Information Technology Reference
In-Depth Information
be verified (iii) if the objects feature the given attri-
butes, and (iii) if the described relations hold among
the given objects. Inference machines can find
out whether these two requirements are fulfilled
on the given data. The process of concluding if a
certain expression (description of a critical situ-
ation) is correct, incorrect or correct to a certain
degree of accuracy is called inductive reasoning.
The JDL Data Fusion Model defines and
describes a process for becoming aware of the
current situation and reason about possible future
situation evolution. The term situation detection is
broadly used within the scientific community, but
focuses in most cases only on the technology or
methodology for detecting situations. Especially
in the graphics and video processing community
a big amount of approaches to detect situations
based on a given picture or video stream exists.
Impact analysis and decision making steps are
usually not involved. Therefore, a situation detec-
tion system can be compared to an L1/L2 situation
awareness system. For detecting situations several
kind of approaches are available: (iii) rule based
approaches following if... then... conditions, which
are described by (Storf, Becker, & Riedl, 2009),
(iii) supervised learning approaches defined by
(Kotsiantis, 2007), where classifiers are trained
using historical or simulated data, (iii) the usage
of finite state machines (Mahajan, Kwatra, Jain,
Kalra, & Banerjee, 2004) and (iv) the usage of
ontology-based reasoning (Kokar, Matheus, &
Baclawski, 2009). Due to the manifold require-
ments for AmI systems, each of the available ap-
proaches for detecting situations has its eligibility
and is more or less feasible for certain purposes.
Besides, assertions about a given expression,
most inference machines provide the functionality
to list recursively all possible prerequisites which
are necessary to fulfill an expression. Within the
eHealth domain this can be used for monitoring
vital signs, like blood pressure, blood sugar, pulse
rate etc. A doctor can define critical ranges for
these values and a smart reasoning system using
inference can conclude when these values get
toward the defined critical ones. Then the system
can suggest counter measures, like doing exercises
or relaxing or maybe informing some nursery or
healthcare personnel in advance.
Modeling of eHealth Processes
To support interoperability concerning health data
processing as well as the interdisciplinary ex-
change of these data among institutions, adequate
Health Service Environment (HSE) models can
be used (see Figure 5). To guarantee that a HSE
model covers the varying needs of several system
partners and services and to assure the necessary
coverage, an analysis of the different healthcare
data gathering systems and the healthcare docu-
ments and data has to be conducted first. Using
the results, a transfer model and a data model can
be created. A possibility for enhancing standard-
ized data exchange is to base the transfer model
Figure 5. Architecture of our HSE model
Search WWH ::




Custom Search