mode that does not meet the diverse needs of older people, especially for older people with limited mobility or poor vision and

Hearing impairment makes use more difficult. Furthermore, the existing

The alarm mechanism is slow in an emergency and has a limited range, so it is unable to notify the responsible personnel quickly and effectively, and the best

The opportunity for rescue was missed.

Content of the invention

In view of the aforementioned shortcomings of the state of the art, the present

Invention of an intelligent interaction device and system based on a digital home care service, which aims to solve the problem that the existing

The care service system lacks intelligence and personalization, and it is difficult to provide accurate and effective services that correspond to actual needs and the

to assess the health status of older people. Regarding the interaction experience, most devices have a single interaction mode that addresses the diverse needs of older adults.

does not meet the needs of older people, and especially not for older people with

It is difficult to operate for people with mobility problems or poor vision and hearing. Furthermore, the existing alarm mechanism is slow in an emergency.

Response speed and a limited notification range mean it is unable to quickly and effectively notify the responsible personnel and the best

Missing the opportunity for rescue, among other problems.

To achieve the aforementioned purpose of the invention, the present invention provides the following technical solution:

An intelligent interaction device based on a digital home

Elderly care service, which includes a mobile robot body, and further includes:

Module for recording physiological parameters: including biosensor arrays;

Embedded controller: including ARM architecture processor;

Wireless communication module: including Wi-Fi/5G dual-mode transmission for the

Data interaction with the cloud server; LU602011

Multimodal interaction interface: including speech recognition unit, touch-

Display and LED warning light strip;

Emergency module: with a local sound and light alarm and a GSM emergency call system;

The module for acquiring physiological parameters is connected to the embedded controller via an I2C bus, and the algorithm for analyzing health risk is based on a neural LSTM network to perform time series modeling of somatic

To reach data.

Furthermore, the multimodal interaction interface includes:

A 3D ToF camera for capturing body movements and microexpressions of the

User;

A directional loudspeaker to enable directed voice announcements;

A haptic feedback unit with a piezoelectric ceramic vibration module;

Each interaction channel is controlled by an attention mechanism-

Fusion algorithm prioritized.

An intelligent interaction system based on a digital home

Elderly care services and includes the following:

A distributed monitoring terminal used in a domestic environment with smart, wearable devices and environmental sensors;

An edge computing node equipped with an FPGA acceleration chip for the

is equipped with real-time processing of sensor data streams;

Algorithms for analyzing health risks, implemented in an embedded controller for

Algorithms are configured to analyze health risks and interactive decision logic;

Emergency response strategy, configured in the emergency response module;

Cloud service platform, built with a user profile database and a

Knowledge graphs of senior care service strategies;

Mobile interaction terminal that supports two-way access to the app and WeChat applet;

The system conducts communication between devices via the MQTT protocol and uses OAuth 2.0 for access control.

Furthermore, the algorithm for analyzing health risks includes:

A feature extraction layer that uses a neural convolutional network to process heterogeneous data from multiple sources;

A risk prediction layer that uses an XGBoost classifier to classify the

used for health condition;

A sophisticated strategy generation layer that creates a service plan based on...

Reinforcement Learning dynamically adapts;

The aforementioned algorithm for analyzing health risks matches a physiological

Baseline of different users through migration learning.

Furthermore, the emergency strategy includes:

A tiered response strategy that categorizes the risk level into four levels.

Release mechanism subdivided;

Multi-channel alarm that synchronously alerts a local device, a relative terminal, and a

community center triggers;

Ambient control that automatically adjusts the interior lighting as well as door and/602011

activates window locks.

Furthermore, the cloud service platform includes:

A data fusion engine that performs feature-level fusion of physiological data,

collects environmental data and user behavior data;

A service recommendation system that uses a collaborative filtering algorithm to create a personalized pension plan;

A security protection system that uses homomorphic encryption technology to process sensitive health data;

The system is connected to third-party medical resources via an API interface.

Furthermore, the security protection system includes:

A hardware trust base that integrates a physical, non-clonable functional unit (PUF);

A secure boot chain that uses a digital signature to verify the integrity of the firmware;

A data desensitization module that implements differentiated data protection processing for the collected data.

