CN-121667652-B - Intelligent pillow sleep monitoring and personalized improving device based on digital twinning

CN121667652BCN 121667652 BCN121667652 BCN 121667652BCN-121667652-B

Abstract

The invention provides an intelligent pillow sleep monitoring and individuation improving device based on digital twinning, which relates to the technical field of medical health information, and aims to quickly judge reflecting conditions and make urgent intervention through hard-wired logic, meanwhile, a individuation intervention strategy is generated through establishing a dynamic digital twinning body to be executed, and after each intervention, an intervention commonality score is calculated by collecting user implicit physiological feedback data, a differential parameter update is executed based on the intervention commonality score, and when new characteristic data appear, a structure self-evolution is carried out, an internal representation component is removed and created, and through a dual mechanism of the parameter update and the structure self-evolution, the digital twinning body can always accurately describe a sleep image of dynamic change of a user, so that continuous, effective and more-used individuation service is provided, the technical problem of adaptive attenuation of a static model is fundamentally solved, the adaptability to the user is improved, and a good intervention effect is ensured.

Inventors

SUN YU
Request for anonymity
WANG SUJIE
Qian Linze
XU PANPAN

Assignees

浙江大学温州研究院

Dates

Publication Date: 20260512
Application Date: 20260206

Claims (9)

