Situation-Aware Ambient Assisted Living and Ambient Intelligence Data Integration for Efficient Eldercare - E-Healthcare Systems and Wireless Communications - page 336

putting incoming data into responsible pipeline(s).
A patient, who is monitored by the system, has at
least one pipeline running, which only operates on
the patient's data. This prevents mixing up data
from one patient with other monitored patients.
Since pipelines operate in parallel, they can be
used to improve scalability by using state-of-the
art processing facilities, like multi-core CPUs.
Anything can be used as filter, like an ordinary
processing unit to check a specific data value
or a complex finite state machine or reasoning
component. Filters can interact with services, like
alarm services for detected critical situations, or
data exchange services for data export.
Every component outside the iResidence TM
core system is modeled as service which can be
hosted locally or remotely for load balancing
and performance reasons. iResidence TM provides
a data exchange service, alarm services, home
automation services and management services.
The management services offer a management
console and a web user interface to administer the
patient's data and the corresponding processing
pipes. They are used by healthcare personnel to
administer patients in a nursing home. The data
exchange service is used for communication with
the environment or third party components and
consists of four main components : (i) a service
responsible for permanently storing collected data,
(iii) a service communicating with IHE compli-
ant systems, (iii) a service for generating alarms
(sending SMS to DECT phones or mobile phones),
and (iv) a service for home automation (remote
switchable power connectors).

The amount of gathered data can, in addition,
be used to analyze ADLs. An ADL can be a simple
activity, like entering and leaving a restroom, or
be more complex and longer lasting, like cooking
a meal. BehaviorScope (Bamis, Lymberopoulos,
Teixeira, & Savvides, 2009) is a project from the
Yale University which addresses these challenges.
BehaviorScope uses a high-level grammar to de-
scribe activities in advance and is able to success-
fully generate summaries of everyday activities
and trigger notifications at run-time.
The iResidence TM system uses a similar ap-
proach. Domain experts can describe situations by
using a domain specific grammar (Hasse, 2010).
Basic activities are described as terminal symbols,
like EXIT, TOILET and use of WASHBASIN,
COOKER, FRIDGE, TV, BED, MEDICINE, as
well as PC (see Figure 8).
A parser, based on ANTLR (Parr, 2010), is
used to identify the described situations by ana-
lyzing an incoming data stream. The recorded
activities may vary in their duration, can remain
incomplete or can be executed simultaneously or
in a varying order. The detected situations are
displayed as shown in Figure 9. Such an ADL
analysis can help care personnel in their daily
work, for example to early detect disorders in the
daily living of a patient.

Figure 8. Grammar for describing activities of daily living (Hasse, 2010)