Information Technology Reference
In-Depth Information
used in sensors can be categorized into two groups: rechargeable and
nonrechargeable. Often in harsh environments, it is impossible to recharge
or change a battery. Current sensors are developed to be able to renew their
energy from solar or vibration energy [4,14]. The alkaline battery has a wide
voltage range and large physical size while the lithium battery provides a
constant voltage supply but with very low nominal discharge currents. The
nickel metal hydride battery can be recharged but with a significant decrease
in energy density [4].
The data fusion has occupied the interests of many researchers and there
are many published papers. Here we present a sample of these papers.
However, none of these have attempted to formulate the data fusion as a
scheduling problem and try to find its boundaries.
Joint directors of laboratories' (JDL) model [1] is a conceptual data fusion
model, which is defined as three levels of fusion: object, situation, and threat
refinement. Also, the model provides a common framework and terminol-
ogy for the data fusion community to unify concepts and terminology. The
object refinement level defines the objects identified in the sensing area.
Situation refinement defines the relationships between objects and events.
The threat refinement projects future threats, vulnerabilities, and opportuni-
ties for operations.
The work in paper [8] describes current state of data fusion schemes for
wireless sensor networks such as: Kalman filtering [2], blind beamform-
ing [6], transferable belief model [13], filter-based techniques [3], and linear
mean square estimator [10]. The methods in [2,6,10,13] are all based on the
JDL model for data fusion; however, they are only designed to work at level
one, which refers to the object refinement.
The work in paper [16] focused on identifying a route for a mobile agent
through a subset of sensor nodes, which indicate the presence of a target by
utilizing the signal strengths of the sensor nodes. They formulate the mobile
agent routing problem in the distributed sensor network as a combinatorial
optimization problem involving the cost of communication and path loss.
Their objective is to maximize the sum of a signal energy received at the
visited nodes while minimizing the power needed for communication and
path losses. One of the most important aspects of mobile agents that was not
addressed by this work is the security.
In paper [15], the researchers developed a computation technique based
on a particle swarm optimization (PSO) method for obtaining the optimal
power scheduling scheme for data fusion in a wireless sensor network when
the distributed sensors collected and observations are correlated.
Trying to make sense of the observed data is the focus of the paper [5],
where the authors proposed a three-layer observer network architecture,
consisting of the application (top), services, and sensors (bottom). The
goal of the observer network is to automate the reasoning of accepting or
rejecting data.