1. The intelligent pillow sleep monitoring and individuation improving device based on digital twinning comprises a processor and a memory, and is characterized in that the executing method of the device comprises the following steps: Acquiring user physiological and behavior data which are acquired in the intelligent pillow through a sensor array in a non-sensing way, judging whether to trigger conditional reflex based on hard-wired logic, if so, executing emergency intervention, otherwise, constructing a dynamic digital twin body based on the user physiological and behavior data; Setting privacy invasiveness indexes on the basis of the expected intervention effects of N candidate intervention strategies based on dynamic digital twin simulation, performing multi-objective evaluation on the basis of the expected intervention effects, wherein the multi-objective evaluation comprises validity evaluation, continuity evaluation and user preference evaluation, calculating comprehensive utility scores of each candidate intervention strategy, generating evaluation results based on the comprehensive utility scores, screening the candidate intervention strategies based on the comprehensive utility scores, selecting candidate intervention strategies with the comprehensive utility scores being greater than or equal to a preset utility threshold value to form a candidate strategy set, selecting candidate intervention strategies with the lowest privacy invasiveness indexes in the candidate strategy set as basic strategies, judging the basic strategies based on the validity evaluation, directly taking the basic strategies as personalized intervention strategies if the validity evaluation of the basic strategies is greater than or equal to the validity threshold value, selecting candidate intervention strategies with the highest validity evaluation from the candidate strategy set as supplementary strategies if the validity evaluation of the basic strategies is smaller than the validity threshold value, and fusing the basic strategies and the supplementary strategies according to time sequence logic to generate distributed fusion instructions as personalized intervention strategies; converting the personalized intervention strategy into a control instruction and transmitting the control instruction to the intelligent pillow, wherein the intelligent pillow jointly executes the control instruction by combining with external environment equipment; And acquiring implicit physiological feedback data of the user after the emergency intervention and control instruction execution, calculating an intervention consensus degree score, and carrying out parameter updating and structure self-evolution on the dynamic digital twin body according to the intervention consensus degree score.
2. The intelligent pillow sleep monitoring and individualizing apparatus based on digital twinning of claim 1, wherein the sensor array comprises micro-vibration sensors, pressure sensors, microphones and reflective photo-sensors, and the user physiological and behavioral data comprises heart rate data, respiration rate data, sleep posture data, body movement data, environmental data, snore intensity and blood oxygen saturation data.
3. The digital twinning-based intelligent pillow sleep monitoring and personalization improvement device of claim 2, wherein the hardwired logic comprises a first judgment condition and a second judgment condition; acquiring real-time respiratory rate data, and triggering a first judgment condition to perform emergency intervention if the respiratory rate signal is identified to be interrupted and the interruption time exceeds a first preset threshold value; And acquiring the real-time snore intensity, and triggering a second judgment condition to perform emergency intervention if the snore intensity is identified to exceed a second preset threshold and the blood oxygen saturation is continuously reduced.
4. The intelligent pillow sleep monitoring and personalization improvement device based on digital twinning of claim 1, wherein the simulating the expected intervention effect of N candidate intervention strategies based on dynamic digital twinning sets a privacy invasiveness index for each candidate intervention strategy, and performing multi-objective evaluation based on the expected intervention effect, comprising: Simulating N candidate intervention strategies based on the user physiological and behavior data, inputting each candidate intervention strategy into a dynamic digital twin body to be deduced, and generating an expected intervention effect, wherein the expected intervention effect comprises a change track of the user physiological and behavior data in a future specified time window; And carrying out multi-objective evaluation on the expected intervention effect, calculating the comprehensive utility score of each candidate intervention strategy, and generating an evaluation result based on the comprehensive utility score, wherein the multi-objective evaluation comprises validity evaluation, continuity evaluation and user preference evaluation.
5. The intelligent pillow sleep monitoring and personalization improving device based on digital twinning according to claim 1, wherein the converting the personalized intervention strategy into a control command and issuing the control command to the intelligent pillow, and the intelligent pillow jointly executing the control command with external environment equipment comprises: acquiring a personalized intervention strategy and analyzing to generate a control instruction, wherein the control instruction comprises a first type instruction and a second type instruction; the first type of instruction is used for being connected with the intelligent pillow and driving the intelligent pillow to execute a personalized intervention strategy; the second type of instruction is connected with the external environment equipment based on the intelligent pillow, and the external environment equipment is driven to execute the personalized intervention strategy based on the intelligent pillow selection.
6. The intelligent pillow sleep monitoring and personalization improvement device based on digital twinning of claim 1, wherein the user implicit physiological feedback data comprises heart rate data changes, respiratory rate data changes, and frequency changes of body movement data within a preset time window after execution of emergency intervention and control instructions.
7. The intelligent pillow sleep monitoring and personalization improvement device based on digital twinning of claim 1, wherein the obtaining implicit physiological feedback data of the user after execution of the emergency intervention and control instruction, calculating an intervention consensus score, comprises: comparing the implicit physiological feedback data of the user with the expected intervention effect, and calculating the intervention fitness; analyzing the sleep stage based on the implicit physiological feedback data of the user to generate a sleep stage improvement condition; An intervention co-insight score is calculated based upon the intervention compliance and the sleep stage improvement situation.
8. The intelligent pillow sleep monitoring and personalization improvement device based on digital twinning according to claim 7, wherein the parameter updating and structure self-evolution of the dynamic digital twinning according to the intervention commonality score comprises: when the scores of the presidentifiability are larger than or equal to the positive learning threshold value, the dynamic digital twin is subjected to parameter updating by adopting a first learning rate; when the scores of the prespecified factors are smaller than the aggressive learning threshold and larger than or equal to the conservative learning threshold, the parameters of the dynamic digital twin are updated by adopting a second learning rate; and when the scores of the prespecified factors are smaller than the conservative learning threshold, carrying out parameter updating on the dynamic digital twin by adopting a third learning rate, wherein the first learning rate is greater than the second learning rate and greater than the third learning rate.
9. The intelligent pillow sleep monitoring and personalization improvement device based on digital twinning of claim 8, wherein the parameter updating and structure self-evolution of the dynamic digital twinning according to the intervention co-intelligibility score comprises: Setting a structural evolution threshold and an observation time window, and starting the self-evolution of the structure of the dynamic digital twin body if the average value of a plurality of pre-intelligibility scores is continuously lower than the structural evolution threshold in the observation time window; m characterization components are arranged in the dynamic digital twin body, the contribution degree of each characterization component to the expected intervention effect is calculated, a contribution degree threshold value is set, and the characterization components with the contribution degree lower than the contribution degree threshold value are removed; Analyzing based on the physiological and behavioral data of the user, extracting the characteristic data which does not exist in the dynamic digital twin, creating a new characterization component based on the characteristic data, and initializing the connection relation with other characterization components.

