Information Technology Reference
In-Depth Information
a
8
t
i
-6
a
7
t
i
-1
t
i
-5
a
2
t
i
-3
t
i
-4
t
i
-2
a
5
a
5
c
t
i
-6
t
i
-7
a
1
a
1
a
7
a
8
a
2
a
4
a
4
t
i
-5
t
i
a
2
a
1
a
7
a
5
a
8
a
5
a
8
a
7
a
1
a
4
a
2
a
4
2
*
p
t
i
-4
a
1
memory
a
2
a
5
a
7
a
8
a
2
a
1
a
5
a
8
a
4
a
7
2
*
p
a
2
t
i
t
i
-1
t
i
-2
t
i
-3
t
i
-4
t
i
-5
t
i
-6
a
4
t
i
-3
a
5
a
8
a
7
a
4
a
1
a
2
t
i
-2
a
5
a
8
a
1
known to
a
2
at
t
i
, flocking
unknown to
a
2
at
t
i
a
4
a
2
a
7
Fig. 4.4
DDIG builds a memory of the current movement history of nearby nodes. Example sensor
node
a
2
with communication range
c
roams from
t
i
−
6
to
t
i
and encounters nearby nodes. Whenever
a
2
is connected to nearby nodes within 2
∗
p
-disc is only illustrated for
t
i
), it pulls their position
history.
a
2
received histories from
a
4
at
t
i
−
4
, from
a
8
at
t
t
−
3
, from
a
5
at
t
i
−
1
(entire history illustrated
for
t
i
−
1
), and finally from
a
1
and
a
7
at
t
i
.Memory
a
2
reflects the corresponding build-up of local
knowledge. Note,
a
2
at
t
i
remembers previous positions of nodes it is currently not connected to (
a
4
,
a
5
,and
a
8
) (Republished from Laube, P., Duckham, M., & Palaniswami, M. (2011). Deferred Decen-
tralized Movement Pattern Mining for Geosensor Networks.
International Journal of Geographical
Information Science, 25
(2), 273-292, Taylor & Francis, DOI:10.1080/13658810903296630.)
∗
p
(2
in the given tasks. Imperfection came in the form of error of omission for small
communication ranges. Imperfection also came in the form of latency (deferred
processing) allowing the nodes to collect information and collaborate for its process-
ing.
DDIG
showed that both in simulated and real data scenarios, detection error
could be reduced by longer latency times. Latency is discussed in Sect.
4.3.3.3
.
4.3 Decentralized Movement Analysis Principles
This section revisits the research included in this topic and enumerates a set of infor-
mation processing principles that illustrate the opportunities for decentralized spatial
computing arising from mobility in geosensor networks. The principles listed here
take advantage of properties of the studied phenomena or the models used for their