Furthermore, the service recommendation system includes:

Construction of a user feature vector space with dimensions that include physiological

Indicators include lifestyle habits and preferences;

Constructing a diagram for the association of service elements, where the

Node weights reflect synergistic effects between service elements;

Implementing dynamic detection of interest drift and updating the

Recommendation strategy through a sliding window mechanism.

Furthermore, the edge computing node includes:

A framework for processing streaming data that uses the Apache Kafka

Real-time data pipeline supported;

A model reduction module that uses knowledge distillation techniques to develop AI-

To compress models;

A unit for optimizing energy consumption that dynamically adjusts computing resources according to the workload.

Furthermore, the system's working procedure includes the following steps:

Collecting physiological data of the user and environmental parameters in

Real-time through a multi-source sensor network;

Implementation of data preprocessing at the edge nodes, including the

Outlier detection and data normalization;

Upload to the cloud for in-depth analysis to generate a report evaluating the

to create health risk profiles;

Matching service resources based on user profiles and providing

Interaction instructions via a multimodal interface;

Optimization of the model based on feedback data to create a closed-loop system

To form a cycle of services.

The advantageous effects of the present invention are:

An intelligent interaction device and system based on the digital home care service of the present invention can accurately and in real time capture physiological data and environmental parameters of the user with regard to health monitoring/©02011 and, in combination with advanced algorithms for in-depth analysis, it can

Assess health risks accurately and in a timely manner, and establish a solid foundation for the

It offers disease prevention and intervention. The multimodal interaction interface integrates multiple interaction modes, fully considers the operating habits and needs of older adults, provides a natural and convenient interaction experience, and reduces usability difficulties. The emergency module develops a hierarchical

A response strategy and a multi-channel alarm mechanism ensure that the responsible personnel are quickly notified in an emergency and that smart home devices are networked and controlled to create favorable conditions for rescue and protect the lives of elderly people. The service recommendation system accurately matches service resources based on user profiles, dynamically optimizes the recommendation strategy, offers personalized retirement solutions, and improves service satisfaction.

At the same time, edge computing and collaborative data processing in the cloud improve the system's responsiveness and data processing capabilities, while also ensuring data security. Overall, the device and the system comprehensively improve the

Quality and efficiency of home care services and create a safer, more comfortable and convenient home care environment for elderly people.

Description of the enclosed drawings

Figure 1 shows a schematic diagram of the three-dimensional structure of the intelligent

Interaction device based on the digitized home-based elderly care service of the present

invention based;

Figure 2 is a schematic diagram of the intelligent interaction device on the

Basis of the digitized home care services of the present invention

Figure 3 is a schematic diagram of the core architecture of the intelligent

Interaction device of the present invention;

Figure 4 is a schematic diagram of the multimodal interaction interface of the present invention;

Figure 5 shows a schematic diagram of a cloud server of the present invention;

Figure 6 shows a schematic diagram of the edge computing layer of the present

Invention;

Figure 7 shows a schematic diagram of the data acquisition layer of the present

Invention;

Figure 8 shows a schematic diagram of the health risk analysis and emergency response of the present invention;

Figure 9 shows a diagram of the service recommendation system and the workflow of the present invention.

A comparison table of the attached markings: 1, a mobile robot body; 2, a physiological parameter acquisition module; 3, an embedded controller; 4, a wireless communication module; 5, a multimodal

Interaction interface; 6, an emergency response module; 7, a three-dimensional ToF camera; 8, a directional speaker; and 9, a haptic feedback unit.

Detailed description

Specific embodiments of the present invention are described below in

The connection is described in more detail with the attached drawings. The same parts and

Components are identified by the same symbols in the accompanying drawings. LU602011

It should be noted that the terms "forward," "backward," "left," "right," "up," and "off" used in the following description refer to the orientations in the accompanying drawings. The terms "inside" and "outside" refer to the five directions toward and away from the geometric center of a given component.

To make the content of the present invention easier and clearer to understand, the technical solutions in the embodiments of the present invention are described in the

The following, in conjunction with the accompanying drawings, clearly and completely describe the embodiments of the present invention.

