03-05-2012, 03:45 PM
Improving the Performance of Wireless Ad Hoc Networks Through MAC Layer Design
Improving the Performance of Wireless.pdf (Size: 890.91 KB / Downloads: 42)
Abstract
In this paper, the performance of the ALOHA and
CSMA MAC protocols are analyzed in spatially distributed
wireless networks. The main system objective is correct reception
of packets, and thus the analysis is performed in terms of outage
probability. In our network model, packets belonging to specific
transmitters arrive randomly in space and time according to
a 3-D Poisson point process, and are then transmitted to their
intended destinations using a fully-distributed MAC protocol.
A packet transmission is considered successful if the received
SINR is above a predefined threshold for the duration of
the packet. Accurate bounds on the outage probabilities are
derived as a function of the transmitter density, the number of
backoffs and retransmissions, and in the case of CSMA, also the
sensing threshold. The analytical expressions are validated with
simulation results. For continuous-time transmissions, CSMA
with receiver sensing (which involves adding a feedback channel
to the conventional CSMA protocol) is shown to yield the
best performance. Moreover, the sensing threshold of CSMA
is optimized. It is shown that introducing sensing for lower
densities (i.e., in sparse networks) is not beneficial, while for
higher densities (i.e., in dense networks), using an optimized
sensing threshold provides significant gain.
Index Terms—Ad hoc networks, Poisson point process, MAC
protocols, outage probability.
I. INTRODUCTION
IN the design of wireless ad hoc networks, various techniques
are applied to efficiently allocate the scarce resources
available for the communication links. Using an appropriate
medium access control (MAC) protocol is one such
technique. Taking into account the system’s quality of service
(QoS) requirements, a MAC protocol for ad hoc networks
shares the medium and the available resources in a distributed
manner, and allows for efficient interference management.
In this paper, we consider a spatial network model in which
nodes are randomly distributed in space, and we address
the problem of interference through MAC layer design. The
ALOHA and CSMA MAC protocols are employed for communication,
and the success rate of packet transmissions is
investigated. In particular, we ask the following questions: (a)
Given a fixed signal-to-interference-plus-noise ratio (SINR)
threshold for each transmitter (TX) receiver (RX) link in the
Manuscript received March 3, 2010; revised June 29, 2010 and October 7,
2010; accepted October 8, 2010. The associate editor coordinating the review
of this paper and approving it for publication was O. Dabeer.
M. Kaynia and G. E. Øien are with the Dept. of Electronics and Telecommunications,
Norwegian University of Science and Technology, Trondheim,
Norway (e-mail: {kaynia, oien}[at]iet.ntnu.no).
N. Jindal is with the Dept. of Electrical and Computer Engineering,
University of Minnesota, MN, USA (e-mail: nihar[at]umn.edu).
Digital Object Identifier 10.1109/TWC.2010.110310.100316
network, what is the probability of successful transmission
for ALOHA and CSMA, (b) can the performance of CSMA
be improved by introducing feedback between the TX and
RX and allowing the RX to make the backoff decision, and
© does CSMA have an optimal sensing threshold which
minimizes the outage probability (OP) for received packets?
We consider a network in which packets are located randomly
in space and time according to a 3-D Poisson point
process (PPP), consisting of a 2-D PPP of TX locations in
space and a 1-D PPP of packet arrivals in time. The packets,
which are assumed to be of constant length, are forwarded by
each TX over a nonfading channel to a RX a fixed distance
away. In order to derive precise results, we focus exclusively
on single-hop communication, as in [1], [2], [3]. All multiuser
interference is treated as noise, and our model uses
the SINR to evaluate the performance (in terms of OP) of
the communication system. The only source of randomness
in the model is in the location of nodes and concurrent
transmissions, which allows us to focus on the relationships
between transmission density, OP, sensing threshold, and the
choice of MAC protocol.
A. Related Work
There has been a notable amount of research done on the
performance of ALOHA in ad hoc networks. A number of
researchers have analyzed slotted ALOHA using a Poisson
model for TX locations, considering transmission capacity and
success probability of the network [2], [4], [5]. Ferrari and
Tonguz [6] have analyzed the transport capacity of slotted
ALOHA and CSMA, showing that for low transmission densities
the performance of slotted ALOHA is almost twice that
of CSMA. Also, it is established that CSMA is advantageous
only at high transmission densities. Other related works have
considered the performance of ALOHA and CSMA in terms of
throughput and bit error rate [6], [7], [8]. Some of these also
assert CSMA’s superiority over ALOHA, which is naturally
followed by tradeoffs in other domains such as transmission
rate and delay [8] [9]. The seminal work of Gupta and Kumar
[7] considers the transport capacity of a Poisson distributed
ad hoc network, which resembles a slotted version of our
model. However, their analysis focuses on a deterministic
SINR model, and employs a deterministic channel access
scheme, thereby precluding the occurrence of outages. Weber
et al. [5] revise this model by considering a stochastic SINRbased
model, within which they find tight lower and upper
bounds to the OP of slotted ALOHA as a function of the
1536-1276/11$25.00 ⃝c 2011 IEEE
KAYNIA et al.: IMPROVING THE PERFORMANCE OF WIRELESS AD HOC NETWORKS THROUGH MAC LAYER DESIGN 241
node density. We consider the model used in [5], and extend
it to also cover unslotted systems.
Despite all the research done on MAC protocols thus
far, only a limited number of works have considered an
interference channel that is both stochastic and continuous
in time [2], [4], [5], [10]. Perhaps the closest work is that
of Hasan and Andrews [4], where the success probability of
slotted ALOHA is analyzed within such a stochastic ad hoc
wireless network model. Success probability is defined as the
probability that a transmission is received successfully at the
RX. This is equal to 1−OP. In [4], a scheduling mechanism
is assumed that creates an interferer-free guard zone, which is
in effect a theoretical circle around the RX, within which no
interfering TXs are allowed. By means of geometrical analysis,
the density of successful transmissions is maximized under an
outage constraint. We adopt the concept of guard zones in our
analysis, with the difference that instead of incorporating into
the protocol a guard zone within which no TX are permitted,
we consider actual MAC protocols that employ virtual guard
zones in order to make the backoff decision and evaluate the
OP. Other analytical models may also be used for performance
evaluation [11], [12]. In [11], the throughput and fairness of
CSMA/CA is evaluated based on a Markovian analysis. In
[12], an analytical framework is introduced to evaluate the
per-flow throughput of CSMA in a multi-hop environment.
However, in both of these works, a single communication link
is considered. Thus, the complexity of various links’ performance
and decisions being inter-dependent, as in interference
channels, is ignored. This is considered in our model.
A good choice of the sensing threshold of CSMA,