wiree ess se soless sensor networks for hmdilhome medical...

62
Wireless Sensor Networks for H M di l Home Medical Care Care John A. Stankovic BP America Professor Department of Computer Science Department of Computer Science University of Virginia http://www cs virginia edu/wsn/medical/ http://www .cs.virginia.edu/wsn/medical/ http://wirelesshealth.virginia.edu/

Upload: others

Post on 30-Jul-2020

1 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Wireless Sensor e ess Se soNetworks for

H M di lHome Medical CareCare

John A. StankovicBP America Professor

Department of Computer ScienceDepartment of Computer ScienceUniversity of Virginia

http://www cs virginia edu/wsn/medical/http://www.cs.virginia.edu/wsn/medical/http://wirelesshealth.virginia.edu/

Page 2: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

What’s Wrong With WiresWhat s Wrong With Wires

And we don’t want a patient tethered to a bed or fixed medical device

Page 3: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

OutlineOutline

• Problems and Vision

• Univ. of Virginia AlarmNet System– Architecture– Main Ideas/Results

• Current Work and Summary• Current Work and Summary

Page 4: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

The ProblemsThe Problems

A i P l ti• Aging Populations• High Cost of Medical Careg• Lack of Facilities• Quality of Life Issues• Quality of Life Issues

• Solution: Home Health CareCCRCCCRCAssisted Living

Page 5: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Vision - Smart Living SpaceVision Smart Living Space• Humans-in-Loop

H• Heterogeneous• Evolution• Open• Privacy

Page 6: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Large Scale DeploymentsLarge Scale Deployments

Page 7: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

State of ArtState of Art

• UCLA, Harvard, Yale, GaTech, MIT, Univ of Washington, Johns Hopkins, Imperial College, U. of Geneva, UPenn, UVA, etc.

• GE Health, Intel, Philips, Verizon, IBM, etc.• West Wireless Health Center• West Wireless Health Center• Wireless Life Sciences Alliance

• Europe, Asia, US

Page 8: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

“3” Open Questions3 Open Questions

S l• Scale– Numbers of sensors– Number of smart home units– Number of facilities– Number of functions on body networks– Numbers of body networksNumbers of body networks

• Activity Recognition (AR) not accurate enoughenough

• Safety

Page 9: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Goals - A System ViewGoals A System View

• Tailored to patient• Evolves over timeEvolves over time• Seemlessly integrate heterogeneous

t h ltechnology• Largely Passivea ge y ass e• 24/7 Monitoring and Care

Page 10: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

BenefitsBenefits

Id tif l b h i• Identify normal behaviors• Identify anomalous behaviors• Detect medical problems (depression) early• Improve quality of lifey• Monitor adherence to and effectiveness of

treatments• Detect dangerous situations• Maintain privacyMaintain privacy• Longitudinal studies

Page 11: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

AlarmNetAlarmNetAssisted Living and Residential Monitoring Network• Assisted Living and Residential Monitoring Network

• In-Lab TestbedIn Lab Testbed

• Privacy – deployed in 8 homes• Detecting Falls – students• CAR – 22 patients in Assisted Living• Sleep Study 10 subjects• Sleep Study - 10 subjects• Body Sensor Networks

• Deployment Plans – Depression in the Elderly– Deployed in one home

Page 12: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

AlarmNet ArchitectureAlarmNet Architecture

PDAs

Innte

InternetNurses Stations

Page 13: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

With HarvardWith Harvard

Page 14: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

With Harvard

Page 15: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

With MARC UVAMedical School

Page 16: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Sleep MonitoringSleep Monitoring

• Sleep motion (restlessness and agitation)

• Sleep quality

Page 17: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Using Physiological SignalsUsing Physiological Signals

• EEG: measures brain waves• EOG: measures eye movements• EMG: measures electrical activity

f lof muscles

Di d t• Disadvantages– Expensive

Uncomfortable– Uncomfortable– Measure once/twice

Page 18: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Wearable Devices in Home Environments

• Actiwatch

• Headband - Zeo

• Disadvantage• Disadvantage– Users need to wear

the devicesthe devices

Page 19: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Non-Wearable SolutionsNon Wearable Solutions

• Pressure PadsDi d t– Disadvantage

• Not entirely comfortable• Do not infer body positionsDo not infer body positions

• Cell Phone Apps• Cell Phone Apps– Built-in accelerometers are

used– Disadvantage

• Not robust

Page 20: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

WISPWISP

• Combines RFIDCombines RFID technology with sensorssensors

• Used to sense light, temperature andtemperature and accelerationPowered and read by• Powered and read by RFID readers

Page 21: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

WISP Instrumented Mattress

Page 22: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Body Position InferenceBody Position Inference

F diff b d i i i i f f h• For different body positions, orientations of one or more axes of the accelerometers with respect to gravity are different

• We combine the readings from all three tags to infer body position

Page 23: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Body Position InferenceBody Position Inference

