A MAC protocol based on adaptive beamforming for ad hoc networks

Abstract

This paper presents a novel slotted MAC (medium access control) protocol for nodes equipped with adaptive antenna array in ad hoc network. The protocol relies on the ability of antenna to uses DOA (direction-of-arrival) information to beamform by placing s in the direction of interferers thus maximize SINR (signal to interference and noise ratio) at the receiver. We studied the performance of the protocol using joint simulation in OPNET and Matlab. We studied the impact of variable number of antenna elements, DOA algorithm, and ing. The performance of our new protocol is compared against one of the recent directional MAC protocols [R.R.N.H.V. Romit Roy Choudhury, Xue Yang, 2002]. We observe that despite the simplicity of our protocol it achieves high throughput.

Full text

Page 1

A MAC protocol based on Adaptive Beamforming
for Ad Hoc Networks
Harkirat Singh and Suresh Singh
Department of Computer Science
Portland State University
harkirat@cs.pdx.edu
Abstract—This paper presents a novel slotted MAC (Medium
Access Control) protocol for nodes equipped with adaptive an-
tenna array in ad hoc network. The protocol relies on the ability
of antenna to uses DOA (Direction-Of-Arrival) information to
beamform by placing nulls in the direction of interferers thus
maximize SINR (Signal to Interference and noise ratio) at the
receiver. We studied the performance of the protocol using
joint simulation in OPNET and Matlab. We studied the impact
of variable number of antenna elements, DOA algorithm, and
nulling. The performance of our new protocol is compared
against one of the recent directional MAC protocols [5]. We
observe that despite the simplicity of our protocol it achieves
high throughput.
I. INTRODUCTION
Recently, there has been increasing interest in developing
MAC protocols for use in ad hoc networks where nodes
are equipped with directional antennas. Antenna models used
include sectored fixed beam antennas, idealized adaptive ar-
ray antennas, and steerable directional antennas. As previous
researchers have shown, using directional antennas increases
throughput because of better spatial reuse of the spectrum
(see [1], [5], [2], [3]). However, we note that these previous
works have not fully exploited the benefits of adaptive array
antennas (or smart antennas) such as the ability to form nulls
in the direction of interferers (resulting in high SINR) and the
ability to determine the direction of transmitters (Direction
of Arrival). We show that by exploiting these capabilities of
smart antennas, a simple protocol can yield throughputs that
are 2x – 4x higher than one of the recent protocols [5]. We also
note that our simulations use realistic antenna models unlike
the idealized models used in many (with the exception of [2])
papers and, despite this, our protocol out performs most of
these existing protocols.
There are two key capabilities of antenna arrays that we
exploit in developing DOA-MAC: the ability to form directed
beams and place nulls in given directions, and the ability to
determine the direction of arrival of signals from multiple
transmitters. In DOA-MAC, a small initial portion of the slot
is used for finding the direction of various transmitters (all of
which transmit directionally). This is done by requiring each
transmitter to transmit a pure tone1(no source and destination
id) towards its intended receiver for a short interval prior to
This work is funded by the NSF under grant ANIR-0125728.
1The tone format is a combination of the DSSS PLCP Preamble and PLCP
header as given in IEEE 802.11.
transmitting the packet. The receiver runs a DOA algorithm
which provides information about the received signal strength
and direction of the different transmitters. This information is
then used at each receiver to guide beamforming (beam and
nulls) for the remaining duration of the slot. Upon correct
packet reception, a receiver sends an ACK using the already
formed beams.
Our work here differs from all of the above papers in the
following ways: our adaptive array antenna model is made up
of a linear array of antenna elements and we exploit DOA
information as well as the nulling capability of the antenna
to maximize SINR at the receiver. This gives us the ability
to develop a simple protocol (DOA-MAC) that performs very
well.
The remainder of the paper is organized as follows: in the
next section we describe our system model and provide a
brief overview of adaptive array antennas. Section III describes
related work. Section IV illustrates our protocol DOA-MAC
in more detail. Section V presents results of the simulation.
II. SYSTEM MODEL
We assume that each node is equipped with an adaptive
array antenna system which is composed of a linear array of
elements. For simplicity, we assume that the antenna array
is perpendicular to the x-y plane in which the nodes lie. The
reason for this assumption is that the beam formed by the
antenna is symmetric about the antenna axis and is thus inde-
pendent of the direction in which a node is “facing”. Figure
1 provides a schematic of an adaptive array antenna system.
As illustrated in the figure, the antenna consists of
elements separated from each other by a known distance
can assume that a transmitter is located far enough away from
the receiver that all the signals
antenna elements are parallel. However, since the elements are
separated by distance
different. Let
each signal
can be written as,
?
?
antenna
? . We
????????? arriving at the different
? , the phase of the different signals is
?
? denote the phase and gain that is added to
?
?
????? . Then
??????? , the output sent to the receiver,
???????????
?
?
?????
?????????????????
is the phase propagation factor,
is an arbitrary gain constant. The weights
?
?
?????
??????? ?????"!$#&%('
?*)&+-,?.0/
where
wavelength, and
12?4365?7688
is the
?

