Ultra-Wideband Interference Mitigation Using
Cross-layer Cognitive Radio
Rashid A. Saeed, Sabira Khatun., Borhanuddin Mohd. Ali, Mohd. Khazani Abdullah
Department of Computer and Communications Systems Engineering
Engineering Faculty, Universiti Putra Malaysia,
43400 UPM, Serdang, Selangor, Malaysia
Abstract. Cognitive Radio (CR) is an emerging approach for a more efficient
usage of the precious radio spectrum resources, which it considers an expanded
view of the wireless channel by managing and adapting various dimensions of
time, frequency, space, power, and coding. In this paper, we define the system
requirements for cognitive radio, as well as the general architecture and basic
physical and link layer functions. In order to self-adapt the UWB pulse shape
parameters and maximize system capacity while co-exist with in-band legacy
NB systems (WiFi and UMTS) in the surrounding environments.
1 Introduction
Ultra-Wideband (UWB) technology is one of the promising solutions for next-
generation wireless communications in multimedia-rich environments. UWB systems
spread the transmitted signal power over an extremely large frequency band, and the
power spectral density (PSD) of the signal is very low. Due to the wide bandwidth of
the transmitted signal, UWB signal energy will spread over the frequency bands
allocated to other radio systems.
The idea behind Cognitive radio is that performance can be improved and interference
reduced if wireless systems were aware of other RF signals in their environment. The
improvements accrued from this technology could be dramatic, while communication
engineers have historically though of channel capacity and Shannon’s law simply in
terms of bandwidth, a cognitive radio takes an expanded view of the channel by
managing and adapting time, frequency, space, power, and coding parameters.
Cognitive radios can ascertain their location and adapt to real time conditions of their
operating wireless environment, including the ability to sense spectrum usage by
neighboring devices, change operating frequency, adjust output power and even alter
transmission parameters and characteristics [1], [2].
Recent studies have shown that UWB signals can significantly affect the operation of
underlying narrowband systems [3], [4]. The coexistence of power-controlled ultra-
wideband systems with UMTS, GPS, DCS1800, and fixed wireless systems was
presented in [3]. In [4], the authors show that UWB devices operating at the peak
allowable power can significantly impact the achievable signal to noise ratio of
802.11a WLAN and UMTS clients in modified Saleh-Valenzuela channel model. In
[5] aggregate UWB interference to UMTS link is calculated as a function of distance
between the communicating devices, also, the realistic channel model is taken into
A. Saeed R., Khatun S., Mohd. Ali B. and Khazani Abdullah M. (2006).
Ultra-Wideband Interference Mitigation Using Cross-layer Cognitive Radio.
In Proceedings of the 5th International Workshop on Wireless Information Systems, pages 76-85
Copyright
c
 SciTePress

account. The calculation is that interference to noise ratio of less than -6dB to -10dB
is achieved.
The term Cognitive Radio was first defined by Mitola [5]. General interference
mitigation methods, which are not limited to UWB only, are presented in [6] and [7].
In [6], collaborative and non-collaborative coexistence mechanisms are proposed. A
Cognitive Radio approach for usage of virtual unlicensed spectrum based on spectrum
pooling idea is developed by Danijela [7]. In [8] a review of possible scenarios is
addressed, where the UWB technology is proposed as the natural platform for CR. A
briefly present for a possible Cognitive Radio network analysis tool based on a game-
theoretic approaches been discussed in [9].
In this paper, we define a cognitive radio system model for UWB in-band interference
(IBI based on new FCC waiver grant. For our well known no previous work discussed
coexistence based on new FCC waiver rule. Also expressions for adaptive pulse
shape to counteract IBI and at the same time to grantee low spectral emission over
continuous background NB transmissions.
This paper is organized as follows. Section II describes FCC rules for UWB, section
III describe our proposed system model for UWB cognitive radio, Numerical results
in section IV, and conclusions are given in section V.
2 FCC Rules for UWB-Coexistence
In April 2002 the Federal Communications Commission (FCC) released UWB
emission masks and introduces the concept of coexistence with traditional and
protected radio services in the frequency spectrum, which allows the operation of
UWB systems mainly in the 3.1 to 10.6 GHz band, limiting the power level emission
to -41dBm/MHz.
In March 2005, the FCC granted the waiver request filed by the MBOA [10]. Which
it is approved the change in measurement for the all UWB technologies (neutral
approach). The FCC’s waiver grants effectively removes the previous transmit power
penalties for both frequency-hopping and gated UWB technologies (TH and DS),
which it can transmit at higher power levels and then sit quiet, as long as they still
meet the same limits for average power density. Table I shows the FCC waiver new
rules.
Table I. Pre- and Post-FCC Waiver Ruling Effects.
Power Requirement Method of Measurement
Pre-Waiver
Ruling
-41 dBm/MHz Power measured in always-on
mode
Post
Waiver
Ruling
-41 dBm/MHz Only average power measured;
systems now allowed to burst and
then sit quiet
Effect No change; spectrum holders still
protected from interference
Both FH and gated UWB systems
benefit
77

