IEEE TRANSACTIONS ON BROADCASTING, VOL. 59, NO.

4, DECEMBER 2013

685

Dynamic Gain Management for On-Channel Repeaters

Michael Mao Wang, Member, IEEE

AbstractStability and output signal quality are always two

major concerns with the deployment of on-channel repeaters.
This paper proposes a gain management scheme and evaluates
the performance of the proposed gain management technique for
on-channel repeaters. The gain manager uses a loop gain metric
as an indication of stability and output signal quality (SNR) of
the repeater to continuously and accurately monitor the repeater
status, and adjust the repeater gain accordingly. The repeater
output SNR and stability can thus be tightly controlled. It is
demonstrated that the proposed gain management ensures the
safe and robust operation of the repeater against the dynamics
of the environment.
Index TermsAmplify-and-forward relays, on-channel (onfrequency) repeaters, single frequency network, stability control.

I. Introduction

HE increasing spread of multimedia broadcasting services, such as DVB-T/H [1], [2], and FLO-TV [3], and
wireless WAN, such as EVDO and WCDMA, require efficient
and economic network technologies for successful delivery
[4], [13]. The purpose of an on-channel repeater [7], [6]
is to extend the coverage of the network in areas where
service quality is poor. Repeaters are an essential part of an
efficient and effective wireless communications system, providing enhanced coverage to isolated areas. A repeater may be
considered as an amplify-and-forward relay with no decoding
or scheduling ability. The on-channel or on-frequency repeater
receives the original signal over the air, amplifies it and then
re-transmits it on the same frequency/channel as received. In
this way, the repeater system improves reception in areas of
poor coverage, especially in homes and buildings, thereby
extending the coverage and achieving the hole filling purposes.
From a cost-benefit tradeoff perspective, the low complexity
of repeaters makes them an attractive low-cost alternative to
other expensive relays.
Sending and receiving concurrently on the same channel, the
effective gain of an on-channel repeater is limited primarily
by the amount of isolation between the transmitter and the
receiver antennas. There is a need for high isolation between
transmit and receive antennas to minimize coupling and to
keep the system stable. The isolation of the receive (donor)
antennas and the transmit (coverage) antennas of the repeater
is necessary but usually not sufficient to reduce the impact

Manuscript received April 26, 2013; revised July 27, 2013; accepted August
7, 2013. Date of publication October 21, 2013; date of current version
December 10, 2013.
The author is with Qualcomm Research, San Diego, CA 92121 USA
(e-mail: [email protected]).
Color versions of one or more of the figures in this paper are available
online at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TBC.2013.2284417

of unwanted echoes or the feedback channel. Digital echo

cancellation techniques can be used to compensate for the
lack of antenna isolation and further reduce the unwanted
echoes in the feedback path or leakage channel between the
transmit and receive antennas, thereby improving the isolation
figure [7][13]. However, the performance of these techniques
depends heavily on the dynamic multipath characteristics of
the feedback channels.
The ability to ensure repeater stability has been the critical
issue that is limiting the wide deployment of repeaters.
In addition, as the wireless market continues to expand,
the use of repeaters that potentially create interference on
carrier networks becomes a growing and pressing issue.
Conventional repeater products do not have a robust method
of oscillation prevention and detection methods and must be
carefully engineered and installed to avoid oscillations. Due
to the lack of stability control mechanism, the repeater status
in a dynamic wireless environment can hardly be guaranteed,
and when isolation/environment changes, a conventional
repeater can either distort the network signals or even
oscillate. Repeater oscillation can act as a jammer within the
network, blocking a channel or a number of channels within a
network; leading to costly trouble-shooting for the carrier, and
degraded services for subscribers. Therefore, stability control
via appropriate gain management is a highly desirable feature
for modern repeaters. Unfortunately, there are very few
publicly available repeater stability and output signal quality
control techniques in the literature [14], [15]. This paper
attempts to fill this gap with a gain management technique
that employs a loop gain metric to continuously monitor the
repeater stability and promptly adjust the gain accordingly
to ensure the repeater stability as well as the repeater output
signal quality. The goal of the design is to ensure that a
repeater is unconditionally stable, and to maintain output
signal quality at a desired level in any dynamic environment.
The remainder of the paper is organized as follows: In
Section II, we briefly review the architecture and components
of modern repeaters; Section III describes in detail the design
of the gain management in both the link level and the device
level perspectives. In Section IV, we evaluate gain management
performance. Finally, conclusions are given in Section V.

