OPTICAL MIMO MULTIMODE FIBER LINKS
Channel Measurements and System Performance Analysis
Andreas Ahrens, Jens Pankow, Sebastian Aust and Steffen Lochmann
Hochschule Wismar, University of Technology, Business and Design, Philipp-M¨uller-Straße 14, 23966 Wismar, Germany
Keywords:
Multiple-Input Multiple-Output System, Singular-value Decomposition, Bit allocation, Wireless transmission,
Optical fibre transmission, Multimode fiber.
Abstract:
Wireless communication is nowadays one of the areas attracting a lot of research activity due to the strongly
increasing demand in high-data rate transmission systems. The use of multiple antennas at both the transmitter
and receiver side has stimulated one of the most important technical breakthroughs in recent communications
allowing increasing the capacity and dropping the bit-error rate. However, multiple-input multiple-output
(MIMO) systems are not limited to wireless MIMO systems and can be observed in a huge variety of trans-
mission links and network parts and have attracted a lot of attention since the mid 90’s. In the field of optical
MIMO transmission systems, multi-mode (MM) fibre offers the possibility to transmit different signals by dif-
ferent mode groups. The perspective of the MIMO philosophy within the field of optical transmission systems
is elaborated in this contribution based on channel measurements within a (2 × 2) MIMO system. For the
channel measurements the second optical window and a fibre length of 1, 4 km was chosen. Computer simu-
lations on an overall data rate of 10,24 Gbps underline the potential of multi-mode fibres in optical high-data
rate MIMO communication systems and show that in order to achieve the best bit-error rate, not necessarily
all MIMO layers have to be activated.
1 INTRODUCTION
The increasing desire for communication and infor-
mation interchange has attracted a lot of research
since Shannons pioneering work in 1948. A possible
solution was presented by Teletar and Foschini in the
mid 90’s, which revived the MIMO (multiple-input
multiple-output) transmission philosophy introduced
by van Etten in the mid 70’s (Telatar, 1999), (Fos-
chini, 1996), (van Etten, 1975), (van Etten, 1976).
Since the capacity of MIMO systems increases
linearly with the minimum number of antennas at
both the transmitter as well as the receiver side, wire-
less MIMO schemes have attracted substantial atten-
tion (McKay and Collings, 2005) and can be con-
sidered as an essential part of increasing both the
achievable capacity and integrity of future genera-
tions of wireless systems (K¨uhn, 2006), (Zheng and
Tse, 2003). However, the MIMO technique isn’t lim-
ited to wireless communication and a lot of scenar-
ios can be described and outperformed by the MIMO
technique.
In comparison to the wireless MIMO channel,
the optical fibre is an important type of a fixed-line
medium, which is used in several sections of telecom-
munication networks, where single- and multi-mode
fibres are distinguished (Singer et al., 2008), (Winters
and Gitlin, 1990).
Optimizing the transmission on high-data rate
links is in particular of great practical interest for
delivering voice or video services in mobile IP (In-
ternet Protocol) based networks in the access do-
main. Unfortunately, the inherent modal dispersion
limits the maximum data speed within the multimode
fiber (MMF). In order to overcome this limitation,
the well-known single-input single-output systems,
also called SISO systems, should be transferred into
systems with multiple-inputs and multiple-outputs,
also called MIMO systems (Hsu and Tarighat, 2006),
(Singer et al., 2008). Taking finally into account that
delay-spread in wireless broadband MIMO transmis-
sion systems isn’t any longer a limiting parameter,
MMF links should be well suited for high-speed data
transmission (Raleigh and Cioffi, 1998), (Raleigh and
Jones, 1999).
Different research groups, e.g. (Sch¨ollmann
and Rosenkranz, 2007), (Sch¨ollmann et al., 2008),
have adapted the MIMO technique on optical com-
128
Ahrens A., Pankow J., Aust S. and Lochmann S..
OPTICAL MIMO MULTIMODE FIBER LINKS - Channel Measurements and System Performance Analysis.