As shown in images 1 to 9, an intelligent interaction device on the

Basis of a digitized home-based elderly care service:

Mobile robot body 1: The mobile robot body 1 serves as a physical

Support platform for the entire intelligent interaction device and was developed under

Utilizing a sophisticated, custom-designed mobile robot architecture with wheels. Its mechanical structure was carefully designed for flexibility.

To ensure steering capability and stable mobile performance, and can easily adapt to the complex terrain and spatial arrangement in the domestic environment.

Equipped with large, non-slip wheels and advanced sensors for

By avoiding obstacles, it ensures safe and smooth movement even in tight spaces.

Hallways or densely furnished rooms.

Physiological parameter acquisition module 2: The biosensor array integrates a variety of high-precision biosensors, such as a highly sensitive heart rate sensor, an accurate

A blood oxygen sensor and a reliable blood pressure sensor. These sensors are arranged according to strict ergonomic principles to ensure good contact with the user's skin and are capable of accurately and consistently measuring the user's physiological parameters.

to capture real-time data. Thanks to a stable connection to the integrated Controller3 via I2C-

Data transmission via bus is efficient and accurate, creating a reliable data basis for subsequent health analyses.

Embedded Controller 3: A high-performance processor with ARM architecture is selected, e.g., a high-performance, low-power STM32H7 series chip.

Power consumption. The processor is capable of smoothly executing the health risk analysis algorithm based on an LSTM neural network to thoroughly process and analyze the collected physiological parameter data. During the implementation of the algorithm, the raw data is first preprocessed to

Storgerdusche to remove, and then fed into the trained LSTM model, which is able to accurately capture the time series characteristics of the physical signs data and generate accurate results for assessing health risk.

Wireless communication module 4: A communication module is used that...

Dual-mode transmission Wi-Fi/5G supported, e.g. the Qualcomm Snapdragon X65 modem.

The module features fast and stable data transmission and is able to quickly and accurately process the data from the embedded controller 3 according to the MQTT standard.

The protocol is uploaded to the cloud server. The OAuth 2.0 access control mechanism ensures the security of the data transmission and prevents data loss and unauthorized access.

Access. When receiving instructions from the cloud, the wireless

Communication module 4 analyzes the MQTT message in a timely manner and transmits the instructions accurately to the embedded control unit 3, thus enabling efficient interaction between deh}602011

The device and the cloud are reached.

Multimodal Interaction Interface 5: Speech Recognition Unit: It integrates the latest

Generation of the speech recognition SDK from Tech Data Xunfei and features a highly precise

Speech recognition function and a powerful noise reduction function.

The speech recognition unit is able to quickly and accurately convert voice commands into text information, which is then forwarded to the integrated Controller3 for processing.

Touch display: the selection of high-resolution, high-contrast capacitive touch displays

The display is clear and offers sensitive operation. Users can operate it intuitively.

View information such as health risk assessment reports, service plans, etc., and personalize operating settings via the touch display.

LED warning strip: The LED warning strip consists of bright, multicolored LEDs.

Luminous beads that can display different colors and flashing frequencies depending on the health risk. For example, if the health risk is low, the green light flashes slowly; if the risk increases, the light color changes to orange or red, and the

The flashing frequency increases to visually warn the user and surrounding personnel. 3D ToF Camera 7: The high-precision 3D ToF camera is capable of capturing the user's body movements and micro-expressions in real time. In-depth analysis of the captured image data achieves precise gesture and facial expression recognition, providing the user with a more natural and comfortable interaction experience.

Directional Speaker 8: The Directional Speaker 8 utilizes advanced acoustic technology.

Directional technology to accurately transmit speech to the user's location and

To avoid interference with the environment. Regardless of where the user is in the

If he is in the room, he can clearly hear the device's voice prompts and interactive instructions.

Haptic Feedback Unit 9: The piezoelectric ceramic vibration module in the haptic feedback unit © has a fast response time and adjustable intensity.

Vibration intensity. When the user operates the device or the device provides feedback, the piezoelectric ceramic vibration module can generate appropriate vibration feedback according to the various interaction scenarios, thus...

Improve user perception and interaction experience.

Prioritization of interaction channels: Each interaction channel is assigned a priority by the

The fusion algorithm of the attention mechanism was assigned a priority.

