Rohling H. (Ed.) OFDM: Concepts for Future Communication Systems
Подождите немного. Документ загружается.
206 5 System Level Aspects for Multiple Cell Scenarios
1
2
3
:
:
:
:
iteratively repeated interference presubtraction
(1)
ˆ
g
1
()
1
d
matched filtering
BS 1
(1)
s
BS 2
BS 3
() ()
1,1 3,1 *
IU
ˆˆ
HH
() ()
3,1 1,1 *
IU
ˆˆ
HH
()
()
1
ti
()
()
1
1-ti
…
()
1
t
()
1,1 *
U
ˆ
H
()
3,1 *
U
ˆ
H
(2)
ˆ
g
1
()
2
d
(2)
s
() ()
1,2 2 ,2 *
IU
ˆˆ
HH
() ()
2,2 1,2 *
IU
ˆˆ
HH
()
()
2
ti
()
()
2
1-ti
…
()
2
t
()
2,2 *
U
ˆ
H
()
1,2 *
U
ˆ
H
(3)
ˆ
g
1
()
3
d
(3)
s
() ()
2,3 3,3 *
IU
ˆˆ
HH
() ()
3,3 2,3 *
IU
ˆˆ
HH
()
()
3
ti
()
()
3
1-ti
…
()
3
t
()
3,3 *
U
ˆ
H
()
2,3 *
U
ˆ
H
MS
BS
significant useful
channels for MS 1
significant interference
channels for MS 1
Figure 5.23: Decentralized implementation of JT with partial CSI in the DL of an
exemplary 3-cell scenario
where the channel energy scaling factor is
ˆg
(k
M
)
=
k
A
ˆ
H
(k
M
,k
A
)
U
ˆ
H
(k
M
,k
A
)
∗
U
. (5.39)
Firstly, at the BSs corresponding to the significant useful channels of the MS k
M
,we
calculate the matched filtering data estimates r
(k
M
)
considering only the significant
useful CSI. Secondly, significant interferences from other MSs to the BSs antennas
considered for MS k
M
in the first step are reconstructed considering the significant
interference CSI, and subtracted from the corresponding matched filtering data es-
timates r
(k
M
)
. This step is implemented in an iterative way, and the preliminary
detection results
ˆ
d
(k
M
)
(i − 1) from other MSs are required.
In the DL, the matrix-vector-notation form of the iterative algorithm for JT with
partial CSI can be described by
t(i)=
ˆ
G
−1
·
d −diag
ˆ
R
·t(i −1)
(5.40)
and
s =
ˆ
H
∗T
U
· t . (5.41)
5.4 Interference Reduction 207
0 0.2 0.4 0.6 0.8 1
0
0.1
0.2
0.3
0.4
0.5
0.6
MF(c), omni
MF(c), sector
JT(4,2), omni
JT(4,2), sector
r/R
Prob(γ ≤ 10dB)
Figure 5.24: Outage probability of SIR for MS moving from cell center to the “bad vertex” in
the center cell in the DL of 19 cell cellular system including one center cell and two
tiers of cells around it with cluster size 3
The corresponding decentralized implementation of the iterative algorithm for each
MS k
M
can be described as
t
(k
M
)
(i)=
1
g
(k
M
)
·
d
(k
M
)
−
k
A
k
M
=k
M
ˆ
H
(k
M
,k
A
)
I,k
M
ˆ
H
(k
M
,k
A
)
∗
U
· t
(k
M
)
(i −1)
(5.42)
and
s
(k
A
)
=
k
M
ˆ
H
(k
M
,k
A
)
∗
U
· t
(k
M
)
. (5.43)
Firstly, for each MS k
M
at its considered BSs corresponding to its significant useful
channels, the significant interferences from this MS to other MSs are predicted and
compensated. Through this process, the predistorted data symbol t
(k
M
)
is obtained
in an iterative way considering significant interference CSI, and the preliminary
predistorted data symbol t
(k
M
)
(i − 1) from other MSs are required. Secondly, the
transmitted signals s
(k
A
)
at the BS antennas corresponding to the significant useful
CSI of the MSs are calculated from the predistorted data symbols t
(k
M
)
of all MSs.
Taking a 3-cell omni-DAS with a single antenna at each BS as an exemplary
scenario, the above decentralized signal processing for JT with significant CSI is
visualized in Fig. 5.23. For example, the signal processing for MS 1 is performed at
its neighboring BS 1 and BS 2 corresponding to its two significant useful channels.