II. Repeater Overview

Fig. 1 illustrates a typical repeater architecture where a
repeater receives the signal over a wireless link using a
directional antenna, i.e., the donor antenna. The received signal
is amplified by the repeater, which is then transmitted using a
coverage antenna that may be directional or omnidirectional.

c 2013 IEEE
0018-9316 

686

IEEE TRANSACTIONS ON BROADCASTING, VOL. 59, NO. 4, DECEMBER 2013

Fig. 1. A typical repeater block diagram (black) and the proposed gain management module (red).

In between receiving its signal and re-transmitting, the repeater

introduces a delay of a few microseconds.
Ideally, we want a repeater to have a large gain
(e.g., > 70 dB), maintain the input SNR at the output, and have
robust boot-up and operation. However, the repeater gain is
limited by the isolation between the amplified signal and its
feedback path, which can be provided by two different mechanisms: isolation between the donor and coverage antennas
(passive isolation) and isolation by feedback echo cancellation (active isolation). The output SNR of a repeater mainly
depends on echo cancellation residual error, and practically
may also depend on RF noise figure. Repeater may oscillate
due to isolation changes as a result of degraded cancellation
performance or antenna isolation. Special mechanisms therefore must be in place to ensure the safe operation of a repeater.
A. Antennas
The main problem with on-channel repeaters is the effect of
strong feedback signals from the coverage (or transmit) antennas towards the donor (or receive) antennas. These feedback
signals can interfere with the weak received signal (i.e., the
remote signal) of the repeater causing oscillations and system
instability. This unwanted feedback signal is also known as
coupling loop interference and arises when the gain of the
amplifier is larger than the isolation between the receive and
transmit antennas.
A first solution to the coupling loop interference problem is
to increase the isolation of the transmit and receive antennas
of the repeater. Depending on the design (physical isolation
between the transmit and receive antennas of the repeater) and
the operating environment (reflections from the surrounding
environment), antenna isolation typically ranges from 30 to
60 dB. This degree of isolation is necessary but often insufficient for high gain repeaters.
B. Feedback/Echo Cancellation
Digital baseband echo cancellation is a commonly used
technique to effectively increase the isolation in modern
repeaters [17], [18]. It typically provides 35 to 45 dB isolation
in addition to antenna isolation.
Accurate estimation of the feedback channel is the key
to echo cancellation. One approach to channel estimation
is the use of a low power (e.g., more than 15 dB below
transmitted signal) reference signal, p (cf. Fig. 1), inserted into

the transmitted signal [9]. The reference signal is uncorrelated

with the transmitted signal. The feedback channel is then
estimated using the reference signal and the received signal
y, and used to recreate the feedback signals (echoes) which
are subtracted from the input signals. The drawback of using
an inserted reference signal is that the reference signal acts as
an unwanted interference and degrades the output signal SNR
of the repeater. Another approach is to use the original signal,
i.e., p = q, as the reference signal for channel estimation. This
method does not introduce extra interference to the transmitted
signal. In addition, it provides a better channel estimate than
the inserted reference approach since the original signal is
much stronger (e.g., 15 dB stronger) than the inserted reference
signal. The drawback of this approach is that the repeater has
to purposely introduce sufficient delay such that the signal
currently being transmitted is uncorrelated with the remote
signal currently being received. That is, the repeater delay has
to be larger than the remote signal delay spread.
The cancellation performance is directly dependent on the
quality and tracking capability of the feedback channel estimation. To obtain an accurate channel estimate, channel estimates
are usually averaged over time. The longest allowable average
time is limited by feedback channel coherence time. The use
of cancellation in modern repeaters increases the vulnerability
of a repeater to channel dynamics, thereby increasing the need
for careful gain management.
C. Stability and Output SNR
As earlier stated, the repeater output SNR and stability
depend on the repeater effective isolation (combined passive
and active isolation). As shown in Fig. 2, the repeater output
SNR is upper-bounded by the residual interference strength
relative to the remote signal, i.e., SNR. SNR = if the
repeater RF noise is ignored. It is clear that the inverse of
the SNR, i.e., SNR (dB) or (dB) is also related to the
loop gain of the feedback system. The repeater stability thus
requires that < 0. However, since the antenna isolation
and the cancellation performance are both dependent on the
feedback channel, the effective isolation (both passive and
active) in turn is affected by the environment. Changes in the
environment result in changes in isolation, and hence changes
in output SNR and stability. A mechanism that dynamically
controls the output SNR and stability is thus highly desirable.
The lack of this control mechanism in most repeater products

