Databases Reference
In-Depth Information
The Internet-of-Things (IoT) and System-of-Systems (SoS) have the po-
tentials to provide the industry with imminence solutions for integrating and
automating large-scale heterogeneous systems. For instance, in plant mainte-
nance, data captured from the sensors and field devices can be transmitted
immediately to the processing network and be made directly available to the
plant engineer for analysis, trigger a request for parts procurement to the
finance department, or trigger a delivery request to the warehouse manager.
Although conceptually this approach offers significant added-value to the day-
to-day operations of industries and enterprises, there remains a gap between
theory and practice that lies at the knowledge level.
With the advent of IoT and SoS models, fully connected environments are
possible. However, the knowledge potentials of high-connectivity between het-
erogeneous systems in a large-scale ecosystem must be extensively explored.
For example, studies on cause and effect relationships can be extended to in-
clude different elements and parameters. Questions such as whether the flood-
ing of a road or a road accident will impact a household's electricity usage or
whether the rapid increase of temperature in a boiler room of a plant will re-
duce the duration of the procurement process. These questions will definitely
be frowned upon because they not a part of the normal human experience.
Nevertheless, an increased interconnectivity achieved through both IoT and
SoS models have the potential to make such questions a reality. In the future,
the aims are to discover the knowledge potentials of infrastructures that are
able to add value to the way interpretations and analysis can be accomplished
in our daily life's operations. This can only be achieved using distributed pat-
tern recognition and analysis schemes when addressing such an Internet-scale
environment.
Developing a capability for large-scale recognition and interpretation
schemes based on both IoT and SoS models will require a detailed understand-
ing of the complex relationship between the devices and information systems
at a bigger scale. A set of accurate recognition models must be formulated
to follow as real systems. This is to provide reasonable approximations of the
resultant behavior. In this perspective, the level of sophistication required in
knowledge of the occurrences of event will be of higher level than what is
currently required, e.g., in weather prediction applications.
10.3 Making a Case
A case for Internet-scale pattern recognition has been made through the
design and implementation of DPR, which has been extensively described in
this topic. Recognition by means of computational intelligence can no longer
be established simply on the basis of algorithmic accuracy and e
ciency. The