The algorithm can analyze the user's current state and interaction needs in real time and dynamically adjust the priority of interaction channels such as speech recognition, touch control, and body motion detection. For example, if the user...

When a voice interaction is performed, the priority of the speech recognition channel is increased; if the

When the user touches the display to operate it, the priority of the touch channel is increased accordingly to ensure the efficiency and accuracy of the interaction.

Emergency Call Module 6: The local sound and light alarm features a loud buzzer and a bright LED light that immediately emits a piercing alarm tone and flashes a bright light when an emergency is detected, quickly attracting the attention of nearby personnel. The GSM emergency call circuit has a built-in high-performance GSM module that connects quickly and reliably to the GSM network and sends an emergency message containing the user's location information and detailed information.

Description of the emergency to pre-configured, associated terminals and municipal

Service centers can be sent. The hierarchical response strategy divides the risk level into vik}/602011

Stages of triggering mechanisms, with each response stage corresponding to a clear operational process and division of responsibilities to ensure that emergency situations can be handled quickly and properly.

The intelligent interactive system includes:

Distributed monitoring terminals: A variety of smart portable devices, such as...

B. smart bracelets and smartwatches, and environmental sensors, such as...

Temperature sensors, humidity sensors, and air quality sensors are usefully employed in the home environment. These devices provide wireless data.

Low-power communication protocols, such as ZigBee or Bluetooth, establish a stable connection with the edge computing node to transmit the physiological data of the

To capture user and environmental parameters in real time and to precisely transfer the data to the edge computing node for processing.

Edge computing node: Equipped with a high-performance FPGA-

Acceleration chip that uses its parallel computing capacity to process sensor data streams into

To process data in real time. The framework for processing streaming data is based on Apache.

Kafka, which can efficiently receive, store, and distribute sensor data to ensure reliable real-time data. The model reduction module utilizes an advanced

Knowledge distillation technology to compress complex AI models into lightweight models, significantly improving the processing efficiency of edge computing nodes while ensuring accuracy. The unit for optimizing the

Energy consumption monitors the workload in real time through intelligent algorithms and dynamically adjusts the processor's operating frequency and voltage to reduce energy consumption and extend the device's lifespan.

Algorithm for health risk analysis: The feature extraction layer uses neural convolutional networks to perform deep feature extraction from heterogeneous data.

to process data from multiple sources; the neural convolutional networks are designed for a large number of users.

The software has been trained on a large dataset and can automatically detect important features in the data.

The risk prediction layer uses the XGBoost classifier, which has efficient and accurate classification capabilities and can precisely categorize the user's health status into different levels. The strategy layer is based on algorithms of

Reinforcement learning, which adjusts the service plan based on health status and...

Continuously optimize user feedback data to ensure personalization and effectiveness of the service. Migration learning technology enables this.

Algorithm for analyzing health risks, quickly adapting to physiological

To adapt the initial situation of the different users, thereby improving the generalizability of the algorithm.

Emergency strategy: A tiered response strategy clearly defines the measures for the different risk levels to ensure rapid action in an emergency.

Multi-channel alarm mechanism with 16 simultaneous local devices via wireless communication,

Devices of family members and municipal service centers to provide comprehensive

To receive alarm notifications. The environment linking function can automatically control smart home devices, such as turning on indoor lighting, etc.

Locking doors and windows, etc., to make the work of the rescue workers easier and the

To ensure user safety.

Cloud service platform: The data fusion engine uses advanced/602011

Data fusion algorithms to combine physiological data, environmental data, and data on

To fuse user behavior at the feature level in order to extract valuable information and provide comprehensive data support for service recommendations and health analyses.

The service recommendation system uses collaborative filter algorithms that...

The user's feature vector space and the mapping of service elements are combined to create personalized retirement plans for the user. The security protection system ensures the security of sensitive health data during transmission and

Storage through homomorphic encryption technology, hardware trust base, secure

Boot chain and data desensitization module, etc. The API interface can be conveniently used with

Third-party health resources are integrated to provide users with more comprehensive information.

To offer healthcare service options.

The service recommendation system includes:

Creation of a user attribute vector space containing multiple dimensions such as physiological indicators, lifestyle habits, and preference settings to...

