Image Processing Reference
In-Depth Information
in [Li] with a computationally simpler algorithm which builds smaller subnetworks, thus providing
lower link maintenance costs and less overhead. However, while this protocol minimizes power con-
sumption for data forwarding, it does not maximize the overall network lifetime, which is usually the
main target in WSN.
7.6.2 Geographic and Energy-Aware Routing (GEAR) Protocol
The GEAR protocol, described in [Yu], uses an energy-aware metric along with geographical
information to efficiently disseminate data and queries across a WSN. Unlike other geographical pro-
tocols not specifically devised for sensor networks, such as the well-known greedy perimeter stateless
routing (GPSR) protocol [Kar], this protocol addresses the problem of forwarding data to each
node inside a target region. his feature enables GEAR to support data-centric applications.
According to the GEAR protocol, each node has to know, besides its own location, the geographical
position and residual energy of all its neighbor nodes. This can be accomplished through a low-
frequency (and low-cost) hello message exchange. In addition, each query has to specify the target
region,i.e.,theareainwhicheverynodeshouldreceivethemessage.Eachnodemaintainsthelearned
costs,
h
(
N
,
R
)
,foreachneighbor-regionpairwherepacketshavetobeforwarded.First,whenthere
is no
h
entry for the neighbor
N
, this cost is computed as a function of the distance between
N
andthecentroidoftheregion
R
and the energy consumed at node
N
. After the node has selected the
next hop neighbor
N
min
,thecost
h
(
N
,
R
)
,wherethelasttermisthe
communication cost from
N
to the selected neighbor
N
min
. he estimated cost can subsequently be
updated through feedback from the receiver node. In fact, after each packet is delivered, the learned
costissentbacktothelasthop.hus,ifthedestinationisreachedwith
n
hops, ater
n
subsequent
packets to the same target the correct cost is propagated to the source node. hanks to this mechanism
it is possible to avoid holes simply by forwarding packets to nodes with minimum learned costs. hus,
the forwarding rule when the target region is not reached is always to send data to the neighbor
N
i
whose
h
(
N
,
R
)
is set to
h
(
N
min
,
R
)+
C
(
N
,
N
min
)
is minimum. If there are no holes, the learned cost will only represent a combination
of consumed energy and distance, so it will be equivalent to the estimated cost. In the presence of
a hole, on the other hand, the updated learned cost will act as a “resistance” to following the path
toward that hole.
Since the objective of this protocol is to disseminate queries inside the target region
R
,data
forwarding does not end when a packet reaches that region, as data must be forwarded to every
node inside
R
. To efficiently achieve this behavior two different mechanisms are proposed, i.e., the
Recursive Geographic Forwarding or the Restricted Flooding algorithm.
RecursiveGeographicForwardingisusedwhennodedensityishigh.hisisarecursiveapproach
in which if a node
N
receives a packet destined to its region
R
,itsplits
R
into four different subregions,
and sends four copies of the packet, each targeted at one of these subregions. he recursive algorithm
terminates when the current receiver is the only node inside the target region. his algorithm works
well when node density is high, but with low densities it is inefficient and in some cases can never
terminate and keeps routing uselessly around an empty target region before the packet's hop-count
exceeds a certain bound. hus when node density is low, the use of restricted flooding is suggested.
Restricted flooding exploits the broadcast medium of the wireless channel and only sends one
broadcast message to all its neighbors, but every node in its transmission range receives this broadcast
message whether it is an intended receiver or not.
hisprotocolachievesenergyeiciencybymeansofthelearnedcostthattakesresidualenergyinto
consideration along with geographical information about neighbors. By minimizing the energy cost
between neighbors, an approximation of the lowest energy cost path is found. In order to improve
network lifetime, the energy consumed so far is taken into account rather than the consumption of
a single transmission. This protocol, compared with GSPR, features reduced energy consumption
and a higher packet delivery ratio, at the expense of higher delay, as it takes longer paths in order to
(
N
i
,
R
)