IEEE TRANSACTIONS ON BROADCASTING, VOL. 59, NO. 4, DECEMBER 2013

687

Fig. 3. Illustration of the effective feedback channel with feedback cancellation. The canceller removes the feedback signals, and as a result spreads
the residual energy over the entire cancellation window.

Fig. 2. Repeater input/output signal levels. Here the remote signal is the
desired signal to amplify.

is one of the key factors that limit the wide deployment of

on-channel repeaters.

corresponding reference signal is transmitted can be estimated

by

2


 N1

(1)
r [n]q[n + Ndelay + ]



Nch

n=0

where N is the correlation length in samples, Ndelay is the total

feedback delay, Nch the channel estimation length, and
p[n]
, 0n < N
(2)
r[n] = 
N1

2
|p[n]|
n=0

III. Gain Management

The proposed repeater gain management includes 1) link
level gain control and 2) device level gain management. They
are the main focus of this section.
A. Link Level Gain Control
The main goals of the proposed repeater gain control include
robust boot-up, oscillation prevention, and accurate output
SNR control. As weve already stated that the repeater gain is
limited by the repeater isolation whereas isolation is dependent
on the repeater operating environment. Ideally, the repeater
gain should change in accordance to the repeater isolation to
ensure the repeater stability as well as the output SNR. We
hence introduce a gain management mechanism, as shown in
Fig. 1 (red), to dynamically control the repeater gain such that
the repeater stability and the desired output SNR are ensured.
1) Loop Gain Metric: It is clear that in order to control
the repeater stability as well as the output SNR, we need
an estimate of the repeater loop gain as the metric for gain
control. The target value of the loop gain metric is typically in
the range of -20 to -15 dB, depending on the stability margin
and the target output SNR. This means that the noise floor
of the estimate metric has to be lower than -20 dB, which
is indeed challenging. Moreover, an issue that makes the
design of the metric even more challenging is the diffusion or
spreading effect of the canceller. Lets consider a simplified
repeater diagram as in Fig. 3. The canceller removes the
feedback signals, and as a result spreads the residual energy
over the entire cancellation window, which makes the effective
feedback energy more difficult to detect.
We devise the following estimation scheme. The feedback energy of the reference signal at the point where the

is the normalized reference signal with unit energy. Here p

N1

|r[n]|2 = 1, (1) represents the
is the reference signal. Since
n=0

amount of the feedback energy from the reference signal p.

The ratio between the feedback energy and the transmitted
energy corresponding to the reference signal is then

2

 N1
  r [n]q[n + Ndelay + ]


Nch n=0
=
N1

|p[n]|2
n=0


2

 N1
  p [n]q[n + Ndelay + ]


Nch n=0
=
,

2
N1

  |p[n]|2 



(3)

n=0

which is an estimate of the loop gain. Therefore,


|R()|2
Nch
=
|S|2
is used as the loop gain metric, where
2
N1





p [n]q[n + Ndelay + ]
R() = 



(4)

(5)

n=0

and

2


N1

2
|p[n]| 
S=



(6)

n=0

Fig. 4 plots the block diagram of the loop gain metric

generation. The down sampling by a factor M removes the

688

IEEE TRANSACTIONS ON BROADCASTING, VOL. 59, NO. 4, DECEMBER 2013

Fig. 4. Loop gain metric generation block diagram. The down sampling by a factor of M removes the correlated samples from the transmit filter to reduce
the computation complexity. The Reset signal clears the contents of Counter N as well as the R and S modules as defined by (5) and (6). When the Counter
N expires, it sends out the Ready signal indicating that the loop gain metric is available. It also resets the R and S modules. The Median-m filter is used
to remove outlier noise of the loop gain metric that may otherwise give rise to large false adjustment of the repeater gain.