To comprehensively represent user characteristics. Construction of associations between

Services, determination of node weights by analyzing the synergy effects between

Services and provision of a scientific basis for

Service recommendations. Implementation of the dynamic mechanism for detecting deviations in interests, monitoring user behavior data, and feedback.

Real-time information through the sliding window, timely adjustment of the

Recommendation strategy and ensuring that the recommended pension system always meets the actual needs of the user.

This includes edge computing nodes:

The framework for streaming data processing builds an efficient data pipeline based on Apache Kafka to enable fast transmission and processing of

to access sensor data. The Lightweighting Module model migrates knowledge from large

Teacher models to small student models through knowledge distillation technology, in order to the

To reduce model complexity and increase processing speed, the power consumption optimization unit dynamically adjusts computing resources to the workload by, for example, switching off unused hardware modules, reducing the processor frequency, etc., to minimize energy consumption.

Specific implementation of a working method of an intelligent interaction system based on digital home care services:

A multi-source sensor network collects physiological data from the user and

Real-time environmental parameters

The intelligent wearable devices and environmental sensors in the distributed

Monitoring terminals collect data in real time according to a preset sampling frequency and accuracy. For example, the smart bracelet collects data every 30 seconds.

Heart rate and the temperature sensor provide room temperature data every 10 minutes. The recorded data

Data is quickly transferred to the edge computing node via wireless communication protocols to ensure the data is up-to-date and accurate.

Data preprocessing takes place at the edge node, where the edge computing node handles the

receives sensor data and immediately performs outlier detection, using statistical

Machine learning methods and algorithms are used to identify and remove obviously erroneous data. The data is then standardized to convert data with different scales into a uniform range and prepare for subsequent in-depth analysis. The preprocessing process is efficient and accurate, ensuring that the data quality meets the requirements of the algorithms for analyzing...

This corresponds to health risks.

The pre-processed data is uploaded to the cloud for in-depth analysis and securely and reliably transferred to the cloud server via the MQTT protocol. On the cloud-

A high-performance algorithm for analyzing health risks runs on the server, which...

thoroughly evaluates and analyzes data. The algorithm combines historical and

Real-time user data is used to generate a detailed health risk assessment report that includes health status, potential health risks, and recommended actions.

Includes service solutions.

Matching service resources based on the user profile offers a

Interaction guidance via a multimodal interface, and the cloud service platform matches the user with suitable senior care service resources based on

Information is stored in the user portrait database, including health status, the

The service plan is then sent to the smart interactive device via the wireless communication module, taking into account the user's lifestyle habits and preferences. The smart interactive device provides the user with five detailed interactive guides through a multimodal interface, such as clear voice instructions and intuitive touchscreen controls.

Displays and obvious LED warnings to ensure that the user is able to easily understand and execute the service plan.

To optimize the model based on the feedback data, in order to achieve a closed-loop system

To form a service loop, the system continuously collects feedback data from the user, such as the user's satisfaction rating with the service plan and changes in physiological parameters. By analyzing this feedback data, the system is able to...

Problems in the algorithm for analyzing health risks and in the

To recognize interaction decision logic and optimize it in a timely manner. For example, the recommendation strategy is adjusted according to user feedback.

The service plan was adjusted and the risk prediction model optimized according to changes in physiological parameters. This mechanism of continuous...

Optimization creates a closed loop of services to continuously improve the system's performance and the quality of services, and to offer users a better quality of home care.

The intelligent interactive device and system, based on the digital home care service of the present invention, can operate stably and efficiently, provide a comprehensive and personalized care service for elderly people, and effectively improve the safety and comfort of home care.

The foregoing is only one preferred embodiment of the patent in the present case.

invention and is not intended to limit the patent of the present invention, and all

Modifications, equivalent replacements, improvements, etc., made within the scope of the spirit and principles of the patent for the present invention, fall within the scope of

Scope of protection of the patent for the present invention.