3 System Model
A cognitive radio (CR) is a radio that can change its transmitter parameters based on
interaction with the environment in which it operates. This interaction may involve
active negotiation or communications with other spectrum users (spectrum sensing),
decision making (spectrum adaptation) within the radio, and share these information
within the network members (Co-operation).
As shown in Figure 1,
System starts with discovery the wireless channel so as to
sense the available spectrum resources. Here we consider two types of interference to
UWB receiver; co-existence systems interference (NBI) and UWB to UWB
interference. There are two types of co-existence interference, background
interference and burst interference. The signals from GSM base-stations and WLAN
access points are considered as background noise since they are transmitting almost
continuously, which can be mitigated by using a high-pass filter or antennas, whose
frequency transfer functions show strong out-of-band attenuation and use new FCC
waiver rules. The term burst interference is used for narrow band interferences (NBIs)
that transmit their data burst-wise, as, e.g., WLAN and UMTS-FDD nodes (shown in
Table II).
Fig. 1. Cognitive Radio system model.
Jamming signal
Spectrum
Sensin
g
Co-existence UWB
Burst Background
FCC
Limitation
Spectrum
adaptation
Contention
Protocol
Send
Discovery mode
Interaction
mode
Adaptation
mode
Broadcast co-
existence info
Co-operation mode
78

Table II. In-band burst interference Narrowband Systems.
IEEE 802.11n WLAN UMTS-FDD Downlink
In-band NBI (GHz)
2412-2472
2.11 – 2.17
PHY Technique OFDM and MC-CDMA WCDMA
Modulation 16-QAM QPSK
Maximum Power 20dBm +33 dBm
Symbol Rate (
s
f )
16.25Msps 3.84 Mcps
Data Rate 100Mbps 384 kbps
Channel Coding convolutional code convolutional code
Pulse filtering Root raised cosine, roll-off =
0.23
Root raised cosine, roll-off
= 0.22
3.1 Spectrum Sensing
Spectrum sensing is best addressed as a cross-layer design problem since sensitivity
can be improved by enhancing radio RF front-end sensitivity, exploiting digital signal
processing gain (matched filter) for specific NB signal, and network cooperation
where users share their spectrum sensing measurements [9].
In next section we discuss the effect of NB receiver filters and their bit error rate
3.1.1 Effect of NB Receive Filter
In a typical NB radio, using digital modulation, the same symmetric baseband pulse-
shape is used at both the transmitter and the receiver, namely the root-raised cosine
filter (RRCF) with transfer function
)( fH
rrcf
with nominal bandwidth W, roll-off
factor or excess bandwidth parameter denoted by
Ω
)10(
≤
Ω
≤
, and overall
bandwidth
)1( Ω+W [11].
Taking into account the fact that the front-end band-pass filter and the low-pass filter
following the demodulator typically have larger bandwidth than the root-raised cosine
filter (RRCF) transfer function
)( fH
rrcf
, we note the overall filter for the UWB
impulse-train is given by,
[]
)()(
)()()(5.0)(
fHfP
fHffPffPfG
rrcf
rrcfcc
=
++−=
(1)
where
)( fP is the FT of the Gaussian monocycle, and
c
f is the carrier frequency for
the NB system, where for UMTS RRC roll-off=0.22
3.1.2 BER Analysis
Let
)(tr denote the received signal, after down-conversion to baseband. Consider the
usual matched filter (MF) plus threshold receiver for BPSK signaling, which is
optimal for the AWGN channel. In the k-th symbol interval,
)(tr can be written as:
)()()()( titntsbEtr
kb
++=
(2)
79