D i t i i f h b d iti f th• During training, for each body position of the subject, we construct a 9-tuple from the readings of the three tagsof the three tags

• We train a Bayesian classifier with these tuples

• We use this classifier to infer body positions during sleep

Page 24: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Controlled Experiments for Body Position Inference

10 bj t• 10 subjects• 3 different mattresses

• Each subject lies in each of the 4 body positions for 2.5 minutes eachminutes each

• For each position, we use the data from the first 2 p ,minutes for training and next 30 seconds for evaluating accuracy of body position inference

Page 25: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

ResultsResults

• 3 settings:t1 diff ti t b t– set1: differentiate between

the bed being empty or occupied

– set2: differentiate between empty lying and sittingempty, lying and sitting

– set3: differentiate between all lying positions, empty and sitting

Page 26: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Realistic Overnight Experiments

• 6 nights• DDR pads (sense pressure)

used as baseline system• Also compare with an

iPhone application: SleepiPhone application: Sleep Cycle

• We also recorded the video• We also recorded the video of the 6 nights’ sleep

Page 27: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Evaluation by Cross Validation

• 6 Evaluation sets• In each set we train our system based onIn each set, we train our system based on

5 nights of data and evaluate the performance of the remaining nightperformance of the remaining night

Page 28: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Movement Detection Evaluation

DDR pad

• Ground Truth

p

– Validated the performance of DDR d bDDR pads by comparing with 3 hours video

WISPhours video

– DDR pads are considered ground gtruth

Page 29: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Body Position InferenceBody Position Inference

• Ground Truth– Collected from the recorded video

Acc rate ithin 5%– Accurate within 5%

Page 30: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Medical StudiesMedical Studies

• Correlation between sleep movement and agitation with incontinence in dementia gpatients

Combine with acoustic and wetness sensors– Combine with acoustic and wetness sensors

Page 31: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

AlarmNet ArchitectureAlarmNet ArchitectureDust

hLightPollenHumidityHumidityTemperatureMotion

PDAsActivities

Innte

InternetNurses Stations

Page 32: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Key PointsKey Points

• Self-configuring - Highly flexible (radio shack model))

N t b dd d l t• New sensor types can be added later

• Contributes to Activity Recognition (AR)

Page 33: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

AlarmNet ArchitectureAlarmNet Architecture

PDAs

InP i nte

InternetNurses Stations

PrivacySecurity

Page 34: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

AlarmGateAlarmGate

• Netbridge device (Stargate)– single board computerg p– embedded Linux

400MHz Xscale– 400MHz Xscale– mote daughterboard– wireless ethernet

Page 35: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Privacy - Many Stakeholders

P ti t• Patients• Patients family and friendsy• Doctor – what advantages for them in

treating patientstreating patients• Nurse• Technician

O d l• Orderly• Admin• Social Worker

Page 36: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Privacy - Many Data TypesPrivacy Many Data Types

• Personal medical data• Personal activity dataPersonal activity data• Environmental data• Contextual data• Longitudinal data• Longitudinal data• System Performance data

Page 37: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Authorization Framework

User’s Request

ReplyRT

Device

Request Authorizer Database

Policy Manager Context Manager

Data mining analysisInconsistency

Analyze AuditTrail

Manager analysisyCheck

Privacy Policy Context Request

HistoryLOGAuditTrail

Violate PrivacyOn Heart Attack

Page 38: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Fingerprint And Timing-based Snoop attackFingerprint And Timing-based Snoop attack

Adversary

KitchenBedroom #2 Fingerprint and Timestamp Snooping Device

Locations and

Living Room

Bathroom

T1

T2

Timestamps Fingerprints Locations and Sensor Types

?

?Living Room

Bedroom #1

T2

T3

… …

?

?

…Front Door

V S i i J St k i K Whit h P t ti Y D il I HV. Srinivasan, J. Stankovic, K. Whitehouse, Protecting Your Daily In-Home Activity Information fron a Wireless Snooping Attack, Ubicomp, 2007.

Page 39: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

ADLADL

ADL i f d• ADLs inferred:– Sleeping, Home Occupancy– Bathroom and Kitchen Visits– Bathroom Activities: Showering,

Toileting, Washing– Kitchen Activities: Cooking AdversaryKitchen Activities: Cooking

• High level medical information

Adversary

Fingerprint and Timestamp Snooping Device

inference possible

HIPAA i h lth

Timestamps Fingerprints Locations and Sensor Types

• HIPAA requires healthcare providers to protect this information

T1

T2

T3

?

?

?

… …

?