Not only the smart reception/transmission technique based on JD and JT with
partial CSI as discussed above, we have also investigated and compared omni-DASs
and sector-DASs which have been introduced in Section 5.4.2. The numerical results
of the system performance in the DL with different combinations of transmission
strategies and BS antenna layouts are shown in Fig. 5.24. It can be seen that itera-
tive JT with N
u
significant useful channels and N
i
significant interference channels
indicated by “JT (N
u
,N
i
)” has much better performance as compared to intracell
matched filtering (MF) indicated by “MF(c)”,bothinomni-DASandsector-DAS.
208 5 System Level Aspects for Multiple Cell Scenarios
−10 0 10 20 30 40 50
−10
−5
0
5
10
15
20
25
upper performance bound with perfect CSI
adaptive scheduling
with imperfect CSI
random scheduling
A
B
10log
10
{γ
E
}/dB
10log
10
{E{γ
(1)
}}/dB
Figure 5.25: Average SNIR E{γ
(1)
} as a function of γ
E
, applying JT(57, 1026), 10log
10
{γ
N
} =
20dB, A: significant channel selection and JT are based on perfect CSI, B: significant
channel selection and JT are based on imperfect CSI
The sector-DAS outperforms the omni-DAS even more when applying iterative ZF
JT with significant CSI as compared to intracell MF.
5.4.5 Impact of Imperfect CSI on Cooperative
Communication Based on JD/JT
The impact of imperfect CSI on cooperative communication scheme is investigated
with the help of numerical results. We take a 19 cell cellular system including one cell
in the center and two tiers of cochannel cells around it as the reference scenario. The
performance of the MS in the center cell is assessed. Adaptive scheduling techniques
have been intensively investigated in [14, 15]. In [16], the impact of imperfect CSI
on the adaptive MS scheduling in the proposed cooperative communication scheme
has been investigated. From Fig. 5.25, we can clearly see that with increasing im-
perfectness of the CSI, the system performance when applying adaptive scheduling
approaches that of applying random scheduling. Under the realistic assumption that
the significant channel selection and the JT with partial CSI are based on imperfect
CSI, the system performance sharply decreases when the extent of imperfectness of
CSI increases. The performance degradation caused by imperfect CSI used in the
significant channel selection and JT is much larger than the performance degrada-
tion caused by replacing adaptive scheduling with random scheduling. In Fig. 5.26
and Fig. 5.27, the MS in the center cell is located at the fixed position of the “bad
vertex”, while the other MSs are randomly distributed in the other cells. In Fig. 5.26,
taking the upper system performance bound obtained by applying perfect CSI in
the whole cooperative transmission scheme as the reference, we can see the system
performance degradation caused by the imperfect CSI used in the significant channel
selection and JT with partial CSI, respectively. In Fig. 5.27, assuming imperfect CSI
is applied in the whole cooperative transmission scheme, both omni-DAS and sector-
DAS cooperative communication schemes considering different partial CSI selection
strategies with different numbers of significant channels are investigated. As a func-
5.4 Interference Reduction 209
−10 0 10 20 30 40
−10
−5
0
5
10
−10 0 10 20 30 40
−5
0
5
10
15
(a) omni-DAS (b) sector-DAS
upperbound
upperbound
JT based on
imperfect partial CSI
imperfect partial CSI
significant channel selection
JT based on
imperfect partial CSI
significant channel selection
based on imperfect CSI
10log
10
{γ
E
}/dB
10log
10
{γ
E
}/dB
10log
10
{E{γ
(1)
}}/dB
10log
10
{E{γ
(1)
}}/dB
Figure 5.26: Average SNIR E{γ
(1)
} as a function of γ
E
, applying JT(4, 2), 10log
10
{γ
N
} = 20dB
21 21.5 22 22.5 23
8.2
8.4
8.6
8.8
9
9.2
9.4
9.6
9.8
10
22.5
23
23.5
24
24.5
12.8
13
13.2
13.4
13.6
13.8
14
14.2
MF(c)
JT(4,2)
JT(4,4)
JT(57,1026)
MF(c)
JT(4,2)
JT(4,4)
JT(57,1026)
(a) omni-DAS
(b) sector-DAS
10log
10
{γ
E
}/dB
10log
10
{γ
E
}/dB
10log
10
{E{γ
(1)
}}/dB
10log
10
{E{γ
(1)
}}/dB
Figure 5.27: Average SNIR E{γ
(1)
} as a function of γ
E
, 10log
10
{γ
N
} = 20dB
tion of the extent of imperfectness of the CSI, we can decide how much CSI should be
considered in the cooperative communication scheme to obtain optimum system per-
formance with moderate implementation complexity. Additionally, the influence of
the imperfectness of CSI on the advanced JD/JT algorithms, i.e., SIC in the UL and
THP in the DL with quantization, has been investigated. In Fig. 5.28 and Fig. 5.29
considering the exemplary channel matrix H
UL
= H
T
DL
=(1010.1; 10 2 0.1; 10 5 1),
the input SNR has been adjusted in such a way that BER is always 10
−3
when ap-
plying different JD/JT algorithms with perfect CSI. In Fig. 5.28, ZF, SIC, ML, and
optimum MF with the ideal assumption of no multiuser interference are considered
for JD in the receivers. In Fig. 5.29, ZF, THP, and optimum MF with the ideal
assumption of no multiuser interference are considered for JT in the transmitters.
