Subcarrier Suppressed Transmission Scheme for Satellite/Terrestrial Integrated Mobile Communication System

Abstract

In this paper, we propose a subcarrier suppressed transmission scheme that improves the frequency utilization efficiency of STICS (Satellite/Terrestrial Integrated Mobile Communication System), which shares the same frequency band between satellite and land mobile systems. Applying our proposal to the OFDM-based land system overcomes the resource assignment limits of the frequency sharing scheme and yields new land spectrum resources. Computer simulations elucidate the interference suppression performance and the improvement in total average frequency utilization efficiency.

Full text

Page 1

Subcarrier Suppressed Transmission for OFDMA
in Satellite/Terrestrial Integrated Mobile
Communication System
Jun Mashino
NTT Access Network Service systems Laboratories, NTT Corporation
1-1 Hikari-no-Oka Yokosuka, Kanagawa, Japan
{mashino.jun, sugiyama.takatoshi}@lab.ntt.co.jp


Abstract—In this paper, we propose subcarrier suppressed
transmission for OFDMA (Orthogonal Frequency Division
Multiple Access) in order to improve frequency utilization
efficiency of STICS (Satellite/Terrestrial Integrated Mobile
Communication System), which shares the same frequency band
between satellite and land mobile systems. In OFDMA, we
introduce a channel rotation scheme that achieves a performance
level for multiusers that is similar to that for a single user case.
Based on computer simulation, we show the interference
suppression performance and clarify the improvement of the
total average spectral efficiency.
Takatoshi Sugiyama
Keywords-component; Subcarrier Suppression, Interference
Suppression, OFDM, FEC, Metric Masking, Frequency Utilization
Effieicncy
I.
been
INTRODUCTION
several There have investigations on
Satellite/Terrestrial Integrated Mobile Communication Systems
(STICSs), which incorporate satellite and land mobile
communication systems, with the purpose of achieving a highly
reliable and spectral-efficient wireless system [1]-[3]. In STICS,
the satellite communication system (satellite system hereafter)
covers mountain/ocean areas,
communication system (land system hereafter) covers
urban/suburban areas including indoor users, and both satellite
and land systems share the same frequency band in order to use
frequency resources efficiently. The targeted frequency band
here is a 2- GHz frequency pair band (1980-2010 MHz and
2170-2200 MHz) since it is already assigned to both Mobile
Satellite Service (MSS)
Telecommunication (IMT) systems.
The frequency sharing method for STICS was investigated
in [1]. This method divides frequency resources into multiple
sub-bands for frequency reuse so that land cells inside a
satellite cell do not use the same frequency resources used by
the satellite cell to avoid co-channel interference. Note that the
optimum bandwidth of the sub-band is different between these
two systems. However, in the conventional scheme
"unavailable" sub-bands emerge since the spectra of land sub-
bands that are overlapped by the outer satellite cell are not
activated, even if there is only slight spectrum overlapping.
Therefore we previously proposed a subcarrier suppressed
transmission scheme that can use those unavailable frequency
and the land mobile
and International Mobile
resources.[4] It is based on interference suppression techniques,
which suppresses some subcarriers after generating original
coded OFDM signals in the land mobile transmitter. It
maintains the transmission bit rate by using Forward Error
Correction (FEC) and FEC Metric Masking technique in order
to suppress the overlapping satellite signal spectrum for the
land mobile receiver and to improve the degradation due to the
suppression. These interference suppression techniques do not
change the air interface. In [4] the performance is showed in
the case of single-user OFDM signaling and that the proposed
scheme activates unavailable land sub-bands. In this paper, we
apply the scheme to a multiuser OFDMA scenario, where
subcarriers are divided among several users in the land mobile
uplink. We show that the scheme is also applicable to a
multiuser scenario by employing a channel rotation scheme
that is efficient in terms of the average subcarrier suppression
rate among users.
The rest of this paper is organized as follows. In Section II,
the targeted system model is described. In Section III, we
introduce the proposed scheme in detail. Section IV briefly
describes the simulation parameters. In Section V, we evaluate
the proposed scheme, and Section VI presents our conclusions.
II.
SYSTEM MODEL IN STICS
According to frequency-band regulation for MSS, we
assume that both satellite and land systems employ Frequency
Division Duplex (FDD) as the duplexing scheme. Two
Figure 1. System model.
relay
Feeder (out of scope)
Land Mobile
Satellite
Frequency
Sharing
BS
MSL
MSS
SS
ES
f2
F1
F1 and f2 are different frequency
but belong to same frequency band
978-1-61284-231-8/11/$26.00 ©2011 IEEE
This full text paper was peer reviewed at the direction of IEEE Communications Society subject matter experts for publication in the IEEE ICC 2011 proceedings

