Abstract
Continuous patient
monitoring requires real-time processing of signals over a long period of time,
varying from few hours to many days, depending on the patient’s medical
condition and the symptoms being monitored. For example, in fall monitoring,
movement-related information is required to distinguish normal activities of
daily living (ADL) from abnormal. It requires continuous monitoring during
daytime, in which a person is engaged in many day-to-day activities, as shown
in the extensive survey on monitoring human movements to detect falls.
Monitoring of sleep-related fall and other disorders such as apnea, requires
data for 8 h during the night, in which a person is often found less active.
With the study of sizeable data collected over the past from passive
monitoring, researchers are now interested in studying physiological signals
for early detection, prevention, and prediction of anomalous events.
The passive monitoring
systems consist of simple off-the shelf cameras and on/off sensors placed on
doors, toilets, beds, chairs in the individual’s living area, to monitor his/her
activities. In addition, the system provides reminders for medication and
similar instructions. When people interact with the environment, infrared or
pressure sensors on the floor or bed are triggered and the system is able to
recognize an abnormal activity, such as a fall.
There are many limitations with this approach:
it requires installation time; has limited coverage capability (only where
there are sensors) and privacy violations. Both the camera and the
environmental sensor devices require prebuilt infrastructures which enable
their use only in hospitals and houses, but not in the outdoor environment. A
large amount of redundant data is collected and often incidents are missed. For
example, it is not possible to detect a fall occurring outside a floor sensor.
No comments:
Post a Comment