Databases Reference
In-Depth Information
FIGURE 8.1: Fully distributed DHGN configuration for a fine-grained net-
work.
The DHGN deployment must be able to perform single-hop communication
between adjacent nodes for the message exchange process. To ensure that
e
cient energy-communication utilization is achieved during the recognition
process, the physical distance should be taken into account when implementing
the DHGN distributed recognition application in WSN. A DHGN deployment
in fine-grained networks will be presented in the next chapter.
8.1.1.2
Clustered Configuration
The clustered configuration maps each DHGN subnet over a single process-
ing node. Each node is capable of conducting the recognition process based
on the input subpatterns obtained from the SI module node. In this config-
uration, the recognition process involves the entire input subpattern and the
processing node is expected to acquire high processing capability and storage
capacity. However, because the communication involves only message com-
munication from the SI module node to each of the processing nodes, the
communication costs between sensor nodes are minimized. Figure 8.2 shows
the clustered configuration of the DHGN algorithm for WSN.
Each processing node in the clustered DHGN configuration can perform
recognition on each subpattern independently. The node must be able to pro-
vide su
cient processing and storage capacity to conduct the recognition pro-
cess. This configuration is intended to be used on coarse-grained networks,