Page 2

frequency operation modes are considered for STICS [1]-[3].
One is the normal mode in which the same bands are used for
the satellite uplink (downlink) and land uplink (downlink), and
the other is the reverse mode in which the same bands are used
for the satellite uplink (downlink) and land downlink (uplink).
From the viewpoint of adjacent channel leakage interference
from the satellite uplink against the land system, the reverse
mode is a better choice since land base stations have a much
higher antenna directivity than land terminals with non-
directional antennas. Hence in this paper, we focus on a
frequency band sharing scenario for the satellite downlink and
land uplink as indicated in Fig. 1. Note that the satellite system
comprises an Earth Station (ES), Satellite Station (SS), and
Mobile Station (MSS), and the land system comprises a Base
Station (BS) and Mobile Station (MSL).
A. Satellite System
In order to support a large volume of satellite traffic and
achieve higher area-spectral efficiency, satellite cells are
constructed using large deployable multi-beam satellite
antennas. These high-gain antennas can be used to downsize
mobile terminal stations. Although each beam has very sharp
directivity to the desired satellite cell, there are residual antenna
sidelobes and they are harmful to neighboring cells. Reference
[1] shows that at least 7 frequency sub-bands are needed for
frequency reuse. Thus, the frequency bandwidth per satellite
sub-band is 30 MHz/7 = 4.3 MHz since STICS has a 30 MHz x
2 pair band as shown in Fig. 2. Note that the “f” indicates a
satellite sub-band, shown as f1.
B. Land System
Since IMT is also assigned to the MSS band, it is highly
desirable that land system employs IMT. In this paper, we
focus on an OFDM-based high-speed wireless system as an
IMT system, for example, the Mobile WiMAX system (FDD
Profile [5]). To adhere to IMT regulations, basically a 5.0-MHz
bandwidth channel must be used and each center frequency is
fixed. Thus, the frequency bandwidth per land sub-band is 5
MHz. STICS can accept 6 sub-bands in total since it has a 30
MHz x 2 pair band; see Fig. 3. It is larger than 3 that is the
necessary number of sub-bands for frequency reuse. Note that
the “F” indicates a land sub-band, shown as F1.
C. Frequency Sharing between Satellite and Land systems
A conventional frequency sharing scheme segregates land
sub-bands and satellite sub-bands inside the satellite cell in
order to avoid co-channel interference as shown in Fig. 4(a).
Fig. 4 illustrates frequency sharing inside a satellite cell using
f2 as an example. The conventional scheme generates
unavailable land bands due to the bandwidth difference
between the satellite and land systems. To avoid interference,
land bands F1 and F2 are not available inside the satellite cell
using f2. Note that in STICS, a land sub-band might be
overlapped by a satellite sub-band at the rate of m/7 (m=1,2,…,
6) = 14, 29, 43, 57, 71, 86% of its bandwidth because there are
7 satellite sub-bands and 6 land sub-bands. One land-band is
unavailable for the land cells inside a satellite cell using f1 or
f7. Two land sub-bands are unavailable for the land cells inside
a satellite cell using f2, f3, f4, f5, or f6.
In [4], new land signals are assigned to a part of land sub-
band F1, denoted as F1’, inside the satellite cell using f2 as
shown in Fig. 4(b). Similarly a new signal is allocated to part of
F2 in f3, and so on. These are achieved using interference
suppression techniques for the land transmitter and land
receiver, and frequency utilization efficiency is improved. The
details regarding the interference suppression techniques are
given in the following section.
III.
PROPOSED SCHEME
A. Proposed Interference Suppression Scheme for Transmitter
We assumed Coded-OFDM (COFDM) [6] signaling that
employs FEC coding and a subcarrier interleaver. Land MSLs
that have land sub-bands that are overlapped by a sub-band of
the outer satellite cell transmit signals after suppressing the
subcarriers on the overlapped bandwidth. The suppression is
easily implemented using subcarrier puncturing; see Fig. 5. Fig.
5 shows block diagram of transceiver. Since subcarrier
suppression is done after generating a full-bandwidth COFDM
signal in the transmitter, it does not degrade the transmission
bit rate and does not affect anything regarding the existing IMT
air interface. Gain adjustment is executed in order to reallocate
power of the suppressed spectrum on transmit subcarriers.
Therefore average transmit power level is always constant after
subcarrier suppression.
(a) Conventional