DOI: 10.5220/0003437201280132
In Proceedings of the International Conference on Data Communication Networking and Optical Communication System (OPTICS-2011), pages
128-132
ISBN: 978-989-8425-69-0
Copyright
c
2011 SCITEPRESS (Science and Technology Publications, Lda.)
munication channels. The experimental equaliza-
tion of crosstalk within frequency non-selective op-
tical MIMO systems has attracted a lot of research
(Sch¨ollmann and Rosenkranz, 2007), (Sch¨ollmann
et al., 2008). By contrast, frequency selective
MIMO links require substantial further research,
where spatio-temporal vector coding (STVC) intro-
duced by RALEIGH for wireless MIMO channels
seems to be an appropriate candidate for optical trans-
mission channels too (Raleigh and Cioffi, 1998),
(Raleigh and Jones, 1999).
In this contribution the spatial multiplexing (SM)
is implemented at the transmitter side via different
sources launching light with different offsets into the
fibre. At the receiver side, i.e. at the fibre end, various
spatial filters are implemented (Pankow et al., 2011).
By launching light with different offsets into the fi-
bre, different mode groups are activated, which prop-
agate along the fibre with different speed and atten-
uation. Together, with the crosstalk between the dif-
ferent mode groups, the classical MIMO channel is
formed, where the most beneficial choice of the num-
ber of activated MIMO layers and the number of bits
per symbol offer a certain degree of design freedom,
which substantially affects the performance (Ahrens
and Benavente-Peces, 2009).
Against this background, the novel contribution
of this paper is that based on channel measurements
within a (2 × 2) optical MIMO system, the perspec-
tive of the MIMO philosophy within the field of op-
tical transmission systems is elaborated. Our results
show that even for relatively long (e. g. 1,4 km) trans-
mission lengths high data rates (e. g. 10,24 Gbps)
are feasible and that the choice of the number of bits
per symbol and the number of activated MIMO layers
substantially affects the performance of a MIMO sys-
tem, suggesting that not all MIMO layers have to be
activated in order to achieve the best BER.
The remaining part of this paper is organized as
follows: Section 2 introducesthe optical MIMO chan-
nel. The crosstalk impact within the optical MIMO
channel is studied in section 3, while the associated
performance results are presented and interpreted in
section 4. Finally, section 5 provides some conclud-
ing remarks.
2 OPTICAL MIMO CHANNEL
In order to comply with the demand on increas-
ing available data rates, systems with multiple in-
puts and multiple outputs, also called MIMO sys-
tems (multiple-input multiple-output), have become
indispensable and can be considered as an essential
10µm
10µm
TX
2
TX
1
ξ
RX
2
RX
2
RX
1
r
Figure 1: Forming the optical MIMO channel (left: light
launch positions at the transmitter side with a given eccen-
tricity ξ, right: spatial configuration at the receiver side as a
function of the mask radius r).
part of increasing both the achievable capacity and
integrity of future generations of communication sys-
tems (K¨uhn, 2006), (Zheng and Tse, 2003).
In this work the potential of the MIMO philoso-
phy in optical channels is elaborated, based on chan-
nel measurements. Basically, light launched at differ-
ent positions within the fibre activates different mode
groups, which propagate along the fibre with differ-
ent speed. Low order mode groups, activated by light
launched into the center of the fibre, lead to a power
radiation pattern concentrated at the center of the fiber
end whereas higher order modes, activated by light
launched at an off-center position, e. g., with a given
eccentricity ξ, within the fibre, lead to power radiation
pattern concentrated at the off-center of the fiber-end.
Therefore, by launching light into the fibre with given
eccentricities, as highlighted in Fig. 1, different spa-
tially separated power distribution pattern can be ob-
tained at the receiver side to form the optical MIMO
channel.
According to Fig. 1, the optical input TX
1
was ad-
justed to launch light into the center of the core (center
launch condition), whereas for the optical input TX
2
a given eccentricity ξ was chosen (off-center launch
condition). The activated modes can be separated at
the fibre end by the corresponding power distribu-
tion pattern. Fig. 2 illustrates the simulated power
distribution pattern by activating low- and high-order
modes. The simulations are in good agreement with
the measured power radiation pattern as depicted in
Fig. 3 for different parameters of the eccentricity ξ.
Now, spatial ring filters at the end of the transmission
line as depicted in Fig. 1 have been applied for chan-
nel separation. These spatial filters have been pro-
duced by depositing a metal layer at fiber end-faces
and subsequent ion milling (Pankow et al., 2011),
(Pankow et al., 2010).