We can see the sensitivity of different JD/JT algorithms to the imperfectness of CSI
described by the variance σ
2
Δ
of the channel errors.
210 5 System Level Aspects for Multiple Cell Scenarios
−60
−40
−20
0
20
40
10
−4
10
−3
10
−2
10
−1
10
0
ZF
SIC
ML
Opt. MF
10log
10
{σ
2
Δ
}
BER
Figure 5.28: BER as a function of the extent
of the imperfectness of CSI in
the UL applying JD algorithms
−60
−40
−20
0
20
40
10
−4
10
−3
10
−2
10
−1
10
0
ZF
THP
Opt. MF
10log
10
{σ
2
Δ
}
BER
Figure 5.29: BER as a function of the extent
of the imperfectness of CSI in
the DL applying JT algorithms
0
10
20
30
40
10
−5
10
−4
10
−3
10
−2
ZF
suboptimum
optimum
10log
10
{1/σ
2
Δ
}
BER
Figure 5.30: BER as a function of the extent
of the perfectness of CSI with
10log
10
{1/σ
2
N
} = 40dB
0
10
20
30
40
10
−4
10
−3
10
−2
ZF
suboptimum
optimum
10log
10
{1/σ
2
N
}
BER
Figure 5.31: BER as a function of PSNR with
10log
10
{1/σ
2
Δ
} = 15dB
5.4.6 Advanced Algorithm Based on Statistical Knowledge
of Imperfect CSI
Exploiting the statistical knowledge of imperfect CSI and the given alphabet S of
the transmitted vector s, we propose an optimum detection scheme following the
rationale of maximum likelihood (ML) as
ˆs =argmax
s∈S
!
p
e|s,
ˆ
h
"
, (5.44)
where p
e|s,
ˆ
h
is calculated as the marginal PDF of p
e, h|s,
ˆ
h
as
p
e|s,
ˆ
h
=
p
e, h|s,
ˆ
h
dh =
p(e|s, h)p(h|
ˆ
h)dh . (5.45)
In (5.45), p(h|
ˆ
h) can be calculated according to the Bayes’ theorem as
p(h|
ˆ
h)=
p(
ˆ
h|h)p(h)
#
p(
ˆ
h|h)p(h)dh
. (5.46)
5.4 Interference Reduction 211
Cooperative Communication
diversity techniques
with no CSI
adaptive
scheduling
significant
channel
selection
estimated full
channel matrix
of whole system
ˆ
H
F
: K
M
× K
A
K
M
= N
M
· K
estimated full
channel matrix
with considered users
ˆ
H : K
M
× K
A
K
M
= K
estimated partial
channel matrix
ˆ
H
U
,
ˆ
H
I,k
M
including significant
CSI from
ˆ
H
JD and JT
with partial
CSI
H
U
H
I,k
M
imperfect CSI
in realistic system
Figure 5.32: Cooperative communication with different levels of knowledge of CSI
Its application in two-user interference channels is presented. In Fig. 5.30 and
Fig. 5.31, numerical results show the improvement of the system performance when
applying optimum detection as compared to ZF and suboptimum detection where
estimated channel coefficients are directly taken as perfect channel coefficients in
ML detector.
5.4.7 System Concept Based on Different Levels of
Knowledge of CSI
We aim at achieving high system performance in realistic mobile radio cellular sys-
tems. On one side, the system performance strongly depends on the available knowl-
edge of CSI. On the other side, suitable communication strategies based on different
levels of knowledge of CSI can be chosen. The system concept of cooperative com-
munication in multiuser MIMO systems is proposed as shown in Fig. 5.32. Diversity
techniques can be applied when no knowledge of CSI is available. Considering the
influence of imperfectness of CSI in realistic systems, adaptive scheduling techniques
are applied to select the users in each subcarrier. Applying the OFDM transmis-
sion technique, in the considered subcarrier significant CSI is selected and applied
in JD and JT. In this way, cooperative communication with partial CSI concerning
imperfect CSI and significant CSI can maintain significant useful contributions and
eliminate significant interference with reduced communication load and computa-
tional load.