(b) Proposed
Figure 4. Frequency sharing scheme.
F4
F5
F6
Satellite cell
f1
f3
Satellite cell
f1 f2 f3f4f5 f6f7
F1 F2F3
MSS
IMT
30MHz
Freq.
Satellite cell
f2
F3
Land cell
Land cell
Not used to prevent co-channel interference
F4F5F6
F4
F5
F6
Satellite cell
f1
f3
Satellite cell
f1f2 f3 f4f5f6 f7
F1 F2F3
MSS
IMT
30MHz
Freq.
Satellite cell
f2
F3
Land cell
Land cell
Assign new land signals while preventing co-channel interference
F4F5F6
F1’
F1’
Figure 3. Sub-bands for land system (IMT).
Freq.
IMT
30MHz (band for STICS)
5.0MHz/ch
F1 F2F3 F4F5 F6
Figure 2. Sub-bands for satellite system (MSS).
Freq.
f1f2f3f4 f5f6 f7
MSS
30MHz (band for STICS)
4.3MHz/ch
This full text paper was peer reviewed at the direction of IEEE Communications Society subject matter experts for publication in the IEEE ICC 2011 proceedings

Page 3

TABLE I.
MAJOR SIMULATION PARAMETERS
Common Parameter
Frequency band 2.0 GHz
ITU-R M.1225 (3.9 power law) on ground
Free space propagation loss in space
ITU-R M.1225 Ped.B on ground only
6 path-Reyleigh(????? 7.2 ? 10??)
Propagation loss
Channel model
Satellite Communication System
Modulation
QPSK
Convolutional Coding 1/2
w/ Viterbi Decoder
3.7 MHz
10 dB
-106 dBm/MHz
Root-raised-cosine filter (roll-off: 0.35)
100 km
FEC
Symbol rate
CNRreceived
Noise density
BPF
Cell radius
Land Mobile Communication System (Mobile WiMAX)
Number of users
Channel rotation
1st modulation
2nd modulation
Subcarrier spacing
Subcarrier permutation
3 or 17
On or off
QPSK, 16QAM
OFDM (427 subcarriers)
10.94 kHz
UL-PUSC in IEEE 802.16e [10]
Convolutional Coding 1/2, 3/4
w/ Viterbi Decoder
-106 dBm/MHz
Inverse-Chebychev type IIR filter
(meets the Mobile WiMAX requirement
[11])
Zero Forcing
Ideal
3.3 km
FEC
Noise density
BPF
Freq. domain equalizer
Synchronization
Cell radius
B. Use of Channel Rotation
We assume that the uplink transmission of WiMAX is
based on OFDMA, which enables multiuser access by dividing
subcarriers. It is possible that the number of suppressed
subcarriers is different among users when applying the
proposed interference suppression scheme to an OFDMA
system, since the suppressed bandwidth is concentrated at the
far left or right of the land sub-bands. This may degrade the
transmission performance.
We therefore introduce channel rotation into the scheme.
Channel rotation is a technique to achieve frequency diversity.
Similar to frequency hopping, the scheme hops subcarriers
assigned to each OFDM symbol by changing the mapping from
logical subcarriers to physical subcarriers. Channel rotation
defined in the IEEE 802.16e standard [7] is applied every 3
OFDM symbols, which is the minimum allocation unit in the
time domain, and it is switched on and off. Channel rotation is
applied between subcarrier interleaving and subcarrier
suppression as indicated in Fig. 5.
C. Proposed Interference Suppression Scheme for Receiver
The subcarrier suppressed signal is similar to an OFDM
signal in which one of the subcarriers is severely degraded due
to frequency selective fading. Hence, a land receiver can
correct errors due to subcarrier suppression if the metric,
which is input to the FEC decoder, is weighted according to
each received subcarrier power level. However, an adjacent
satellite signal received on the overlapped bandwidth may
degrade the error correction ability, since a large absolute
metric is improperly calculated for the suppressed subcarriers.
We therefore use the FEC Metric Masking technique, which
was proposed for superposed transmission [8]. The technique
neglects the power of the received signals on the overlapped
bandwidth, and sets corresponding metrics for the subcarrier on
the overlapped bandwidth with neutral values. When the Log-
Likelihood Ratio (LLR) is used as the metric, the neutral value
is zero. Fig. 5 shows a block diagram of the land receiver. After
the symbol demapper, calculated metrics are replaced through
the zero replacer only for the satellite sub-bands. This approach
improves the error correction performance of FEC since
reliable subcarriers are relatively emphasized.