Together, with the crosstalk between the different
mode groups, the classical MIMO channel is formed
(Fig. 4) (Pankow et al., 2011), (Pankow et al., 2010).
OPTICAL MIMO MULTIMODE FIBER LINKS - Channel Measurements and System Performance Analysis
129
Figure 2: Simulated mean power distribution pattern (left:
by the LP
01
mode, right: by activating all solutions of LP
81
modes); the dotted line represents the 50µm core size.
Figure 3: Measured mean power distribution pattern as
a function of the light launch position (left: eccentricity
ξ = 0µm, right: eccentricity ξ = 18µm); the dotted line rep-
resents the 50µm core size.
u
s 1
(t)
u
s 2
(t)
u
k 1
(t)
u
k 2
(t)
g
11
(t)
g
21
(t)
g
12
(t)
g
22
(t)
Figure 4: Electrical MIMO system model (example: n = 2).
3 CROSSTALK IMPACT
In this section it is studied, how the crosstalk impact
depends on the MIMO system configuration. There-
fore at this point the eccentricity of the transmitter
side MIMO configuration as well as the mask ra-
dius of the ring filter configuration at the receiver side
are investigated in an exemplary system according to
Fig. 5. It is assumed, that each MMF input is fed
by a system with identical mean properties with re-
spect to transmit filtering, pulse frequency f
T
= 1/T
s
,
the number of signalling levels and the mean transmit
power P
s
.
u
q 1
(t)
u
q 2
(t)
G
s
(f)
G
s
(f)
u
s 1
(t)
u
s 2
(t)
G
11
(f)
G
12
(f)
u
k 1
(t)
u
k 11
(t)
u
k 12
(t)
Figure 5: Electrical system model of transmitter and MMF
with crosstalk (example: n = 2).
The source signals u
q1
(t) and u
q2
(t) traverse the
transmit filters with the transfer function G
s
( f). Then
the wanted transmit signal u
s1
(t) passes the MMF,
modelled by the transfer function G
11
( f) and causes
the signal u
k11
(t) with power P
k11
at the MMF out-
put, whereas the crosstalk signal u
k12
(t) (with power
P
k12
), which is fed into the MMF with a given ec-
centricity ξ, originates at the MMF output, after the
transmit signal u
s2
(t) passed the filter with the trans-
fer function G
12
( f), which models the coupling from
optical input 2 to the output 1 (Fig. 5).
In general, the MIMO performance is affected by
both the mask radius r and the eccentricity ξ. As
highlighted in Fig. 6, the power of the wanted signal
u
k11
(t) at the MMF output, distributed in the inner
ring, increases monotonically with rising mask radius
r, whereas at the same time the power of the wanted
signal u
k22
(t), distributed in the outer ring, decreased
with increasing mask radius and increased eccentric-
ity. In addition a channel asymmetry can be observed
which is caused by the larger differential mode atten-
uation of the higher order mode groups. From this
point of view it can be concluded that a mask radius
in the range of 5µm to 15µm should be chosen in order
to have adequate power at both outputs.
From a practical point of view the power P
k12
(ξ,r)
of the crosstalk signal u
k12
(t) at the MMF output is an
interesting indicator for the strength of the crosstalk
disturbance, which depends on the mask radius r
and the eccentricity ξ. In order to assess the effect
of crosstalk on the wanted signal not only the pure
crosstalk signal power is of interest, but rather the be-
haviour of the powers of the wanted signal and the
crosstalk signal to each other. This behaviour may be
investigated by a signal-to-crosstalk-interferenceratio
(SCIR)
ρ
k11
(ξ,r) =
P
k11
(0,r)
P
k12
(ξ,r)
and ρ
k22
(ξ,r) =
P
k22
(ξ,r)
P
k21
(0,r)
.
(1)
Since MIMO makes use of the interference for chan-
nel improvement the SCIR should not be chosen as
high as possible like in orthogonal transmission. Re-
ferring to Fig. 7 this means for lower order modes
OPTICS 2011 - International Conference on Optical Communication Systems
130
0 5 10 15 20 25
0
0.2
0.4
0.6
0.8
1
r (inµm)
normalized output power
P
k11
(0µm, r)/P
max
P
k22
(10 µm,r)/P
max
P
k22
(14 µm,r)/P
max
P
k22
(18 µm,r)/P
max
Figure 6: Measured electrical signal power P
kνµ
(ξ,r) at the
MMF output as a function of the mask radius r for given
parameters of the eccentricity ξ.