5.4.8 Outlook
The main focus of our work is the interference management in multiuser cellular
systems from the interference cancellation signal processing point of view. Other
interference management strategies such as resource allocation and power control can
further reduce the influence of interference. In future work, various combinations
of cooperative communication considering JD/JT, resource allocation, and power
control can be investigated to jointly improve the performance of interference-limited
cellular systems.
212 5 System Level Aspects for Multiple Cell Scenarios
Cooperative reception/transmission with only partial CSI is of high interest for
practical realization in future mobile radio cellular networks. Based on our con-
tributions about significant CSI selection, various criteria of significant CSI can be
further investigated and applied in the realistic cellular systems. Since researchers
are paying more and more attention to the cooperative communication concept
with selected CSI [8, 9,19–23], in future work different proposals can be compared.
Further investigations could indicate which criterion for significant CSI selection is
more beneficial to system performance and which one is more beneficial to imple-
mentation complexity. Combinations of different proposals and improvements of the
significant CSI selection algorithms are expected. Some concrete design guidelines
of the selection algorithms could also be proposed.
Advanced nonlinear JD/JT algorithms, i.e., SIC with quantization in the UL and
THP in the DL, with various significant CSI selection algorithms can be further
investigated with respect to analytical and numerical results in future work. Dis-
tributed signal processing of these algorithms in the cooperative communication
scheme could be implemented, and its implementation complexity can be assessed.
Statistical signal processing based on limited knowledge of CSI has also been inves-
tigated based on a real-valued two user interference channel model. Its application
can be extended to realistic multiuser cellular systems in future work.
Bibliography
[1] X. You, G. Chen, M. Chen, and X. Gao, “Toward beyond 3G: The FuTURE
project in China,” IEEE Communications Magazine, vol. 43, no. 1, pp. 38–44,
Jan. 2005.
[2] K. Tachikawa, “A perspective on the evolution of mobile communications,”
IEEE Communications Magazine, vol. 41, no. 1, pp. 66–73, Oct. 2003.
[3] G. J. Foschini and M. J. Gans, “On limits of wireless communications in a fading
environment when using multiple antennas,” Wireless Personal Commun.,vol.
6, pp. 311–335, Mar. 1998.
[4] E. Telatar, “Capacity of multi-antenna Gaussian channels,” European Trans-
actions on Telecommunications, vol. 10, no. 6, pp. 585–595, 1999.
[5] J. G. Andrews, “Interference cancellation for cellular systems: A contemporary
overview,” IEEE Wireless Communications, vol. 12, pp. 19–29, April 2005.
[6] R. D. J. van Nee and R. Prasad, OFDM for Wireless Multimedia C ommunica-
tions, Artech House, Boston, 2000.
[7] T. Weber, I. Maniatis, A. Sklavos, and Y. Liu, “Joint transmission and detection
integrated network (JOINT), a generic proposal for beyond 3G systems,” in
Proc. 9th International Conference on Telecommunications (ICT’02),Beijing,
June 2002, vol. 3, pp. 479–483.
Bibliography 213
[8] P. Zhang, X. Tao, J. Zhang, Y. Wang, L. Li, and Y. Wang, “A vision from the
future: Beyond 3G TDD,” IEEE Communications Magazine, vol. 43, no. 1, pp.
38–44, January 2005.
[9] X. Tao, J. Xu, X. Xu, C. Tang, and P. Zhang, “Group cell FuTURE B3G
TDD system,” in Proc. IEEE International Symposium on Personal, Indoo r
and Mobile Radio Communications (PIMRC’05).
[10] A. Klein, G. K. Kaleh, and P. W. Baier, “Zero forcing and minimum mean-
square-error equalization for multiuser detection in code-division multiple access
channels,” IEEE Transactions on Vehicular Technology, vol. 45, no. 2, pp. 276–
287, May 1996.
[11] M. Meurer, P. W. Baier, T. Weber, Y. Lu, and A. Papathanassiou, “Joint
transmission: Advantageous downlink concept for CDMA mobile radio systems
using time division duplexing,” Electronics Letters, vol. 36, no. 10, pp. 900–901,
May 2000.
[12] T. Weber, A. Sklavos, and Meurer M., “Imperfect channel state information in
MIMO transmission,” IEEE Transactions on Communications, vol. 54, no. 3,
pp. 543–552, March 2006.
[13] S. Verdu, Multiuser Detection, University Press, Cambridge, 1998.