Page 2

??? used in this paper only shift the phase of the signal and
leave the amplitude untouched. The representation for the
weights is,
????!
%('
?*)
?
???
/??
For a more comprehensive discussion, please see [4].
As we noted in the introduction, a beneficial feature of
adaptive array antennas is the ability of these antennas to form
nulls in given directions. In fact, given
can form upto nulls. However, the shape of the desired
beam can change dependingon the number of and the direction
of the nulls. Figure 2 illustrates two cases when using
antenna elements with
In the first case, we are forming only two nulls whereas in
the second case we are forming six nulls. As can be seen, the
shape of the beam and side lobes changes. In this work we
are using MMSE (Minimum Mean Square Error) algorithm to
determine weights
We implemented the adaptive array antenna model in MAT-
LAB and interfaced it with the physical layer of OPNET. In
our study we use realistic antenna patterns with the side lobes.
?
elements, an antenna
?????
?
???
?
??????? being the desired direction.
?
? to form nulls appropriately [4].
w
w
w
w
1
2
3
M
Σ
Receiver
S1(t)
S2(t)
S3(t)
SM(t)
d
Variable gain and
phase shifters
Antenna elements
Signal received from
transmitter at each
antenna element
θ
Fig. 1.Schematic of a adaptive array antenna.
III. RELATED WORK
[6] focuses on design the design of Smart antennas for
mobile devices with operating frequency of 20 Ghz. The
authors examine the impact of the antenna design on network
throughput and the impact of mutual coupling2on the perfor-
mance of adaptive algorithms. The paper presents results of
detailed OPNET simulations using a TDMA version of the
802.11 protocol as the MAC layer.
[7] presents a scheduling-based MAC protocols for nodes
equipped with directional antennas. The directional antenna
considered is a multi-beam adaptive array antenna (MBAA)
which is capable of forming multiple beams. The key contri-
bution of the paper is the development of a neighbor tracking
scheme that is then used to schedule transmissions by each
node in a distributed way.
[1] proposes a MAC protocol where nodes are equipped
with
directional antenna elements. Each of the antenna
elements has a conical pattern, spanning an angle of
?
365?7
?
2Mutual coupling results in radiation patterns that have shallower and
shifted nulls, and less accurate AOA, thus deteriorating overall network
throughput.
0.2
0.4
0.6
0.8
1
30
210
60
240
90
270
120
300
150
330
1800
Desired = 45 deg, Nulls = 25, 70 deg
0.2
0.4
0.6
0.8
1
30
210
60
240
90
270
120
300
150
330
1800
Desired = 45 deg, Nulls = 10, 20, 30, 70, 80, 90 deg
Fig. 2.Antenna patters with 8 antenna elements and 2 or 6 nulls.
radians. The
overlapping beam directions so as to collectively span the
entire plane. The MAC protocol is assumed to be capable
of switching any one or all the antennas to active or passive
modes. The paper didnot examine the benefits of nullingor the
impact of side-lobe interference. Furthermore, the propagation
model was rather simplistic because of the assumption of
complete attenuation outside the conical pattern.
[5] studies MAC protocols for Ad hoc network for nodes
equipped with directional antennas where the main beam is
modeled as a cone and the sidelobes as a sphere. They develop
a novel multi-hop RTS to establish links between distant nodes
and
(omni directional gain). The direction in which the main lobe
is to be oriented is determined by the MAC protocol (which in
turn is provided this information by the network layer which
is assumed to be neighbor-aware). Beamwidth of the antenna
is assumed to be a constant. They show that their protocol has
a 4-5x throughput as compared with 802.11.
In [2] a node in promiscuous mode caches AOA information
based on signals received and uses this informationfor sending
RTS. A circular antenna with 6 elements is assumed, and a
node is capable of electronically steering the boresight towards
a specific direction. A constant beamwidth of 45 deg assumed.
?
antennas at each node are fixed with non-
?
) ( directional gain) is assumed to be higher than
?
?

