Nonintrusive Physical Sign Sensing Key Technologies For Health Mornitoring

The progressive aging of population is currently one of the vital population issues for the whole world.Due to body functions degeneration,elderly people are a high-risk group suffering from chronic diseases,which brings a heavy healthcare burden to families and society.In particular,the incidence of chronic respiratory diseases and Dyskinesia is very high in elderly group.The characteristics of chronic respiratory diseases and Dyskinesia can be summarized as: 1)the course of the disease is long,2)early symptoms are not obvious,and 3)pathologic condition gradually changes over time.The reliable assessment of the above two types of disease relies on long-term continuous respiratory and motor function monitoring.Eventhough the traditional method based on clinical diagnosis can achieve accurately physical sign mornitoring,they require patients to wear special monitoring equipments(e.g.,for respiration monitoring: nasal tube,electrode,etc.;for motor function assessment: inertial sensors,reflective marker,etc.),which make it incapable for long-term use in ordinary living environment.Walking and sleeping are two most natural behaviors.Monitoring elders' respiration during sleeping and gait during walking continuously without interfering their daily actions is an ideal choice to tackle the challenges caused by chronic respiratory diseases and Dyskinesia.In this dissertation,we explore the key techniques of nonintrusive physical sign sensing.Specifically,this thesis is devoted to investigate the techniques for:1)Contactless respiration monitoring during sleep.We investigate contactless respiration monitoring using acoustic signal with commodity audio devices.Comparing with other kinds of contactless respiration monitoring method(e.g.respiration monitoring using RF signal and optical signal),acoustic signal can be transmitted and received by commodity speaker and microphone widely available in home environments,which is cost-effective.There are two sensing targets for respiration mornitoring: chest movement displacement and exhaled airflow during respiration.The respiration monitoring methods based on chest movement measurement can not always accurately monitor respiration,since the chest movement displacement is not always a reliable indicator of human respiration.The respiration monitoring method sensing exhaled airflow is sensitive for interference airflow around human body.These two respiratory monitoring methods are suitable for different environments.In this dissertation we investigate both two respiration methods.2)Nonintrusive gait analysis and pathologic gait pattern recognition during walking.We study the gait analysis and gait pattern recognition method based on plantar pressure data,which can be easily obtained through a pressure insole and provides fine-grained movement characteristics from both time domain and kinetic domain.Comparing with healthy gait,pathologic gait(e.g.Parkinsonian gait)caused by Dyskinesia shows specific movement pattern change.Taking the gait of early Parkinson's patients as an example,we analyze and identify the specific gait patterns change from the plantar pressure data.The main contributions in this thesis are as follows:1)Propose a C-FMCW based contactless respiration monitoring approachWe propose a contactless respiration detection system by measuring the chest movement displacement during breathing using acoustic signals with off-the-shelf speaker and microphone.We firstly propose a Correlation based Frequency Modulated Continuous Wave method(CFMCW)which is able to achieve high ranging resolution.Based on C-FMCW,we design and implement a contactless human respiration monitoring system by sensing chest movement displacement with commodity speaker and microphone,which are widely available in home environments.Extensive experiment results show that: 1)the median error of respiration monitoring of the proposed system is lower than 0.35 breaths/min,outperforming the state-ofthe-arts,and 2)the proposed system is robust to various practical factors,including different scenarios,Apnea,respiration rate variation and sensing distance,etc.2)Propose an acoustic Doppler shift based contactless respiration monitoring approachWe propose a contactless respiration detection system by directly sensing the exhaled airflow using acoustic signals with off-the-shelf speaker and microphone.Based on the fact that exhaled airflow can scatter acoustic signal and results in Doppler effect,our system works as an acoustic radar which transmits sound wave and detects the echo Doppler effect caused by breathing airflow.We mathematically model the relationship between the Doppler effect change and the direction of breathing airflow.Based on this model,we design a MDL-based algorithm to effectively capture the Doppler effect caused by exhaled airflow.Extensive experiment results show that that our system accurately mornitors human respiration with the median error lower than 0.3 breaths/min.The results also demonstrate that the system is robust to various practical factors,including different respiration styles(shallow,normal and deep),respiration rate variation,ambient noise,sensing distance variation(within 0.7 m)and transmitted signal frequency variation.3)Propose a fine-grained movement function measurement based Parkinsonian gait pattern recognition approachWe propose a novel computation framework to recognize gait patterns in patients with PD.We firstly put forward a sliding window based method to discriminate four gait phases from plantar pressure data.Based on the gait phases,we extract and select fine-grained gait features which characterize movement stability,symmetry and harmony.Finally,we recognize PD gait patterns by building a hybrid gait pattern recognition model.We evaluate the framework using an open dataset that contains real plantar pressure data from 93 PD patients and 72 healthy individuals.Experimental results demonstrate that our framework can accurately recognize Parkinsonian gait pattern(precision rate 87.9%,recall rate 87.4%)and significantly outperforms the four baseline approaches.4)Design and implement an integrated nonintrusive health sensing systemTwo respiration monitoring systems and gait pattern recognition method mentioned above are integrated to provide nonintrusive physical sign sensing services for elders.Firstly,two respiration monitoring systems are implemented on only one pair of acoustic devices,so that the two systems not only can run simultaneously but also share same hardware.Then,we pack the gait analysis and gait pattern recognition module as a software interface.User can directly use the gait analysis and gait pattern recognition functions by feeding the gait data collected from smart insoles.Based on the fundamental functions,i.e.,respiration monitoring,gait analysis and gait pattern recognition,we design and implement several extended services,such as apnea alarming,abnormal respiration pattern alarming,pathology gait alarming,motor function declining alarming,history data visualization etc.Finally,we test the reliability and stability of the integrated nonintrusive health sensing platform.