[14] A. Kühne and A. Klein, “Adaptive MIMO-OFDM using OSTBC with imperfect
CQI feedback,” in Proc. International ITG/IEEE Workshop on Smart Antennas
(WSA’08), Darmstadt, Feb. 2008.
[15] A. Kühne and A. Klein, “Throughput analysis of multi-user OFDMA-systems
using imperfect CQI feedback and diversity techniques,” IEEE Journal on
Selected Areas in Communications, vol. 26, no. 8, pp. 1440–1451, October 2008.
[16] X. Wei, T. Weber, A. Kühne, and A. Klein, “Joint transmission with imper-
fect partial channel state information,” in Proc. IEEE Vehicular Technology
Conference (VTC’09-Spring), Barcelona, April 2009.
[17] X. Wei, T. Weber, A. Ahrens, and S. Deng, “Decentralized interference man-
agement in mobile radio networks,” Frequenz, vol. 61, no. 11/12, pp. 259–269,
2007.
[18] X. Wei, and T. Weber, “Cooperative communication with partial channel-state
information in multiuser MIMO systems,” AEÜ , 2010.
[19] A. Papadogiannis, H.J. Bang, D. Gesbert, and E. Hardouin, “Downlink over-
head reduction for multicell cooperative processing enabled wireless networks,”
in Proc. IEEE International Symposium on Personal, Indoor, Mobile Radio
Communications (PIMRC’08), Cannes, September 2008.
[20] A. Papadogiannis, D. Gesbert, and E. Hardoin, “A dynamic clustering approach
in wireless networks with multi-cell cooperative processing,” in Proc. IEEE
International Conference on Communications (ICC’08), Beijing, May 2008.
214 5 System Level Aspects for Multiple Cell Scenarios
[21] P. Marsch and G. Fettweis, “A framework for optimizing the uplink performance
of distributed antenna systems under a constrained backhaul,” in Proc. IEEE
International Conference on Communications (ICC’07), Glasgow, June 2007.
[22] P. Marsch and G. Fettweis, “On multicell cooperative transmission in backhaul-
constrained cellular systems,” Annals of Telecommunications, vol. 63, no. 5–6,
pp. 253–269, June 2008.
[23] S. Khattak and G. Fettweis, “Low backhaul distributed detection strategies for
an interference limited uplink cellular system,” in Proc. IEEE 67th Vehicular
Technology Conference (VTC’08-Spring), Singapore, May 2008, pp. 693–697.
H. Rohling (ed.), OFDM: Concepts for Future Communication Systems, 215
Signals and Communication Technology, DOI: 10.1007/978-3-642-17496-4_6,
© Springer-Verlag Berlin Heidelberg 2011
6 OFDM/DMT for Wireline
Communications
With wireline, we mean digital subscriber lines as well as optical transmission. Apart
from the general aspects of multicarrier modulation, the channel properties are very
different. It is thus unavoidable to go into some details of channel transfer character-
istics and disturbances. In DSL, the channel characteristics are determined by the
propagation constant, the characteristic impedance, and the loop structure, further-
more crosstalk from other loops and ingress (RFI, impulse noise) from outside. In
copper loops, the frequency dependent transfer function of the twisted pairs together
with possible reflections at splices lead to dispersion which we will also see in optical
transmission. There, we distinguish chromatic dispersion and polarization mode dis-
persion. Optical transmission especially also suffers from non-linear characteristics
(Kerr effect), which does not have a counterpart in copper cables. In copper trans-
mission, non-linearities are especially due to the limitations of the D/A converters
and power amplifiers, which may lead to clipping of the almost Gaussian distributed
time-domain signal of OFDM. This holds for all OFDM transmission. Other non-
linearities may be due to baluns (transformers) used in DSL, but practically, they
are of minor relevance.
In the following two sections, we will study properties of the channel and specialties
of the multicarrier transmission over twisted pairs and optical fibers. Thereafter, we
consider a new approach for impulse noise cancellation in DSL and the simulation
of optical multicarrier transmission.
6.1 Discrete MultiTone (DMT) and Wireline
Channel Properties
W. Henkel, Jacobs University Bremen, Germany
This chapter describes basic properties of the wireline twister-pair channel and
introduces DMT (Discrete MultiTone), the baseband variant of OFDM. Further
information can, e.g., be found in a student book project [1], or [2–5].
6.1.1 Properties of the Twisted-Pair Channel
The twisted pair (TP) cable channel is characterized by a transfer function that is
increasingly attenuating with frequency and cross-talk functions that increase with
frequency, as well. Furthermore, ingress from radio interferers (RFI), e.g., amateur
radio, and impulse noise have a strong impact. In the following, we describe the