0 5 10 15 20 25
−10
0
10
20
30
r (inµm)
10 · lg(ρ
k
) (in dB)
P
k
11
(0µm,r)/P
k
12
(14 µm,r)
P
k
22
(14 µm,r)/P
k
21
(0,r)
P
k
11
(0µm,r)/P
k
12
(10 µm,r)
P
k
22
(10 µm,r)/P
k
21
(0,r)
Figure 7: Electrical signal-to-crosstalk-interference ratio
(SCIR) ρ
k
at the MMF outputs as a function of the mask
radius r and the eccentricity ξ.
(output 1) a movement towards larger mask radiuses
and vice versa for higher order mode groups (output
2).
Though a relationship between the spatial mode
location and the channel’s impulse response needs to
be established for an exact BER (bit-error rate) trade-
off it can already been concluded from Fig. 6 and
Fig. 7 that mask radiuses in the range of about 5µm
to 15µm should be used for further BER analyses.
4 PERFORMANCE ANALYSIS
For BER analysis, a frequency selective SDM (spa-
tial division multiplexing) MIMO link, composed of
n
T
inputs and n
R
outputs, is considered. The block-
oriented system for frequency selective channels is
modelled by
u = H· c+ w . (2)
In (2), the transmitted signal vector c is mapped
by the channel matrix H onto the received vector
u. Finally, the vector of the additive, white Gaus-
sian noise (AWGN) is defined by w (Pankow et al.,
2011), (Ahrens and Benavente-Peces, 2009). The in-
terference between the different input’s data streams,
which is introduced by the off-diagonal elements of
the channel matrix H, requires appropriate signal
processing strategies. A popular technique is based
on the singular-value decomposition (SVD) (Haykin,
2002) of the system matrix H, which transfers the
whole system into independent, non-interferinglayers
having unequal gains (Pankow et al., 2011), (Ahrens
and Benavente-Peces, 2009).
In this contribution the efficiency of fixed trans-
mission modes is studied regardless of the channel
quality. Assuming predefined transmission modes, a
fixed data rate can be guaranteed.
For numerical analysis it is assumed, that each
optical input within the MMF is fed by a system
with identical mean properties with respect to trans-
mit filtering and pulse frequency f
T
= 1/T
s
. Within
this work, the pulse frequency f
T
is chosen to be
f
T
= 5, 12 GHz. The average transmit power is sup-
posed to be P
s
= 1V
2
and as an external disturbance a
white Gaussian noise with a power spectral density N
0
is assumed. In order to transmit at a fixed data rate, an
Table 1: Investigated transmission modes.
throughput layer 1 layer 2
2 bit/s/Hz 4 0
2 bit/s/Hz 2 2
appropriate number of MIMO layers has to be used,
which depends on the specific transmission mode, as
detailed in Tab. 1 for the investigated (2 × 2) optical
MIMO system.
The obtained BER curves are depicted in Fig. 8 for
the different ASK (amplitude shift keying) constella-
tion sizes of Tab. 1. For the investigations, an eccen-
tricity of ξ = 10µm and a mask radius of r = 15µm
was assumed, which was found to be beneficial for
minimizing the overall BER at a fixed data rate. As-
suming a uniform distribution of the transmit power
over the number of activated MIMO layers, it turns
out that not all MIMO layers have to be activated in
order to achieve the best BERs.
5 CONCLUSIONS
In this paper the perspective of the MIMO philosophy
OPTICAL MIMO MULTIMODE FIBER LINKS - Channel Measurements and System Performance Analysis
131
20 25 30 35 40
10
−6
10
−4
10
−2
10
0
10 · log
10
(P
s
T
s
/N
0
) (indB)
bit-error rate
4 ASK, SISO
(4,0) ASK, MIMO
(2,2) ASK, MIMO
Figure 8: BER when using the transmission modes intro-
duced in Tab. 1 and transmitting 2 bit/s/Hz over frequency
selective optical MIMO channels.
within the field of optical transmission systems is in-
vestigated. Our results, obtained by channel measure-
ments and computer simulations, show the potential
of MIMO techniques in the field of optical transmis-
sion systems. In combination with appropriateMIMO
signal processing strategies, an improvement in the
overall BER was obtained.