Page 40: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

PerformancePerformance

• 8 homes (X10) - different floor plans– Each home had 12 to 22 sensors

• 1 week deployments• 1, 2, 3 person homes• Violate Privacy - Techniques CreatedViolate Privacy Techniques Created

– 80-95% accuracy of AR via 4 Tier Inference FATS l ti• FATS solutions– Reduces accuracy of AR to 0-15%y

Page 41: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Key PointsKey Points

P i i iti l ( t )• Privacy is critical (many types)

• Overridden on alarms

• Use dynamic context and request history

• Inconsistency checking algorithmsInconsistency checking algorithms required

Page 42: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

AlarmNet ArchitectureAlarmNet ArchitectureDustustPollenHumidityTemperatureMotionActivities

PDAs

Activities

InP i nte

InternetNurses Stations

PrivacySecurity

Page 43: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Graphical InterfacesGraphical Interfaces

• PDA real-time query issuer – template based• Circadian Activity Rhythmsy y• Nurse’s station monitoring• Embedded displays• Embedded displays

Page 44: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Depression Detection andMonitoring

• Multi-modal

• Passive

• Combines Objective and Subjective• Combines Objective and Subjective Measures

Page 45: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Depression MonitoringDepression MonitoringPatient Display Caregivers

Di lPatient Display Display

Depression Inference

Eating Sleep Quality Movement Mood Weight Gain/Loss

DBDB

Motion and Contact Sleep Data PHQ-9 Acoustic Weight

Page 46: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations
Page 47: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations
Page 48: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Caregivers DisplayCaregivers Display

Page 49: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

PDA Real-Time QueriesPDA Real Time Queries

AlarmGate SW on stargate

DB

Page 50: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

SenQSenQ

I t ti E b dd d Q S t• Interactive, Embedded Query System– Peer to peer

• Streams – define discover and share• Streams – define, discover and share• Virtual sensors – discover and share• Devices added/deletedDevices added/deleted• Optional Modules• Location Transparencyp y• UI - Developers, Domain Experts, Users• Privacy and Security

A. Wood, L. Selavo, J. Stankovic, SenQ: An Embedded Query SystemFor Streaming Data in Heterogeneous Interactive Wireless Sensor NetworksFor Streaming Data in Heterogeneous Interactive Wireless Sensor Networks,DCOSS, 2008.

Page 51: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

SenQ LayersSenQ Layers

• Loosely coupled layers

Page 52: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Sensor Data Sampling & Processing

Page 53: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Sensor Data Sampling & Processing

• Virtual Sensors– users fuse streams to

make new sensors– sensor drivers can

i l i k S Qrecursively invoke SenQ

Page 54: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Query ManagementQuery Management

Page 55: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

AlarmNet ArchitectureAlarmNet ArchitectureDustustPollenHumidityTemperatureMotionActivities

PDAs

Activities

InP i nte

InternetNurses Stations

PrivacySecurity

Page 56: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Circadian Activity RhythmsCircadian Activity Rhythms

22 ti t• 22 patients• 3 months to 1 yeary• 7 males; 15 females• Ages 49 93• Ages 49-93• All ambulatory• Weekday; weekend; seasonal• Eliminate times when not in facilityEliminate times when not in facility• Learning - 2-3 weeks of normal behavior

Page 57: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Circadian RhythmsCircadian Rhythms60

n)

50

hms

(min

30

40

vity

rhyt

h Bedroom

Kitchen

Living

20

ian

activ

groomBathroom

WC

0

10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Circ

ad

Ci di ti it h th f 70 d

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

24 hour cycle

Circadian activity rhythm per room for 70 days

Page 58: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

AnomaliesAnomalies

• Examples– Retroactively analyzed the anomaliesy y

• Detected “depression” – much more time in bed• Detected increased urination at nightDetected increased urination at night• Detected different behavior upon return from

hospitalizationp

G. Virone, et. al., Behavioral Patterns of Older Adults In Assisted Living,IEEE Transactions on Information Technology in Biomedicine, Vol. 12,No. 3, May 2008.

Page 59: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

SummarySummary

• Wireless Health– Body Sensor Networksy– Environmental and AR Networks

E t M dif Ti• Easy to Modify over Time– Incorporate new technology as it becomes

available– Adapt as medical conditions changep g

• Protects Privacy

Page 60: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations
Page 61: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Diabetes Depression

Eating Level

ToiletingLevel

SleepingLevel

MovementLevel

LightLevel

WeightLevel

eating toileting showering sleeping Light Weight

Kitchenvisits

bedroomvisits

bathroomvisits

Personal location tracking

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

front floor

fridge

microwa

pantry

cook top

sink flush

entranc

sink shower

motion

motion

motion

weight

light light light pressureoo e o a

vey top a c

ee o o o t su e

bed room

kitchen

kitchen

kitchen

kitchen

kitchen

bath room

bath room

bath room

bath room

bed room

kitchen

bath room

bed room

bed room

kitchen

bath room

bed room

Page 62: Wiree ess Se soless Sensor Networks for HMdilHome Medical Carecse.snu.ac.kr/sites/default/files/node--seminar/20110523... · 2018-08-17 · T1 T2 Timestamps Fingerprints Locations

Current ResearchCurrent ResearchData Association (multi person homes) new height• Data Association (multi-person homes) – new height sensor

• Run Time Assurance – safety

R b t AR• Robust AR

• Scaling• Scaling

• Fall Detection

• BSN