where )(ts is the unit energy signal waveform with duration T (corresponding to
)( fH
rrcf
),
b
E is the energy per bit,
k
b {-1, 1} is the unknown bit, )(tn is AWGN
with two-sided PSD No/2, and
)(ti is the interference. Applying the MF, we have
[]
)(
~
)()()()(
0
ςυ ++=
++=
∫
nbE
dttstitntsbEz
kb
T
kb
(3)
Here
n
~
represents the 'noise' term and is zero-mean Gaussian, with variance No/2.
The term
)(
ς
υ
represents the interference. Assuming that the interfering pulse )(ti is
completely contained within the symbol period and has a relative delay of
ς
, we have
∫∫
∞
∞−
−== dfefSfPEdttsti
fj
p
T
ςπ
ςυ
2
0
)()()()()(
(4)
b
E is the energy in the received UWB pulse. For an NB )(FS , )( fP is essentially
constant over the bandwidth of
)(FS , so that
)()()( ςςυ sfPE
cp
≈
(5)
Define the SNR impairment factor,
)()( ςδ sfP
E
E
c
B
p
=
(6)
Matched filter effectively requires priori knowledge of primary user signal at both
PHY and MAC layers, e.g. modulation type and order, pulse shaping, packet format.
However, since we work only in in-band interference for UWB (e.g., WLAN and
UMTS). Such information might be pre-stored in UWB node memory. To achieve
coherency with NB signal by performing timing and carrier synchronization, even
channel equalization. This is still possible since most NB systems have pilots (e.g.,
UMTS), preambles synchronization words (e.g., WLAN) or spreading codes that can
be used for coherent detection.
3.3 Spectrum Adaptation
There are some UWB pulse waveform candidates (i.e. Hermits, PSWF…etc). In our
simulation we use Gaussian doublet with a pair of separated narrow monocycles, a
positive pulse followed by a negative pulse. This offers two degrees of time-cognitive
radio freedom, time separation between the two pulses in the doublet (
n
T ) and
doublet pulse time length (
p
T
).
The Gaussian pulse waveform
)(tp can be expressed as:
⎥
⎥
⎦
⎤
⎢
⎢
⎣
⎡
⎟
⎠
⎞
⎜
⎝
⎛
−
−=
2
2
1
exp
2
)(
σ
μ
σπ
tA
tp
(7)
80

Where μ is the mean and
σ
is the variance of the statistical distribution characterized
by the function. Since infinite pulses cannot be used in practical implementation and
leading to unavoidable overlap between pulses, and ISI. We can introduce the time
length of the pulse
p
T
, related to σ through the linear transformation
π
σ 2=
p
T
(8)
Which pulse is nulled outside the interval
{
}
2,2
pp
TT
−
and the factor A is
introduced so that the total energy of the monocycle is normalized to unity,
i.e.
∫
= 1)(
2
dttp , in the simulations the transmitting and receiving antennas are
modeled as differentiation (derivative) operations [12].
The null frequencies of Gaussian doublet can be controlled by regulating the position
of the Gaussian pulse (the first pulse
)(
0
tp begins at t = 0, the second one
tp (
1 begins
at
n
Tt = ). With spectrum analysis, we have:
)()()(
10
tptptp −=
(9)
⎪
⎭
⎪
⎬
⎫
⎪
⎩
⎪
⎨
⎧
⎥
⎥
⎦
⎤
⎢
⎢
⎣
⎡
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
−−
−−
⎥
⎥
⎦
⎤
⎢
⎢
⎣
⎡
⎟
⎠
⎞
⎜
⎝
⎛
−
−=
2
2
2
1
exp
2
1
exp)(
σ
μ
σ
μ
n
Tt
t
Atp
(10)
Which has a Fourier transform given by
[] {}
πμπ
πσπσπ
5.0)5.0(2exp
)2(
2
1
exp)sin(22)(
2
−+−
⎥
⎦
⎤
⎢
⎣
⎡
−
′
=
n
n
Tfj
ffTAfp
(11)
For this simulation, the Monte Carlo simulation is considered, which an independent
normal variable is generated for each sample, with variance (
smp
var
) given by
0
0
2
var
NE
A
b
N
i
i
smp
∑
=
=
(12)
where N is the number of samples in a pulse,
i
A is the amplitude of a sample of the
pulse, and E
b
/N
0
is the average SNR measured, which it meet the new FCC waiver
rules. These samples are then processed, and the BER estimate is formed by taking
the ratio of the number of errors in the received bits to the total number of transmitted
bits.
Conceptually, a pulse-train (which may be dithered to reduce spectral lines, to
accommodate user codes, to represent data via PPM, and so on) is convolved with the
pulse-shape, so that the power spectrum of the transmitted UWB signal is essentially
given by |P (f)|
2
. The effect of dithering makes the modulation zero-mean and
81

