Image Processing Reference
In-Depth Information
network stack, i.e., just under the network layer. Their objective is to improve the energy efficiency
of routing protocols for wireless networks by coordinating the sleep transitions of nodes. Several
routing protocols in fact try to enhance network lifetime by reducing the number of data transmis-
sions or balancing the transmission power, but neglect idle power consumption. However, several
measurements, e.g., in [Ka] [Ste], show that idle power dissipation should not be ignored, as it
could be comparable to the transmitting or receiving power. herefore, in order to optimize energy
consumption, nodes should turn off their radios. Topology control protocols exploit redundancy in
dense networks in order to put nodes to sleep while maintaining network connectivity. hey can be
applied to standard routing protocols for ad-hoc networks or for WSNs that do not directly han-
dle sleep schedules. Although some of them are designed for wireless ad-hoc networks rather than
WSNs, the typically high redundancy of sensor nodes and the need for maximum energy saving
make WSNs perhaps the most suitable type of networks for taking advantage of these protocols. In
the following sections, a few examples of topology control protocols are given.
7.8.1 Geographic-Adaptive Fidelity (GAF) Protocol
The GAF [Xu] protocol, in order to put nodes into low-power sleep states without excessively
increasing the packet loss rate, identifies groups of nodes that are “equivalent” in terms of routing
cost and turn off unnecessary nodes. This is achieved by dividing the whole area into virtual grids,
small enough that each node in a cell can hear each node from an adjacent cell. Nodes are location-
aware, so each sensor obtains its coordinates on the virtual grid from location information. Nodes
thatbelongtothesamecellcoordinateactiveandsleepperiods,sothatatleastonenodepercellis
active and routing fidelity (which requires that in any cell at any one time there is at least one node
able to perform routing [Xu]) is maintained.
According to the GAF protocol, nodes can be in three different states: discovery, sleeping, and
active. Transitions from one state to another are depicted in Figure ..
In the discovery state nodes are active and exchange discovery messages in order to find nodes
within the same cell. Then, after a discovery timeout
T
d
, a node enters the active state. Each node
only stays in the active state for a defined time
T
a
,thenmovestothediscoverystateagain.Anodein
thediscoveryoractivestatecangotothesleepstateifitindsanequivalentnodewithhigherrank
that handles routing. Finally, the sleep period is also limited, so that after a time
T
s
the node returns
Sleeping
Receive
discovery msg
from high rank
nodes
After
T
s
Active
After
T
d
Discovery
After
T
a
FIGURE .
State transitions of the GAF protocol.
