Databases Reference
In-Depth Information
FIGURE 9.5: Maximum memory consumption for each DHGN subnet as a
function of the pattern size. The DHGN uses minimal memory space when
processing small pattern sizes.
by including a spatio-temporal consideration has been reported in [96]. In
this work, pattern recognition was performed on the overall responses from
the wireless sensor nodes obtained. In this context, the signals sent by each
sensor node will colectively form a spatio-temporal representation in the form
of spatial coordinates and time instances. The detection area used in this
analysis is represented in the form of Cartesian coordinates with the grid size
of 9×5 obtained from the data set in [97]. Figure 9.6 shows a scenario of
spatio-temporal analysis using the proposed scheme.
Note the changes in the distribution of event across different geographical
coordinates over 25 time instances (t01 to t25). Note that the light-shaded
area represents the area in which forest fire was detected using the allocated
wireless sensor node with DHGN scheme, while dark-shaded area represents
the area in which its sensor node responses with non-event signal. The non-
shaded area represents area with no active node at the respective time in-
stances.
This capability of deploying spatio-temporal event detection opens up fur-
ther possibility for enhanced event detection mechanism capable of providing
measures on the distribution of events over a specified time period. Further-
more, simple event tracking approach may be performed by examining the
order of magnitude and the direction of the occurrences of events, as shown
in Figure 9.7.