Page 3

However, it was observed that as the boresight changes the
side lobe pattern changes drastically. They also observe that
using realistic antenna patterns as opposed to a ideal patterns
results in a 36% degradation of throughput.
There have been several other papers that look at the
benefits of using smart antennas in cellular environments see,
for instance, [8], [9], [10], [11], [12]. These papers look at
models where the base station is equipped with one or multiple
adaptive antenna arrays.
IV. PROTOCOL DESCRIPTION: DOA-MAC
In this section we describe the behavior of our protocol.
However, before doing this, we need to make the following
assumptions: (1) We assume that nodes are aware of the
angularlocationof each of theirneighbors (as in[5]) since this
information is needed at transmitters to form directed beams
towards receivers; and (2) For simplicity, we assume that all
nodes use the same constant transmit power.
Consider the case when a node
to node which is its one-hop neighbor. Since
? needs to transmit a packet
?
?
knows
?’s angular direction it can form a directed beam towards
However, in order to maximize SINR at
beam towards
In order to do this,needs to know two things – first, that
attempting to transmit to it, and second, the angular direction
of all the other transmitters that interfere at
is based on the slotted ALOHA model with the addition of a
component that enables receivers to form beams and nulls as
described.
Each slot in DOA-MAC is broken into three minislots. The
protocol then works as follows:
1) The first minislot in a slot is called the DOA-minislot
and it is here that a node identifies the angular direction
of all transmitters that it can hear. All transmitters
transmit a simple tone (i.e., a sine wave) during the
DOA-minislot towards their intended receivers. The
signal received at some receiver is thus the complex
sum of all of these tones. The receiver runs a DOA
algorithm to determine the angular direction of each
of the transmitters and the received power from each
transmitter.
There are several different DOA algorithms that can
be used and the primary difference between them is
fidelity versus computational complexity. For this work
we chose to use MUSIC (MUltipleSIgnal Classification)
[4] which lies somewhere in between all the other
algorithms in terms of complexity and fidelity. Figure
3 shows an example of running MUSIC at the receiver
when there are three transmitters (all using the same
transmit power). As we can see, nodes
to one another in angle w.r.t. the receiver whereas
quite distinct.The output of the MUSIC algorithmshows
that the receiver is unable to distinguish between
because they are close in angle. This can result in a
higher SINR at the receiver because the receiver’s beam
could include both transmitters.
?.
?,
?
needs to form a
? and form nulls towards all other transmitters.
?
? is
?. Our protocol
? and
?
are close
? is
? and
?
30
o
10
o
30
o
a
b
c
Receiver
020406080
Angle (deg)
100120140160180
−6
−5
−4
−3
−2
−1
0
Power (in dB)
MUSIC Spatial Spectrum
a, b
c
Fig. 3.Example of computing DOA using MUSIC.
2) Once a receiver determines the DOA of all transmitters
it can hear, it forms its directed beam towards the one
that has the maximum power and forms nulls in all the
other identified directions.
3) The second (and largest) minislot is the packet trans-
mission slot and it is here that the packets are trans-
mitted. After the receiver has formed its beam and
nulls as described above, it receives the packet from
the transmitter. After receiving the packet, it looks at
the header and rejects the packet if it was not the
intended destination. An example of this happening is
illustrated in Figure 4 where we see that nodes
are transmitting to nodes
node incorrectly chooses to receive
because that transmission is stronger!
? and
?
? and
? respectively. However,
?? ’s transmission
a
b
d
c
Node d mistakenly forms
a beam towards a because
a’s signal is stronger
than b’s signal at d
a has a packet for c
b has a packet for d
Fig. 4.An example of incorrect beamforming leading to a rejected packet.
4) The last minislot is the ACK slot where the receiver
transmits an ACK using the already formed beam to
the sender (if the packet was not rejected and correctly
received). In Figure 4 node
nodebecause it did not receive
but rather mistakenly received
5) When a transmitter does not receive an ACK, it retrans-
mits the packet at a later time (as in slotted ALOHA).
? will not send an ACK to
? or to node
??’s packet
? ’s packet.