correlated samples from the transmit filter to reduce the computation complexity. The Reset signal clears the contents
of Counter N as well as the R and S calculation modules.
When the Counter N expires, it sends out the Ready signal
indicating that the loop gain metric is available. It also resets
the R and S modules. The Median-m module is a non-linear
median filter with length m. It is used to effectively remove
outlier noise of the loop gain metric that otherwise may result
in a large false adjustment of the repeater gain.
2) Loop Gain Control: With the loop gain metric, the
instantaneous stability status and output SNR of the repeater
can be closely monitored and accurately controlled.
Loop gain control is for controlling the loop gain of the
repeater to the desired range such that the stability requirement
and the output signal quality of the carrier are met, and at the
same time the output gain is maximized. First, the target range
of the loop gain should provide sufficient separation from 0 dB
to ensure stability. Second, as earlier stated, since the residual
feedback signal not only causes instability but also presents
itself as interference to the original signal, the loop gain metric
also represents the inverse of the output signal SNR and has
to be controlled to meet the repeater output SNR requirement.
As depicted in Fig. 5, loop gain control logic manages the
loop gain by monitoring if the loop gain
 metric is within a
certain desired range, i.e., Tup , Tdown , corresponding to the
green zone in Fig. 5. Typically, Tdown is set at the value less
than the target output SNR, i.e., Tdown SNRtarget .
If the current loop gain metric is smaller than Tup , i.e.,
< Tup ,

(7)

which corresponds to the blue zone in Fig. 5, i.e., the loop

gain is less than the desired value, indicating there is room for
higher repeater gain, the repeater digital gain, G, is adjusted
up by an amount of

2
+ =  Tup 
(8)
dB, where is a coefficient that controls the step size of the
up adjustment. Whereas if the loop gain metric is greater than
Tdown , i.e.,
> Tdown ,

(9)

which corresponds to the red zone in Fig. 5 and indicates

that the current loop gain is higher than the desired value, the

Fig. 5. Illustration of gain adjustment as a function

of the
2 loop gain metric.

For < Tup , the gain is increased by + =  Tup  . For > Tdown ,
the gain is reduced by = | Tdown |2 . No gain adjustment is necessary
otherwise.

Fig. 6. Gain management flowchart. The gain update procedure is triggered

by any of the three events: loop gain metric update, saturation metric update,
and channel isolation update.

repeater digital gain, G, is then lowered by

= | Tdown |2

(10)

dB, where is a coefficient controlling the down adjustment

step size.
B. Device Level Gain Management
Device level gain management deals with the device limitation to ensure that repeater components, such as ADC, DAC,
and power amplifier are not saturated. Saturation may incur
degradation of output SNR and even instability of the repeater.
1) DAC Saturation Control: The output of the transmit
filter q, which contains mostly the gain adjusted remote signal
as well as the cancellation residual error, is the signal to be

IEEE TRANSACTIONS ON BROADCASTING, VOL. 59, NO. 4, DECEMBER 2013

689

Fig. 7. Illustration of the effectiveness of repeater gain management under various Doppler fading frequencies in the feedback channel. The maximum repeater
gain is set at 70 dB. The top row shows the Doppler frequency of the feedback channel. The switch between Doppler fading frequencies is instantaneous for
stress-testing the gain managers response to channel dynamics.

sent to the DAC, amplified, and transmitted. We need to first

make sure that this signal does not saturate the DAC. That is,
(E)dB + G < TDAC ,

(11)

effective channel gain from the input of the digital gain to

the input of the ADC, it can be obtained directly from the
channel estimate. The ADC headroom is then
Gmax
ADC (E, GISO ) = TADC (GRF + GISO ) (E)dB ,

where
E=

1
Nsat

Nsat


q(n)

(12)

n=1

is referred to as the saturation metric and TDAC is a threshold

related to DAC range. For example, TDAC = 68 dB for a 14-bit
DAC. The DAC headroom for the digital gain is
Gmax
DAC (E) = TDAC (E)dB ,

(13)

which is a function of the saturation metric.

