Image Processing Reference
In-Depth Information
may be unused, so the channel might be not fully utilized. Other classical MAC protocols for WSNs,
such as PMAC [Zhe] or WISEMAC [Hoi], store neighboring information, so in dense WSNs
the memory requirements on sensor nodes might be undesirably high, or only a small subsection of
neighbor information could be maintained.
The Crankshaft protocol tries to solve the problems that classical schedule-based MAC proto-
cols for WSNs have when they are applied to dense WSNs. Both overhearing and communication
grouping are reduced by alternating the power-off of sensor nodes, while the bandwidth exploitation
is enhanced by using receive-slots instead of send-slots. In addition, the proposed mechanism does
not require per-node neighborhood information. he basic idea of the protocol is that nodes should
stay awake listening for incoming packets at fixed offsets from the start of a frame. he name comes
from the analogy with engines, where the time when a piston fires is a fixed offset from the start of
the crankshaft rotation.
The Crankshaft protocol divides time into frames that are further divided into slots. Two types
of slots are defined: unicast and broadcast. Every node has to be awake during the broadcast slots
in order to receive broadcast communications. Instead, each node listens for only one unicast slot
in every frame. he only exception is the sink node that stays awake during all the time-slots of the
frame. A node willing to transmit a packet contends for the medium access, either during a broadcast
slot, or during the unicast slot in which the destination node is active. his is because the sink node is
assumed to be less energy-constrained than the other nodes, and also because most of the traffic in a
WSN is typically destined to it. Nodes do not need explicit slot assignation, as a very simple criterion is
used to decide the listening time-slot. If n is the number of the time-slots inside a frame, the listening
time-slot of each node is obtained by calculating MAC address modulo n . In order to improve energy
efficiency even further, each time-slot is divided into two different parts, i.e., the contention window
and the message exchange window. he source nodes have to wake-up during the former and resolve
the contention, while the destination wakes-up in the second part. This protocol can achieve good
energy efficiency and good convergecast delivery ratio in dense WSNs, but with high latency values.
In addition, the rigid structure of the protocol may limit its applicability.
8.2.15 Correlation-Based Collaborative MAC
A different approach that achieves energy efficiency exploiting spatial data correlation is presented
in [Vur]. This protocol, called a spatial Correlation-based Collaborative MAC (CC-MAC), is
designed for event-driven WSNs. Usually, in such networks, data from spatially separated sensor
nodes is more useful than highly correlated data from close nodes. Reducing the transmission
attempts of correlated (and redundant) data it is possible to save energy, bandwidth, and time. So,
the basic idea of the CC-MAC protocol is to intelligently manage the transmissions, taking into
account the spatial correlated nature of the event information. In fact, it is not necessary that all
the nodes sensing an event transmit their data to the sink, but a smaller set of measurements may
be sufficient. However, as the number of the nodes that transmit their data is reduced, the informa-
tion decoded by the sink is degraded. So, also the reliability of the event detection is reduced. The
aim of the CC-MAC protocol is to exploit the spatial correlation to reduce the transmission attempts
without compromising the event detection performance, by introducing a distortion constraint to
be met. Thus, the minimum number of representative nodes that achieve the distortion constraint
has to be selected. Intuitively, the minimum distortion is achieved when the representative nodes are
located close to the event source, but as far from each other as possible. Another benefit of selecting
representative nodes far from each other is that spatial wireless channel reuse can be achieved. In
order to exploit the spatial correlation in a distributed fashion, an Iterative Node Selection scheme is
proposed, that exploits the statistical properties of the node distribution to compute the correlation
radius, r corr . his algorithm is executed during the network initialization phase. he computed radius
is then used by the distributed MAC mechanisms. Two different MAC mechanisms are defined for
event detection (E-MAC) and packet forwarding (N-MAC), respectively. hey are both based on the
Search WWH ::




Custom Search