Cryptography Reference
In-Depth Information
applications, the data sensed by the sensors has a certain amount of redundancy and
correlation. It would be desirable if the sink could only receive relevant data, for faster
and better decision making. For this purpose, the sensed data should be processed at
intermediate sensors before reaching the sink. The benefit of this in-network process-
ing, such as data fusion, can be seen when vector data rather than scalar data are being
transmitted. For example, in an application monitoring the temperature of a room, the
sensed data is scalar (i.e., integer or real values). Hence, the cost of data communication
is not very high, and the data fusion or aggregation is not as costly. But continuously
sending unnecessary and redundant data will consume a huge amount of energy. If a
sensing application has to send a large amount of data, for example, images, to the sink
for further analysis and processing, it would consume a huge amount of energy. In this
case, it would be more beneficial if those images, sensed by different sensors, could be
aggregated and only a few of them sent. However, it is also true that processing those
images for data fusion requires a considerable amount of energy. Moreover, there will
be a delay due to the processing of those images. Therefore, there is a trade-off between
data communication and fusion, in these types of information-intensive networks,
where the sensed data is not scalar but rather vector.
7. 2 . 3 .1.
Low-Energy Adaptive-Clustering Hierarchy
In terms of the energy efficiency of a WSN, traditional routing protocols based on a
f lat architecture were not very efficient. Hence, alternative methods were considered in
which the clustering approach emerged as a viable option among the existing alterna-
tives. In a cluster-based model, the network is scattered into several clumps or clusters,
and each cluster is supervised by a resource-abundant node called a cluster head. The
cluster head is in charge of all the necessary coordination between the sensors and acts
as a local sink to all the nodes in the cluster. The cluster head could preprocess the
aggregated data before forwarding it to the concerned sink node. As a result, a cluster-
ing-based approach would be beneficial in a WSN environment since the transmission
distance between the ordinary nodes and their respective cluster heads is much smaller,
when compared to their respective sink nodes. One such example of a cluster-based
protocol that aggregates the sensed data into smaller sizes of data so that only mean-
ingful information is conveyed to all the further forwarding nodes is the Low-Energy
Adaptive Clustering Hierarchy (LEACH) (Heinzelman et al. 2002).
LEACH divides a network into several clusters of sensors that are created by using
localized coordination and control, so that it not only reduces the amount of data trans-
mitted to the sink but also makes routing and data dissemination more scalable and
robust. To conserve the energy levels of ordinary sensor nodes, the selection of cluster
heads is not done in a deterministic manner. Instead, it uses a randomized rotational
method in which every node in the network has an opportunity to be the cluster head.
Consequently, constant depletion of energy levels of individual sensors is significantly
reduced. This scheme can be classified into five phases: advertisement, cluster setup,
schedule creation, and data transmission.
Search WWH ::
Custom Search