Information Technology Reference
In-Depth Information
4.3.1 Challenges
Duckham (
2012
) suggests
neighborhood-based
or
location-based
decentralized
spatial computing algorithms. Whereas neighborhood-based algorithms have only
access to minimal spatial information in the form of the identities of neighbors,
location-based algorithms exploit more complex spatial information such as direc-
tion, distance, and location of nodes. Movement challenges the underlying assump-
tions for both types of decentralized algorithms.
First, basic preliminaries about the spatial distribution of nodes and assumptions
about the resulting communication graphs do no longer hold. When nodes move
around, their spatial arrangement can be very heterogeneous, resulting in clusters
of dense deployment separated by large empty gaps. This holds specifically when
mobile nodes interact, hence meet, or are constrained to some form of transportation
infrastructure. Establishing a communication network is difficult when the nodes
show a heterogeneous and potentially unfavorable distribution.
Second, roaming nodes result in constantly changing topological node constella-
tions regarding neighborhood, and changes in distances and direction. Apart from
the danger of temporarily broken communication links, building up and maintaining
data structures with enriched information about neighborhood, distances, or direc-
tions is at best difficult and often impossible. Nodes that are now close by neighbors
move on and can be far away after a short while. Also, many DeSC algorithms
function based on Tobler's first law of geography, assuming that the sensor readings
of neighboring nodes tend to be similar. This assumption is also challenged, since
both the monitored phenomena and the sensing nodes may move about. In essence,
spatial structure and contiguity are no more static but potentially change constantly.
Interacting, exchanging, and enriching information is even more difficult than in
conventional geosensor networks.
Finally, when both the sensing system and the monitored phenomena can be
mobile, it is difficult to know if changing sensor readings are due to an actual change
in the monitored natural or built environment or simply a result of the sensing node
getting a different perspective of an actually unchanged phenomenon.
4.3.2 Specific Decentralized Movement Analysis Principles
4.3.2.1 Mobility Compensation
Mobility compensation is a strategy for enlarging the reach of a sensor node. Nodes
can be limited with respect to the perception area of their sensor (e.g., measuring the
temperature around a node) or with respect to their communication range (maximal
distance to maintain a communication link with a neighbor). Both, perception and
communication range are often modeled and approximated with a disk around the
sensor of a given radius. In static geosensor networks, individual sensor nodes might