Figure 5. Block diagram of land mobile transceiver

For non-overlapping band
MUX
BPF
OFDM
demodulation
Symbol
demapping
(calc. metric)
Zero
replacement
For overlapping band
Subcarrier
deinterleaving
FEC
decoding
Channel
derotation
Output
Bits
FEC
encoding
Symbol
mapping
(PSK/QAM)
Subcarrier
interleaving
OFDM
modulation
BPF
Input
Bits
Channel
rotation
Subcarrier
suppression
Freq.
Time
Freq.
Time
Freq.
Time
Freq.
Time
Overlapping band
Suppression
Freq.
Time
Gain
adjustment
Freq.
Power
Suppression
Freq.
Power
Power reallocation
Transmitter
Receiver
This full text paper was peer reviewed at the direction of IEEE Communications Society subject matter experts for publication in the IEEE ICC 2011 proceedings

Page 4

IV. SIMULATION PARAMETERS
Table I gives the simulation specifications. The satellite
system (downlink of service link) adopts a single carrier QPSK
transmission. The land system (uplink) is compatible with
IEEE 802.16e-2005 OFDMA PHY [9]. Subcarrier allocation
conforms to UL-PUSC [10], which is the mandatory
permutation method in the uplink. Note that full loading is
assumed, which means that the entire bandwidth for the land
system is fully occupied by users. It is also assumed that the
same number of subcarriers is fairly allocated to each user.
Satellite and land signals are adjacently allocated in frequency
domain. The guard band between the satellite and land systems
is (5 MHz－10.94 kHz/subcarrier * 427 subcarriers) / 2 ＝
164.31 kHz, which conforms to the guard bands between the
land spectra.
V.
EVALUATION
A. Effect of Channel Rotation
First, the effect of channel rotation is briefly evaluated
based on the distribution of frequency utilization efficiency per
user as shown in Fig. 6. Fig. 6(a) shows the results for 3 users,
and Fig. 6(b) shows the results for 17 users. The figures show
large differences in frequency utilization efficiency among
users when channel rotation is not used. Especially for 17 users,
it appears that the difference is so severe that the throughput for
one of the users is zero. The degree of fairness is improved
using the channel rotation scheme. The efficiency per user is
completely averaged in the case of the three users, and similar
results are achieved for the case of 17 users. This is achieved
because the channel rotation scheme thoroughly scrambles
user-assigned subcarriers and the suppression rate per user is
averaged. Hence, we can adopt the proposed scheme into an
OFDMA system and for multiusers it achieves similar
performance to that for a single user.
B. Frequency Utilization Efficiency
Fig. 7 shows the user-averaged frequency utilization
efficiency versus the signal-to-noise power ratio (SNR). In
STICS, suppression rate α can accept the discrete values of
14%, 29%, 43%, 57%, 71%, or 86% due to spectrum
overlapping between satellite and land sub-bands, as mentioned
in Section II-C. Note that the proposed scheme transmits a
signal with subcarrier suppression using the channel rotation
technique, and the conventional scheme transmits a signal
without subcarrier suppression. The figure shows that the
proposed scheme is applicable for α of 14% or 29%. In the
proposed scheme the peak frequency utilization efficiency
increases when α is a larger value because the employed
spectrum bandwidth becomes narrower. However, the
proposed scheme for α = 14% and the conventional scheme for
α = 0% can utilize the higher coding rate of 3/4 when the SNR
is sufficiently high, while the proposed scheme for α = 29%

(a) 3 users

(b) 17 users

Figure 6. Distribution of frequency utilization efficiency.
123
User (i)
0.0
1.0
2.0
3.0
w/ ch rotation
w/o ch rotation
Frequency Utilization Efficiency (bit/sec/Hz)
degraded
user
1 2 3 4 5 6 7 8 9 10 1112 1314 151617
User (i)
0.0
1.0
2.0
3.0
Frequency Utilization Efficiency (bit/sec/Hz)
w/ ch rotation
w/o ch rotation
degraded
users
Figure 7. Average frequency utilization efficiency.

0.0
0.5
1.0
1.5
2.0
2.5
3.0
0510152025 30
SNR (dB)
Average Frequency Utilization Efficiency (bit/sec/Hz)
prop. (α=14%)
prop. (α=29%)
prop. (α=43, 57, 71, 86%)
conv. (α=0%)
BLER=0.1
17users
This full text paper was peer reviewed at the direction of IEEE Communications Society subject matter experts for publication in the IEEE ICC 2011 proceedings

Page 5

Figure 8. Average spectral efficiency.