Claims (10)

Claims LU602011 1. An intelligent interaction device based on a digital home-based elderly care service, comprising a mobile robot body, characterized in that it further comprises: A module for acquiring physiological parameters: with a biosensor array; An embedded controller: including a processor of ARM architecture; A wireless communication module: with Wi-Fi/5G dual-mode transmission for data interaction with a cloud server; A multimodal interaction interface: comprising a speech recognition unit, a touch display, and an LED warning light strip; An emergency call module: with a local audible and visual alarm and a GSM emergency call circuit; The module for acquiring physiological parameters is connected to the embedded controller via an I2C bus, and the health risk analysis algorithm is based on an LSTM neural network to achieve time-series modeling of the somatic data. 2. An intelligent interaction device based on a digital home care service according to claim 1, characterized in that the multimodal interaction interface comprises: A 3D ToF camera for capturing body movements and micro-expressions of the user; A directional speaker for directed voice prompts; A haptic feedback unit with a piezoelectric ceramic vibration module; Each interaction channel is assigned a priority by an attention mechanism fusion algorithm. 3. An intelligent interaction system based on a digital home care service, comprising an intelligent interaction device based on a digital home care service according to one of claims 1 to 2, characterized in that it comprises: a distributed monitoring terminal deployed in a home environment with smart wearable devices and environmental sensors; an edge computing node equipped with an FPGA accelerator chip for real-time processing of sensor data streams; a health risk analysis algorithm configured in an embedded controller for health risk analysis algorithms and interactive decision logic; an emergency response strategy configured in the emergency response module; a cloud service platform built with a user profile database and a knowledge graph of senior care service strategies; a mobile interaction terminal supporting two-way access via app and WeChat applet; the described system enables communication between devices via the MOTT protocol and uses OAuth 2.0 for access control. 4. An intelligent interaction system based on a digital home-based elderly care service according to claim 3, characterized in that the algorithm for Health risk analysis includes the following: LU602011 Fine feature extraction layer, which uses a neural convolutional network to process heterogeneous data from multiple sources; a risk prediction layer, which uses an XGBoost classifier to classify health status; a strategy generation layer, which dynamically adapts a duty roster based on reinforcement learning; and a health risk analysis algorithm, which adapts a physiological baseline of different users through migration learning. 5. An intelligent interaction system based on a digital home-based elderly care service according to claim 3, characterized in that: the emergency response strategy comprises: A tiered response strategy that divides the risk level into four stages of triggering mechanisms; Multi-channel alarm that synchronously triggers local devices, relative terminals and community service centers; Environmental control that automatically switches on the interior lighting and the door and window locking devices. 6. An intelligent interaction system based on a digital home-based elderly care service according to claim 3, characterized in that the cloud service platform comprises: a data fusion engine that implements feature-level fusion of physiological data, environmental data, and user behavior data; a service recommendation system that applies a collaborative filtering algorithm to generate a personalized elderly care program; a security protection system that employs homomorphic encryption technology to process sensitive health data; the system can be connected to third-party medical resources via an API interface. 7. An intelligent interaction system based on a digital home-based elderly care service according to claim 3, characterized in that the security protection system comprises: A hardware trust root integrating a physical non-clonable functional unit (PUF); A secure boot chain using a digital signature to verify firmware integrity; A data desensitization module implementing differentiated data protection processing for the collected data. 8. An intelligent interaction system based on a digital home-based elderly care service according to claim 6, characterized in that the service recommendation system comprises: constructing a user attribute vector space whose dimensions include physiological indicators, lifestyle habits, and preference settings; constructing a graph for the assignment of service elements, wherein the node weights reflect synergistic effects between service elements; Implementation of a dynamic detection of interest drift and updating of the recommendation strategy through a sliding window mechanism 9. An intelligent interaction system based on a digital home-based elderly care service according to claim 3, characterized in that: the edge computing node comprises: a streaming data processing framework that supports the Apache Kafka real-time data pipeline; a model reduction module that applies knowledge distillation technology to compress AI models; and an energy consumption optimization unit that dynamically adjusts computing resources according to the workload. 10. An intelligent interaction system based on a digital home-based elderly care service according to claim 3, characterized in that: the system's working method comprises the following steps: collecting user physiological data and environmental parameters in real time via a multi-sensor network; performing data preprocessing at an edge node, including outlier detection and data standardization; uploading to the cloud for in-depth analysis to generate a health risk assessment report; matching service resources based on user profiles and providing interaction instructions via a multimodal interface; optimizing the model based on feedback data to form a closed-loop service cycle.