Monitoring Homes with Wireless Sensor Networks

brewerobstructionAI and Robotics

Nov 7, 2013 (3 years and 9 months ago)

55 views

Monitoring
Homes with
Wireless Sensor Networks


C.H.S. Gibson
1
, T. van Kasteren
2
, B. Kröse
3

1
Professional University of Amsterdam, NL;
2
Universiteit van Amsterdam, NL;
3
Professional
University and University of Amsterdam

c.h.s.gibson@hva.nl



B
ackground
: The escalating cost of nursing care and the projected increase in the number of
elderly patients requiring institutionalized care is an important stimulus for promoting aging
in place. The importance of developing
autonomous systems which are capable of effectively
monitoring activities of daily living in a h
ome environment can hardly be over
estimated.


M
ethod
: This paper describes and discusses the implementation of a home monitoring system
using simple reed switc
h sensors transmitting data through low power wireless network nodes
to a database server. The choice of simple reed switch sensors was dictated by the need to
create a non
-
intrusive system of ubiquitous sensors which would not impact
unduly
on the
daily l
ife of the residents being monitored. Camera observation was excluded as this was
considered by the researchers and by the home residents to be unacceptably intrusive. The
prohibitive
cost of installing and maintaining a system consisting of perhaps a hund
red or
more wired sensors was also important in determining the choice of hardware.


After selecting the hardware the network was configured in a star
topology
. Programs were
developed to read the sensor impulses from the gateway node and store the impulse
s with
a
record
of
the originating node and time stamp in a central database. The resulting test system
was then successfully implemented in an ‘assisted living’ apartment in a residential complex
attached to a nursing home in the Netherlands.


O
bjective
:
By associating patterns of sensor impulses and time stamps with the annotated list
of daily activities, the system can be used to observe activities of daily life (ADLs) in
residences for the elderly. The aim of the system is to enable observation and anal
ysis of the
ADLs to be carried out by autonomous intelligent systems, invoking Bayesian networks to
detect patterns which may be indicative of situations requiring external intervention.


R
esearch project
: The project was initiated at the Digital Life Cent
re of the professional
University of Amsterdam, in Almere in the Netherlands. The installation and programming of
the network hardware was performed
by
students of the Institute of Information Engineering.
The annotated list of daily activities was compile
d under supervision by students from the
department of ergotherapy at the Professional University of Amsterdam. The development of
the Bayesian network analysis is currently being carried out at the University of Amsterdam.


Keywords:

wireless sensor netwo
rk, aging in place, activities of daily life, home monitoring