2) ADC Saturation Control: The total signal presented to
the input of the ADC is the remote signal plus the feedback
(as well as RF noise). Since the feedback interference is much
hotter than the remote signal (cf. Fig. 2), the total energy at the
input of ADC is approximately equal to the feedback energy.
We thus require
(E)dB + (GRF + GISO ) + G < TADC ,

(14)

where GRF is the lumped RF gain, GISO is the antenna

isolation, and TADC is the threshold determined by the ADC
input dynamic range all in dB. Since (GRF + GISO ) is the

(15)

which is a function of both the saturation metric and the

channel isolation.
3) Overall Gain Head Room: Repeater gain is also limited
by other factors like the maximum PA and other physical
limits imposed by other RF components. Assuming the corresponding gain cap is Gcap , the overall gain head room is
thus


max
Gmax (E, GISO ) = min Gmax
DAC (E), GADC (E, GISO ), Gcap ,
(16)
The gain manager must ensure that this headroom is not to be
exceeded at all times, i.e.,
GGmax (E, GISO ).

(17)

C. Gain Update Procedure

The update procedure of the repeater gain is illustrated in
Fig. 6. The gain update procedure is triggered by any of
the update events: loop gain metric update, saturation metric
update, and channel isolation measure update.

690

IEEE TRANSACTIONS ON BROADCASTING, VOL. 59, NO. 4, DECEMBER 2013

Fig. 8. Stress-test of the repeater gain managements responsiveness to large-scale changes in isolation. The maximum repeater gain is set at 70 dB. The
canceller is turned on and off in a random fashion creating a 40 dB swing in effective isolation. This instant drop in isolation is responded by the gain
manager with a quick reduction in repeater gain to maintain the repeater stability and output SNR.

Fig. 9. The test of the repeaters response to large-scale changes in isolation. The maximum repeater gain is set at 70 dB. The canceller is turned on and off
in a random fashion creating a 40 dB swing in effective isolation. This instant drop in isolation is responded by the gain manager with a quick reduction
in repeater gain to maintain the repeater stability and output SNR.

IEEE TRANSACTIONS ON BROADCASTING, VOL. 59, NO. 4, DECEMBER 2013

691

Each time when a loop gain metric update is available or

ready (cf. Fig. 4), the loop gain controller determines if a
gain adjustment is necessary based on the loop gain metric.
If a gain increase of + is suggested, this gain adjustment
will be on hold until the new saturation metric is available
before the adjustment can be finally applied. As soon as the
saturation metric is available, the gain manager updates the
gain headroom and checks to see that the headroom is not
exceeded. That is,


.
G = min Gmax (E, GISO ), G + + ,
(18)

To show how well the loop gain management scheme can

adapt to the changing environment, fading channels with various Doppler frequencies is injected into the repeater feedback
channel as shown in Fig. 7 and Fig. 8, where the effective
isolation changes as the channel changes. In particular, since
the echo cancellation performance is sensitive to channel
variation speed, the faster the channel variation rate, the poorer
the cancellation performance is as shown by the repeater
output SNR in Fig. 7.
In Fig. 7, the gain management is disabled. As a result,
the repeater output SNR varies with the feedback channel
condition and cancellation depth. The required SNR (15dB),
as well as the stability, cannot be guaranteed. It is in fact
completely at the mercy of the environment.
In contrast, it is seen in Fig. 8, in which the gain management is active, that the repeater safely boots up and the
output SNR is maintained above the required level at all times.
The gain manager detects the effective isolation via the loop
gain metric and adjusts the gain accordingly such that the
output SNR is maintained above the desired level regardless
of the channel conditions. Note that, although unrealistic, the
switch between different Doppler fading rates is purposely set
to be instantaneous for stress-testing the gain manager. The
instant changes in channel parameters cause degraded channel estimation, and consequently the degraded cancellation
performance, thereby resulting in sudden changes in the effective isolation of the repeater. As a result, the output SNR
dips at the time of switch. However, the potential instability
is detected by the gain manager, as reflected by the soaringup of the loop gain metric, the repeater gain is thus promptly
reduced and the repeater stability, as well as the output SNR, is
maintained.
To further stress-test the gain management scheme, the
canceller is turned off and on in a random fashion resulting
in a 40 dB sudden swing in repeater effective isolation.
Oscillations are expected without gain management. However,
as seen from Fig. 9, with gain management the repeater is able
to quickly adjust the gain and manages to maintain stability
even under such harsh disturbances.

