Aucun titre de diapositive

1
ALARM-NET WIRELESS SENSOR NETWORKS FOR
ASSISTED-LIVING AND RESIDENTIAL MONITORING
G. Virone, A. Wood, L. Selavo, Q. Cao, L. Fang,
T. Doan, Z. He, R. Stoleru, S. Lin and J.A.
Stankovic Department of Computer Science,
University of Virginia
We have developed a residential wireless sensor
network (WSN) called Alarm-Net for smart
healthcare that will open up new opportunities
for resident health monitoring in assisted or
independent living facilities. This poster
describes our vision of the architecture,
objectives and advantages of Alarm-Net. Results
demonstrate a strong potential that will open new
research perspectives for ad hoc deployment of
multi-modal sensors and improved quality of
medical care.
I. System overview
IV. Real-time resident monitoring Front-ends

• Quality of life
• Well-being, Autonomy, Comfort, Visits
• Cognitions, Nutrition, Hygiene
• Resident Health Remote Monitoring
• Real-time (WSN)
• Long-term for longitudinal studies(back-end
  system)
• Climate Monitoring
• Environmental conditions control
• Pollution detection
• Security
• Detection of at-risk situations
• Alert triggering

A General User Interface, on a based station,
permits the users to access to the resident
medical and personal information.
A PDA displays accelerometer data, patient
pulse-rate, and environmental temperature.
Pills
Assisted Living Facilities
Video cameras
Motion sensors
Architecture
Backbone nodes
Motes (emplaced WSN)
A body network, embedded in a jacket, records
human activities such as walking, eating and
stillness using three 2-axis accelerometers. It
also incorporates a GPS to track outdoor location.
Front-ends Policy Enforcement Point (authenticatio
n) Query issuing
Sensors Policy Enforcement Point (authorization-pr
ivacy) Query processing
Routing
Experimental Smart Living Space

Dedicated data flows
Query Management Data filtering Aggregation Cachin
g
Long-term monitoring applications
Archiving
Alerts
Policy Decision Point (PDP) Authentication Authori
zation Dynamic privacy
An LCD mote, used for wearable applications,
enables sensor readings, reminders and queries,
and can accept rudimentary inputs from the
wearer.
Heterogeneous Power Management
Real-time Monitoring application
Network arbitration Data association Security
WSN R.T. medical apps
Data queries Authentication Privacy
Backbone
DB
II. Multi-sensor data acquisition
V. Long-term resident monitoring
Body scale
In the back-end of the system, a medical
application monitors the Circadian Activity
Rhythms (CAR) to extract activity patterns and
detect behavioral anomalies.
Harvard University
III. Wireless Sensor Network
Emplaced Sensor Network. It includes sensor
devices deployed in the assisted living
environment (rooms, hallways, units, furniture)
connected to a more resourceful backbone. Sensors
communicate wirelessly using multi-hop routing
and may use either wired or battery power.
Bed sensor. Based on an air bladder strip, it
records bed movements, breathing and heart rate.
VI. Research topics
Backbone. It connects traditional systems, such
as PDAs, PCs, and in-network databases, to the
emplaced sensor network. Nodes possess
significant storage and computation capability,
for query processing and location services. Yet,
their number, depending on the topology of the
building, is minimized to reduce cost.
Wearable InterfacesLCD Mote
Wireless Sensor Network. Heterogeneous power
management depending on the life habits of the
resident, topology, routing, network arbitration,
aggregation, caching. Dynamic privacy. The
system monitors and collects patient data,
subject to privacy policies, depending on the
current behavioral status of the resident, and
detected anomalies. Security. Security
mechanisms are present throughout the system.
Data association. To know who is doing what in
a system where biometric identification is not
always accessible and where multiple persons may
be present at the same time. Data fusion.
Back-end software programs to analyze autonomy,
behavior and health status of the resident.
Body networksPulse, EKG
Emplaced SensorNetwork
In-network and Back-end Databases. One or more
nodes connected to the backbone are dedicated
in-network databases for real-time processing and
temporary caching. Back-end databases are located
at the control center for long-term archiving,
monitoring and data mining for longitudinal
studies.
Backbone
Back-end Database
Caregiver InterfacesPC, PDA