Description

Intelligent pillow sleep monitoring and personalized improving device based on digital twinning Technical Field The invention relates to the technical field of medical health information, in particular to an intelligent pillow sleep monitoring and individuation improving device based on digital twinning. Background Along with the rapid development of the technology of the health internet of things, intelligent bedding, particularly an intelligent pillow, is an important carrier for sleep health monitoring and intervention, in the prior art, the sleep quality is monitored through a sensor as disclosed in the publication No. CN112704365B, sleep is assisted when the user is monitored to be insomnia, the breathing rate is monitored through the sensor as disclosed in the publication No. CN107669055A, a sleeper is directly awakened through vibration when the breathing rate is abnormal, and a detection report is generated through detecting various parameters of the sleeping state and sleeping posture of the user through a module as disclosed in the publication No. CN 117643454A. In the prior art, the intelligent pillow is generally integrated with various sensors to monitor physiological and behavioral data such as heart rate, breathing, snoring and sleeping posture of a user, and based on preset simple logic, if the snoring is detected, the intelligent pillow is lifted, and a single intervention is performed. The following technical problems exist in the prior art: Firstly, the decision model in the existing intelligent pillow is usually static and solidified, the intervention effect is generally based on fixed logic, such as detecting abnormal breathing rate, directly waking up, and the like, and the intervention model has dynamic data monitoring and data optimization, but cannot perform structural optimization on the components of the model, cannot perform self-adjustment and optimization according to the unique sleeping habit of a user and long-term data change, so that the suitability is reduced after long-term use, and the intervention effect is gradually attenuated; Secondly, the existing intervention strategy generation mechanism aiming at sleep monitoring is focused on solving single and instant sleep problems such as snore stopping, respiratory interruption preventing and the like, lacks comprehensive trade-off of multi-objective such as intervention effectiveness, sleep continuity and long-term preference of users, and lacks consideration of psychological feelings and privacy boundaries of users during intervention, so that intervention means are hard and abrupt, and even a new problem such as user being interfered and awakened is caused while a problem is solved, so that the acceptance and use viscosity of the users are reduced, and the use experience is reduced. Disclosure of Invention The intelligent pillow sleep monitoring and individuation improving device based on digital twinning comprises a processor and a memory, wherein the executing method of the device comprises the following steps: Acquiring user physiological and behavior data which are acquired in the intelligent pillow through a sensor array in a non-sensing way, judging whether to trigger conditional reflex based on hard-wired logic, if so, executing emergency intervention, otherwise, constructing a dynamic digital twin body based on the user physiological and behavior data; simulating expected intervention effects of N candidate intervention strategies based on the dynamic digital twin body, performing multi-objective evaluation based on the expected intervention effects to generate evaluation results, setting privacy invasiveness indexes based on the evaluation results, performing strategy fusion, and generating personalized intervention strategies; converting the personalized intervention strategy into a control instruction and transmitting the control instruction to the intelligent pillow, wherein the intelligent pillow jointly executes the control instruction by combining with external environment equipment; And acquiring implicit physiological feedback data of the user after the emergency intervention and control instruction execution, calculating an intervention consensus degree score, and carrying out parameter updating and structure self-evolution on the dynamic digital twin body according to the intervention consensus degree score. Further, the sensor array comprises a micro-vibration sensor, a pressure sensor, a microphone and a reflective photoelectric sensor, and the physiological and behavioral data of the user comprise heart rate data, respiratory rate data, sleep posture data, body movement data, environment data, snore intensity and blood oxygen saturation data. Further, the hardwired logic includes a first judgment condition and a second judgment condition; acquiring real-time respiratory rate data, and triggering a first judgment condition to perform emergency intervention if the respiratory rate signal is identified