if a gain up adjustment + is pending. Otherwise,

.
G = min {max (E), G}

(19)

is asserted.
On the other hand, if a gain reduction is ordered by the loop
gain controller, the gain is updated and is applied immediately
without further delay,
.
(20)
G = G .
IV. Performance
In this section, we investigate the performance of the
proposed gain management scheme via a combination of a
simulation system and a prototyping system. The system used
WCDMA signals operating on 1850-1990 MHz. The patchdipole antenna system provides antenna isolation in the range
of -40 dB to -55 dB. Gain ripple over operation frequency
range was 1.5dB. Return loss for each antenna is better than
10 dB over operation range. A baseband FFT-based frequency
domain echo canceller gave an additional 35 to 45 dB active
isolation depth. This ensures that the residual error was 15 to
25 dB below the remote signal. The resolution of the ADC
and DAC was 14 bits, with sampling rate 120 and 240 Msps,
respectively. RF gain GRF 45dB. Gain cap, Gcap , was set
to 70 dB. Including the coverage antenna gain of 16 dB,
the maximum repeater gain was 86 dB. The repeater target
output SNR was 15 dB. The corresponding loop gain metric
thresholds were set to Tup = Tdown = 16dB, to allow an
1 dB margin for thermal and other imperfection
noise. The

conversion of digital gain G in dB to G in linear scale was

via a look up table (LUT). The median filter length m was set
to be three.
The gain was initialized at 20 dB, resulting in an unconditionally stable total loop gain of -20 dB. There was no
echo cancellation during the first 33 s for channel estimation
settlement. As soon as the channel estimation was stabilized,
echo cancellation was activated, thereby lowering the repeater
loop gain, and allowing the gain management to ramp up the
gain accordingly.
Fig. 7 through Fig. 9 plot the repeater total gain, output
SNR, and loop gain metric during boot-up and steady operation under various isolation conditions recorded from the
simulation system. The maximum repeater total gain is set at
70 dB. It is clear that the loop gain metric accurately reflects
the repeater output SNR (the inverse of the output SNR, cf.
Fig. 2).

V. Conclusion
The feedback of the transmit signal to the input of the
repeater in an on-channel repeater not only causes potential
instability but also acts as interference degrading the repeater
output signal quality. The potential danger of oscillation and
degradation of output SNR prevent repeaters from being
widely deployed. Oscillation prevention and output SNR control are thus crucial to the repeater design, and hence is the
main focus of this paper. Although there are different types
of repeaters and echo cancellers available in the literature (not
the focuses of this paper), this paper proposes an effective
gain management scheme for on-channel repeaters. In particular, a loop gain metric is introduced to continuously and
accurately detect the stability status of a repeater, allowing
the gain manager to control the output SNR and prevent
instability by appropriately adjusting the repeater gain in a
dynamic environment. The proposed active gain management

692

IEEE TRANSACTIONS ON BROADCASTING, VOL. 59, NO. 4, DECEMBER 2013

scheme is further tested in a simulation and a prototyping

system. It is demonstrated that the gain management scheme
is highly effective and robust in responding to the dynamics
of environments.