Page 4

TABLE I
OPNET SIMULATION PARAMETERS.
Simulation Parameters
Background Noise + ambient Noise
Propagation model
Bandwidth
Min frequency
Data Rate
Packet Size
Packet Generation
Carrier Sensing Threshold
Minimum SINR
Bit Error
Maximum radio range
-143 dB
Free space
1,000 kHz
2.4 GHz
2000 kbps
512 bytes
CBR
+3dB
9 dB
Based on BPSK Modulation curve
250 m
V. PERFORMANCE STUDY
In our simulation study, we examined three questions:
Does increasing the number of antenna elements improve
throughput?
What is the impact of using a realistic DOA algorithm
(MUSIC) as opposed to an optimal algorithm with an
arbitrary resolution (i.e., in Figure 3 it would correctly
discriminate between nodes a and b)?
Does nulling have any benefits?
The simulationparameters we selected are displayedin table
I. We evaluate the performance of DOA-ALOHA using a 5x5
mesh (as used in [5]) with four pre-defined flows. Figure 5
shows the network topology and flows used for two of these
experiments. For the third experiment, we used a random node
placement on the grid where a node’s position is shifted in the
x-axis and y-axis by adding a displacement randomly selected
from [-150m , +150m] and the flows are as in Figure 5(b).
The traffic is CBR (Constant Bit Rate) which increases (per
flow) from 75kbps to 2Mbps. The packet size is 512 bytes.
In order to examine the impact of the number of antenna
elements on throughput, we plot the aggregate throughput as
a function of data rate of one flow, for the case shown in
Figure 5(a), in Figure 6. We plot the same data for the random
topology case in Figure 7 as well. As we can see, using 16
antenna elements as opposed to 8 does improve throughput in
both cases. This result in not surprising because larger number
of antenna elements results in narrower beams and hence better
spatial reuse. Interestingly, the throughput is higher for the
random topology case when compared with Figure 5(a). This
is because, in Figure 5(a), the flows are aligned and need to
share bandwidth at the second hop whereas in the random
topology case, there is greater potential for spatial reuse since
flows are not aligned.
In order to determine the impact of using MUSIC instead
of optimal DOA and to answer the question about the benefits
of nulling, we focus on the case shown in Figure 5(b). Figure
8 plots the aggregate throughput as a function of data rate of
a single connection for the case when we have 16 antenna
elements (Figure 9 does the same when we use 8 antenna
elements).
We observe that using 16 antenna elements as opposed to
8 elements makes a big difference in aggregate throughput.
?
?
?
(a) Four flows (some alignment)
(b) Randomly selected flows
Fig. 5.5x5 grid topology used to compare performance with [5].
This is because the beamwidth when using 16 elements is
smaller than when using 8 elements which results in more
simultaneous transmissions/slot. For the flows in Figure 5(a),
(when flows are aligned), we did not notice much difference
in the performance of 16 and 8 antenna elements but for
Figure 5(b) and for random topologies we do see a significant
difference. The reason is that when flows are not aligned, there
is a greater potential for spatial reuse with 16 antenna elements
(due to its smaller beamwidth).
0200400600800
Sending Rate (Kbps)
100012001400160018002000
0
500
1000
1500
2000
2500
3000
Sending rate (Tx) vs Aggregate Throughput
Aggregate Throughput (Kbps)
8 elements
16 elements
Fig. 6. Performance of our protocol with optimal DOA in 5(a).
We notice that MUSIC results in poor performance (approx.
8% reduction in throughput with 8 elemenst and 5% reduction
with 16 elements) in comparison to optimal DOA, this is for
the reason as explained in Figure 3 and 4 that MUSIC fails to
distinguish between two nodes located in close proximity. A
high-resolutionalgorithm will be able to discriminate between
these two nodes, however, it will require more training se-
quences and its computationcost will be high. We are currently
studyingthe trade-off between DOA algorithm complexity and
its computation cost. Finally, we observe that Nulling has
greater impact on 8 elements (improvement of approx. 11%)
than 16 elements (improvement of approx. 5%).
Table II summarizes our results and compares them with [5].
We provide the maximum throughput when using the optimal
DOA as well as when using MUSIC. We observe that our
protocol is 2x – 3x better when we use 8 elements and is much
better (3x – 4x) for 16 elements. We note that the beamwidth

