Transcript Dust to Doctors: WSN for Assisted Living

Microsoft Research
Faculty Summit 2007
John A. Stankovic
BP America Professor, UVA
Microsoft Corporation
Problems of interest
Univ. of Virginia AlarmNet
Microsoft interactions
HealthGear
SensorMap
Summary
Home health care
Home health care
(Large scale) assisted living facilities
Home health care
(Large scale) assisted living facilities
Flexible targeting of care
to a person’s health condition
Longitudinal studies
Assisted Living and
Residential Monitoring Network
Nurses Stations
Internet
I
n
t
e
With Harvard
With Harvard
With MARC UVA
Medical School
* With the Univ. of Illinois
Activity of daily living classification
Five two-axis accelerometer network
for classifying activities with a HMM
Dust
Light
Pollen
Humidity
Temperature
Motion
Activities
PDAs
Internet
Nurses Stations
I
Nurses Stations
Privacy
Security
Internet
Patients
Patients family and friends
Doctor – what advantages
for them in treating patients
Nurse
Technician
Orderly
Admin
Social Worker
Personal medical data
Personal activity data
Environmental data
Contextual data
Longitudinal data
System Performance data
User’s
Request
Reply
Request Authorizer
Policy Manager
Inconsistency
Check
Privacy
Policy
Context
Manager
Context
RealTime
Database
Data mining
analysis
Request
History
Nurses Stations
Privacy
Security
Internet
AlarmGate SW
on stargate
DB
Nurses Stations
Privacy
Security
Internet
Circadian activity rhythms (min)
60
50
Bedroom
40
Kitchen
30
Living
room
Bathroom
20
WC
10
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
24 hour cycle
16
17
18
19
20
21
Circadian activity rhythm per room for 70 days
22
23
24
Diurnal/nocturnal
activity
Life Habits
at-home
Learning
period
Support for heterogeneous
health care systems
Utilize Cell phones
BlueGate, WiGate, EtherGate
Unobtrusive body networks
Behavior based systems management
Exploiting circadian activity rhythms
HealthGear
Pulse
Oximeter
BlueGate
Gateway
AlarmNet
Alarms
Any Data
Bluetooth
WiGate
Gateway
POGA
Blood
Pressure
Sensor, ECG,
…
SeeMote
IEEE 802.15.4
Smart
phone
EtherGate
Gateway
PC, DB,
AlarmNet
Backend
StarGate
Gateway
WiFi, Ethernet
HealthGear
Pulse
Oximeter
Smart
phone
Bluetooth
WiGate
AlarmNet
BlueGate
Pulse, SpO2
Tripwire Motion, Light
BP
Scale
WiFi,
Dust
ECG
IEEE 802.15.4
Switch
Environmental
SeeMote
Ethernet
BlueGate
Bluetooth to MICAz, 802.15.4,
2.4GHzRF
Bluetooth to MICA2, 4xx/9xx MHz RF
WiGate
802.15.4 to WiFi (802.11b/g)
EtherGate
802.15.4 to wired Ethernet with POE
NordGate
Nordic RF (ECO mote, Poga2) to
802.15.4 and/or Bluetooth
Antenna
RF, MCU, Accl.
Main board
Li Ion Battery
Classify and detect ADLs - accelerometer
Two versions
Tmote-Mini based – 802.15.4 compatible
(designed/built)
Fast design time using Tmote module
TinyOS compatible
PHY-compatible with ZigBee
Nordic RF chip based (future work)
Smaller form factor/cost (expected 10x20mm w/o battery)
Detection of
multiple people
User
Identification
Alerts and Messages
for the caregivers
Detection of irregular
behaviors and anomalies
Motion sensors
Other Sensors
Context Aware
Power
Management
(CAPM)
Sensor network
Backbone
Activity pattern inference
Through Circadian
Activity Analysis
Base Station
Power consumption
(J)
Power consumption estimation
10000000
Total energy required
consumption with CAPM
8000000
6000000
Total energy required
consumption without
CAPM
4000000
2000000
0
Total energy saved
1
2
3
Users
4
5
Publishing and searching
private data streams
Privacy guarantees
Anonymous aggregates
Quality control for search results
Rank real-time/archived sensor data search results
AlarmNet
Privileged applications
One administrative domain
MetroNet
Public information
Many administrative domains
Must support
Privacy
Mobility
Data annotation
Rank data streams like PageRank
algorithm ranks web pages
All data need not be at central databases
Aggregated streams created
by users – can be ranked
Microsoft
Univ. of Virginia Medical School
Harvard: (some) body networks
Johns Hopkins: Triage mote
Univ. of Illinois: Smart clothing
http://www.cs.virginia.edu/wsn/medical/
© 2007 Microsoft Corporation. All rights reserved. Microsoft, Windows, Windows Vista and other product names are or may be registered trademarks and/or trademarks in the U.S. and/or other
countries. The information herein is for informational purposes only and represents the current view of Microsoft Corporation as of the date of this presentation. Because Microsoft must respond to
changing market conditions, it should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of any information provided after the date of
this presentation. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.