[9] K. M. Nasr, J. P. Cosmas, M. Bard, and J. Gledhill, Performance of an

echo canceller and channel estimator for on-channel repeaters n DVBT/H networks, IEEE Trans. Broadcast., vol. 53, no. 3, pp. 609618,
Sep. 2007.
[10] C. R. Anderson, S. Krishnamoorthy, C. G. Ranson, T. J. Lemon,
W. G. Newhall, T. Kummetz, and J. H. Reed, Antenna isolation,
wideband multipath propagation measurements, and interference mitigation for on-frequency repeaters, in Proc. IEEE SoutheastCon, 2004,
pp. 110114.
[11] R. N. Braithwaite and S. Carichner, Adaptive echo cancellation for an
on-frequency RF repeater using a weighted power spectrum, in Proc.
Eur. Conf. Wireless Technol., 2007, pp. 8285.
[12] P. Larsson and M. Prytz, MIMO on-frequency repeater with selfinterference cancellation and mitigation, in Proc. IEEE Veh. Technol.
Conf., Apr. 2009, pp. 15.
[13] A. A. Baghai and J. Divall, Cost effective solutions for coverage
enhancement, in Proc. Eur. Conf. Mobile Personal Commun., 1993,
pp. 5156.
[14] T. Riihonen, S. Werner, and R. Wichman, Optimized gain control for
single-frequency relaying with loop interference, IEEE Trans. Wireless
Commun., vol. 8, no. 6, pp. 28012806, Jun. 2009.
[15] C. Liu, M. He, and W. Xia, An optimal gain control scheme for
digital on-channel repeater with feedback interference canceller in digital
terrestrial television broadcasting networks, WSEAS Trans. Commun.,
vol. 10, no. 5, pp. 147154, May 2011.
[16] Y.-T. Lee, S.-I. Park, H. Eum, J. Seo, H. Kim, and S. Kim, A Design
of equalization digital on-channel repeater for single frequency network
ATSC system, IEEE Trans. Broadcast., vol. 53, no. 1, pp. 2337, Mar.
2007.
[17] J.-Y. Choi, M.-S. Hur, Y.-W. Suh, J.-S. Baek, Y.-T. Lee, and J.-S. Seo,
Interference cancellation techniques for digital on-channel repeaters in
T-DMB system, IEEE Trans. Broadcast., vol. 57, no. 1, pp. 4656, Mar.
2011.
[18] S. I. Park, H. Eum, S. R. Park, G. Kim, Y.-T. Lee, H. Kim, and
W. Oh, Novel equalization on-channel repeater with feedback interference canceller in terrestrial digital multimedia broadcasting system,
ETRI J., vol. 31, no. 4, pp. 357364, Dec. 2009.
[19] H. Chen, Study of radio frequency synchronized switch for TDSCDMA repeater station, in Proc. Int. Conf. Wireless Commun., Network. Mobile Comput., 2009, pp. 13.

Acknowledgement
The author would like to thank the reviewers for their
constructive comments and suggestions which helped improve
the paper. The author is also grateful to Dr. Gwendolyn Barriac
and Dr. Dhananjay Gore for their helpful discussions.

References
[1] Digital Video Broadcast (DVB): Framing Structure, Channel Coding
and Modulation for Digital Terrestrial Television, ETSI Standard EN
300 744 V1.5.1, 20042011.
[2] G. Faria, J. A. Henriksson, E. Stare, and P. Talmola, DVB-H: Digital
broadcast services to handheld devices, Proc. IEEE, vol. 94, no. 1, pp.
194209, Jan. 2006.
[3] M. R. Chari, F. Ling, A. Mantravadi, R. Krishnamoorthi, R. Vijayan,
G. K. Walker, and R. Chandhok, FLO physical layer: An overview,
IEEE Trans. Broadcast., vol. 53, no. 1, pp. 145160, Mar. 2007.
[4] A. Chakrabarti, C. Lott, D. Ghosh, and R. Attar, Repeaters and remote
radioheads in EVDO networks, in Proc. IEEE Veh. Tech. Conf., Sep.
2010, pp. 16.
[5] K. Salehian, B. Caron, and M. Guillet, Using on-channel repeater
to improve reception in DTV broadcasting service area, IEEE Trans.
Broadcast., vol. 49, no. 3, pp. 309313, Sep. 2003.
[6] K. Salehian, M. Guillet, B. Caron, and A. Kennedy, On-channel
repeater for digital television broadcasting service, IEEE Trans. Broadcast., vol. 48, no. 2, pp. 97102, Jun. 2002.
[7] Y. Lee, J. Lee, S. I. Park, Y. Lee, H. M. Kim, and H. Kim, Feedback
cancellation for T-DMB repeaters based on frequency-domain channel
estimation, IEEE Trans. Broadcast., vol. 57, no. 1, pp. 114120, Mar.
2011.
[8] D. Lee, M. Choi, and S. Choi, Channel estimation and interference
cancellation of feedback interference for DOCR in DVB-T system,
IEEE Trans. Broadcast., vol. 58, no. 1, pp. 8797, Mar. 2012.