Page 5

0200400600800
Sending Rate (Kbps)
100012001400 160018002000
0
500
1000
1500
2000
2500
3000
3500
4000
4500
Sending rate (Tx) vs Aggregate Throughput
Aggregate Throughput (Kbps)
8 elements
16 elements
Fig. 7.
topologies.
Performance of our protocol with optimal DOA in random grid
0200400600800
Sending Rate (Kbps)
100012001400160018002000
0
500
1000
1500
2000
2500
3000
3500
Sending rate (Tx) vs Aggregate Throughput
Aggregate Throughput (Kbps)
Optimal DOA
MUSIC
No Nulling
Fig. 8. Performance of our protocol with 16 elements in 5(b).
used in [5] is
symmetric beams and we define beamwidth for our protocol
as the sum of these two beams.
???
?. In our case, the linear array creates two
VI. CONCLUSION
In this paper we have presented DOA-MAC, a slotted MAC
that uses DOA information at the receiver to beamform in a
way that maximizes SINR. We notice that by exploiting the
benefits of smart antenna a simple protocol like DOA-MAC
can achieve very high throughput. We studied the impact of
using a realistic DOA algorithm as well as the benefits of
nulling. Finally, we compare the performance of our protocol
against [5] and show that our protocol has a throughput of 2x
– 4x higher than the [5].
REFERENCES
[1] J. Y. A. Nasipuri, S. Ye and R. Hiromoto, “A mac protocol for mobile
ad hoc networks using directional antennas,” in IEEE WCNC, 2000.
[2] R. B. Mineo Takai, Jay Martin and A. Ren, “Directional virtual car-
rier sensing for directional antennas in mobile ad hoc networks,” in
ACM/SIGMOBILE MobiHoc 2002, Oct 2002.
0200400600800
Sending Rate (Kbps)
100012001400160018002000
0
500
1000
1500
2000
2500
Sending rate (Tx) vs Aggregate Throughput
Aggregate Throughput (Kbps)
Optimal DOA
MUSIC
No Nulling
Fig. 9.Performance of our protocol with 8 elements in 5(b).
TABLE II
Mesh Figure 5(a)
16 Elements
(
2500kbps
800kbps
Mesh Figure 5(b)
16 Elements
3100kbps
2940kbps
1000kbps
Random Mesh
16 Elements
4300kbps
1000kbps
8 Elements
(
2200
?
?????)
???
???)
Our Protocol (optimal DOA)
[5]
8 Elements
2065
1900
Our Protocol (optimal DOA)
Our Protocol (MUSIC)
[5]
8 Elements
3700 Our Protocol (optimal DOA)
[5]
[3] V. S. Y.B. Ko and N. Vaidya, “Medium access control protocols using
directional antennas in ad hoc networks,” in IEEE INFOCOM 2000,
March 2000.
[4] J. C. Liberti and T. S. Rappaport, Smart Antennas for Wireless Commu-
nications.Prentice Hall, 1999.
[5] R. R. N. H. V. Romit Roy Choudhury, Xue Yang, “Using direc-
tional antennas for medium access control in ad hoc networks,” in
ACM/SIGMOBILE MobiCom 2002, 23 – 28 Sep 2002.
[6] S. Bellofiore, J. Foutz, R. Govindarajula,I. Bahceci, C. A. Balanis, A. S.
Spanias, J. Capone, and T. M. Duman, “Smart antenna system analysis,
integration, and performance for mobile ad-hoc networks (manets),”
IEEE Transaction on Antennas and Propagation, vol. 50, no. 5, pp.
571 – 581, May 2002.
[7] J. G.-L.-A. Lichun Bao, “Transmission scheduling in ad hoc networks
with directional antennas,” in ACM/SIGMOBILE MobiCom 2002, 23 –
28 Sep 2002.
[8] P. S. A. Chockalingam, “Performance analysis of space diversity s-aloha
with steerable beam smart antenna arrays,” in IEEE GLOBECOM ’00,
Nov 2000.
[9] J. Ward and J. R. T. Compton, “Improving the performance of a slotted
aloha packet radio network with an adaptive array,” IEEE Transactions
on Communications, vol. 40, no. 2, pp. 292 – 300, February 1992.
[10] A. S. Acampora, S. Krishnamurthy, and M. Zorzi, “Media access
protocols for use with smart array antennas to enable wireless multime-
dia applications,” in Proceedings 9th Tyrrhenian Workshop on Digital
Communications, September 1997.
[11] C. Sakr and T. Todd, “Carrier-sense protocols for packet-switched smart
antenna basestation,” in ICNP’97, 1997.
[12] R. R. A. Chockalingam, “Mac layer performance with steerable multi-
beam antenna arrays,” in IEEE PIMRC ’98, 1998.