removes periodicities in the data, so that the PSD has no spectral lines. The
monocycle pulse
)(
0
tp with the pulse shape factor
σ
of 0.6 nsec and its doublet
unmodulated pulse in the time domain with time separation of 0.7 nsec were shown
in Figure 1, where Figure 3 shows the 'PSD' |P(f)|
2
, which depicted that there are
spectral lines in the doublet waveform which cause interference to other co-existence
systems.
-1 -0.5 0 0.5 1
0
0.2
0.4
0.6
0.8
1
t nsec
Normali zed ampl i tude
-1 0 1 2
0
0.2
0.4
0.6
0.8
1
t nsec
Normali zed ampl i tude
Fig. 2. Gaussian pulse in the time domain and its Gaussian doublets.
3.2 Cooperative Spectrum Sensing
Here we rely on the variability of signal strength at various locations, which the
performance of matched filter technique is limited by received signal strength which
may be severely degraded due to multipath fading and shadowing and due to hidden
terminal problem and UWB nodes in quite-mode can receive sensing information,
which can be used in on-mode. In large network of UWB nodes with sensing
information exchanged between neighbors would have a better chance of detecting
the NB bursts compared to only individual sensing.
4 Numerical Results
The design procedure detailed in the following sections optimizes the PSD expression
of (9) to minimize UWB interference into a given narrowband systems. This approach
is justified because the utilized pulse shape is only capable of changing this part of the
PSD, which including a spectral notch in the transmitted PSD. The simulation was
done using MATLAB
®
.
4.1 Waveform PSD Smoothing
In our simulation, we assume a monocycle pulse )(
0
tp with
σ
=318.31 ps, producing
a 200 ps wide pulse. The methodology for choosing
σ
to produce a signal that is
compliance with the FCC spectral mask is discussed in detail in [13] and thus, is
omitted here.
82

The term )sin(
n
fT
π
in equation (9) represent the discrete component (spectral lines)
that shown in Figure 2, for deduce spectral lines power, one can adjust the time
between the pulses (
n
T ) so as the term )sin(
n
fT
π
=1, as can be seen in Figure 4.
4 6 8 10 12 14
x 10
9
10
-30
10
-20
10
-10
10
0
Frequency domain
Frequency [Hz]
Normalized ESD
N=1024
T
n
=1nsec
Fig. 3. Energy Spectral Density for Gaussian doublet pulse.
4.2 Effect of UWB Interference
This section introduces the effect of UWB spectrum interference to the narrowband
systems performance degradation. Error probability (
e
P ) curves are presented for
both victim systems (WLAN and UMTS-FDD) utilizing UWB Gaussian doublet
waveform shown in Figure 2.
4 6 8 10 12 14
x 10
9
10
-30
10
-25
10
-20
10
-15
10
-10
10
-5
10
0
Frequency domain
Frequency [Hz ]
Normalized ESD
N=1024
T
n
=1ns
Fig. 4. The effect of adjust the time between pulses (
n
T ).
In Figure 5 the effect of different signal to interference ratio (SIR) are studied in
presence of UWB in-band interference. The pulse length (
p
T ) are varied from 0-
5nsec, the average signal to noise ratio is SNR=15dB. The UWB PSD is evenly
distributed inside the considered WLAN/UMTS-FDD bands. From results one can
notice that WLAN system degrades more than UMTS in presence of UWB spectrum,
83