Inside
f1 MSS cell
Inside
f2 MSS cell
Inside
f3 MSS cell
Inside
f4 MSS cell
Inside
f5 MSS cell
Inside
f6 MSS cell
Inside
f7 MSS cell
Average Spectral Efficiency（bit/sec/Hz）
IMT average
STICS
Total average
IMT
only
0.00.20.4 0.60.81.0 1.2
＋＋12％％ improved
Conv.(α=0%)Prop.(α=0, 14, 29%)
w/o ch rotation
w/ ch rotation
can utilize the lower coding rate of 1/2 due to the required error
correction ability against the suppression rate. This is why the
performance for α = 14% exceeds that for α = 29% when the
SNR is higher than 15 dB.
C. Total Spectral Efficiency Improvement
Fig. 8 shows the total average spectral efficiency obtained
from the results given in Section V-B. Note that we assume that
MSLs are uniformly distributed in the land cells, and are
allocated the same amount of radio resources. Based on the
results from the previous subsection, the available frequency
resource for the land cells inside a satellite cell using f2 (α =
14%), f3 (α = 29%), f5 (α = 29%), and f6 (α = 14%) are
improved using the proposed scheme with channel rotation, as
indicated by the dotted-line boxes in the figure. Thus, the total
average frequency utilization efficiency is improved from 0.97
bit/sec/Hz to 1.09 bit/sec/Hz. The improvement ratio is 12%.
VI. CONCLUSION
We apply the proposed subcarrier suppressed transmission
scheme to the OFDMA uplink in STICS. We clarified that the
channel rotation scheme well averages the subcarrier
suppression rate among users and yields multiuser transmission
performance that is almost the same as that for a single user.
The proposed scheme allocates new land signals on land sub-
bands which are spectrum-overlapped by 14% or 29% by the
satellite sub-bands. As a result, the total average frequency
utilization efficiency is improved by 12%.
REFERENCES
[1] T. Minowa, M. Tanaka, N.Hamamoto, Y. Fujino, N. Nishinaga, R.
Miura, K. Suzuki, “Satellite/Terrestrial
Communication System for Nation's Security and Safety,” IEICE Trans.
Commun., Vol.J91-B, No.12, pp.1629-1640, Dec. 2008.
[2] Y. Fujino, A. Muira, H. Tsuji, H. Watanabe, N. Hamamoto, R. Suzuki,
“Research and Development on Satellite/Terrestrial Integrated Mobile
Communication System for Secured and Safe Society,” Technical report
of IEICE. SAT, vol. 109, no. 254, pp.59-61, Oct. 2009.
[3] M. Umehira, “A System Model for Frequency Sharing Study in Satellite
and Terrestrial Integrated Mobile Communication Systems, ” Technical
report of IEICE. SAT, vol. 108, no. 274, pp.27-32, Oct. 2008.
[4] J. Mashino, Y. Shirato, T. Sugiyama, “Subcarrier Suppressed
Transmission Scheme for Satellite/Terrestrial Integrated Mobile
Communication System,” VTC 2010-Fall, IEEE, Sep. 2010.
[5] WiMAX Forum Mobile System Profile Release 1.0 Approved
Specification Revision 1.6.1, WiMAX Forum, Apr. 2008.
[6] Macedo, A.S., Sousa, E.S., “Coded OFDM for broadband indoor
wireless systems, ” Conference on Communications, ICC 97, IEEE, Vol.
2, pp. 934-938, June 1997.
[7] IEEE Std. P802.16 Rev2/D6, “Data subchannel rotation scheme, ” Air
Interface for Fixed and Mobile Broadband Wireless Access Systems ，
IEEE, p.1004, Oct. 2007.
[8] J. Mashino, T. Sugiyama, “Total frequency utilization efficiency
improvement by superposed multicarrier transmission scheme,” 2009
IEEE 20th International Symposium, PIMRC2009, IEEE, pp.1677-1681
Spt. 2009.
[9] IEEE Std. 802.16e-2005 and 802.16-2004/Cor 1-2005, Air Interface for
Fixed and Mobile Broadband Wireless Access Systems, IEEE, Feb.
2006.
[10] IEEE Std. P802.16 Rev2/D6, “Symbol structure for subchannel (PUSC), ”
Air Interface for Fixed and Mobile Broadband Wireless Access Systems
，IEEE, pp.995-997, Oct. 2007.
[11] WiMAX Forum Mobile Band Class 3A Spectrum Mask, Annex A,
WiMAX Forum, Apr. 2008.

Integrated Mobile

This full text paper was peer reviewed at the direction of IEEE Communications Society subject matter experts for publication in the IEEE ICC 2011 proceedings