which error probability is higher for smaller SIR and higher pulse length (
p
T
). The
difference in NB systems performance is based on the different spectrum location at
the UWB signal related to these systems. In this case study, two UWB monocycles
were produce doublet waveform with
n
T =1.1nsec, which generates spectrum that
overlaps the victim systems spectrums.
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
10
-10
10
-5
10
0
T
p
nsec
P
e
SIR=0.5dB
SIR=1.0dB
SIR=1.5dB
blue -->W LAN
red -->UMTS-FDD
Fig. 5. probability of error (
e
P ) for UWB co-existence interference WLAN (IEEE802.11n)
and UMTS-FDD uplink for different SIR values.
4.3 Spectrum Adaptation
As we want the ESD of UWB interference is as small as possible at the victim
narrowband system’s band, the null frequencies of Gaussian doublet pulse’s ESD is
preferred to be positioned in victim system’s band. For deep interference as caused in
IEEE802.11a and UMTS-FDD center frequencies the proposed cognitive radio
system will respond with notch in the victim frequencies by adapt the number of
samples (N) and pulse length (
p
T
) as can be seen in Figure 6.
2 4 6 8 10 12
x 10
9
10
-20
10
-15
10
-10
10
-5
10
0
Frequency domain
Frequency [Hz]
Normalized ESD
Solid
N=2048
T
p
=0.045 ns
Dots
N=1850
T
p
=0.04 ns
Fig. 6. Normalized ESD for two type of UWB doublet pulse waveform design to mitigate the
U-NII center frequency for IEEE802.11n (solid line) and UMTS-FDD uplink center frequency
(dots line).
84

5 Conclusions
The paper introduces the concept of cognitive radio for IBI UWB, which wireless
systems based on UWB transmission able to self-adapt to the characteristics of the
surrounding in-band systems (e.g., unlicensed WLAN 5.3GHz ISM (industrial,
scientific, and medical) band and UMTS). Cognitive radio sensitivity can be
improved by enhancing radio RF front-end sensitivity, exploiting UWB digital signal
processing gain for specific narrowband signal, and network cooperation where users
share their spectrum sensing measurements.
For future work the other factors such as code and space (distance) cognitive radio
can be considered.
References
1. Joseph Mitola III. Cognitive radio for flexible mobile multimedia communications. In Sixth
International Workshop on Mobile Multimedia Communications (MoMuC’99), San Diego,
CA, 1999.
2. J. Lansford. “UWB Coexistence and cognitive Radios”, Pg 35-39 alleviate spectrum
crowding proceedings Ultra-Wideband Systems, May 2004.
3. R. Giuliano and F. Mazzenga, “On the coexistence of power-controlled ultra-wideband
systems with UMTS, GPS, DCS1800, and fixed wireless systems,” IEEE Trans. on
Vehicular Technology, vol. 54, pp. 62–81, January 2005.
4. M. Hamalainen, R. Tesi, and J. Iinatti, “UWB co-existence with IEEE802.11a and UMTS
in modified Saleh-Valenzuela channel,” 2004 Joint UWBST & IWUWBS, pp. 45–49, May
2004.
5. Joseph Mitola III. Cognitive Radio an Integrated Agent Architecture for Software Defined
Radio. PhD thesis, KTH Royal Institute of Technology, Stockholm, Sweden, 2000.
6. R. W. Brodersen, A. Wolisz, D. Cabric, S. M. Mishra, and D. Willkomm, “CORVUS: A
cognitive radio approach for usage of virtual unlicensed spectrum,” University of
California, Berkley, CA, White Paper, July 29, 2004.
7. D. Čabrić1, S. Mubaraq, D. Willkomm, R. Brodersen1, A. Wolisz “A Cognitive Radio
Approach for Usage of Virtual Unlicensed Spectrum “2005 Joint UWBST & IWUWBS,
pp. 50–54, May 2005.
8. T. Zasowski, F. Althaus, and A. Wittneben “Temporal Cognitive UWB Medium Access in
the Presence of Multiple Strong Signal Interferers” IEEE UWBST 2005, pp. 285–289, Aug.
2005.
9. F. Granelli, H. Zhang, “Cognitive Ultra-wide band radio: a research vision and it is open
challenges” IEEE Spectrum, vol. 41, no. 3, pp. 48-52, March 2004.
10. FCC, “Revision of part 15 of the commission’s rules regarding ultra-wideband transmission
systems,” First Report and Order, ET Docket 98-153, FCC 02-48, adopted/released Feb.
14/ Apr. 22 2002.
11. Proakis, J.G. Digital Communications. McGraw Hill Inc, 2001, chapter 6, pp 333-372.
12. “Ultra-Wideband Opportunities under the New FCC Waiver” Freescale Semiconductor,
Freescale Semiconductor, Document Number: UWBOPPFCCWVRWP, Rev 1,
March/2005
13. H. Shing, P. Orlik, A. M. Haimovich, L. Cimini, and Z. J., “on the spectral and power
requirements for Ultra-Wideband transmission”, in IEEE International Conference on
communications, (Anchorage, AK), 2003.
85