ACKNOWLEDGEMENTS
The authors wish to thank their co-worker, Mr. Ralph
Bornitz, for supporting the measurement campaign.
REFERENCES
Ahrens, A. and Benavente-Peces, C. (2009). Modulation-
Mode and Power Assignment in Broadband MIMO
Systems. Facta Universitatis (Series Electronics and
Energetics), 22(3):313–327.
Foschini, G. J. (1996). Layered Space-Time Architecture
for Wireless Communication in a Fading Environment
when using Multiple Antennas. Bell Labs Technical
Journal, 1(2):41–59.
Haykin, S. S. (2002). Adaptive Filter Theory. Prentice Hall,
New Jersey.
Hsu, R. and Tarighat, A. (2006). Capacity Enhance-
ment in Coherent Optical MIMO (COMIMO) Mul-
timode Fiber Links. IEEE Communications Letters,
10(3):195–197.
K¨uhn, V. (2006). Wireless Communications over MIMO
Channels Applications to CDMA and Multiple An-
tenna Systems. Wiley, Chichester.
McKay, M. R. and Collings, I. B. (2005). Capacity
and Performance of MIMO-BICM with Zero-Forcing
Receivers. IEEE Transactions on Communications,
53(1):74– 83.
Pankow, J., Ahrens, A., and Lochmann, S. (2010). Channel
Measurements and Performance Analysis of Optical
MIMO Multimode Fiber Links. In 4th Baltic Con-
ference Learning in Networks, pages 69–78, Kaunas
(Lithuania).
Pankow, J., Aust, S., Lochmann, S., and Ahrens, A. (2011).
Modulation-Mode Assignment in SVD-assisted Op-
tical MIMO Multimode Fiber Links. In 15th Inter-
national Conference on Optical Network Design and
Modeling (ONDM), Bologna (Italy).
Raleigh, G. G. and Cioffi, J. M. (1998). Spatio-Temporal
Coding for Wirleess Communication. IEEE Transac-
tions on Communications, 46(3):357–366.
Raleigh, G. G. and Jones, V. K. (1999). Multivariate
Modulation and Coding for Wireless Communication.
IEEE Journal on Selected Areas in Communications,
17(5):851–866.
Sch¨ollmann, S. and Rosenkranz, W. (2007). Experimen-
tal Equalization of Crosstalk in a 2 x 2 MIMO Sys-
tem Based on Mode Group Diversity Multiplexing
in MMF Systems @ 10.7 Gb/s. In 33rd European
Conference and Exhibition on Optical Communica-
tion (ECOC), Berlin.
Sch¨ollmann, S., Schrammar, N., and Rosenkranz, W.
(2008). Experimental Realisation of 3 x 3 MIMO Sys-
tem with Mode Group Diversity Multiplexing Limited
by Modal Noise. In Optical Fiber Communication
Conference (OFC), San Diego, California.
Singer, A. C., Shanbhag, N. R., and Bae, H. M. (2008).
Electronic Dispersion Compensation– An Overwiew
of Optical Communications Systems. IEEE Signal
Processing Magazine, 25(6):110 – 130.
Telatar, E. (1999). Capacity of Multi-Antenna Gaussian
Channels. European Transactions on Telecommuni-
cations, 10(6):585–595.
van Etten, W. (1975). An Optimum Linear Receiver
for Multiple Channel Digital Transmission Systems.
IEEE Transactions on Communications, 23(8):828–
834.
van Etten, W. (1976). Maximum Likelihood Receiver for
Multiple Channel Transmission Systems. IEEE Trans-
actions on Communications, 24(2):276–283.
Winters, J. and Gitlin, R. (1990). Electrical Signal Process-
ing Techniques in Long-Haul Fiber-Optic Systems.
IEEE Transactions on Communications, 38(9):1439–
1453.
Zheng, L. and Tse, D. N. T. (2003). Diversity and
Multiplexing: A Fundamental Tradeoff in Multiple-
Antenna Channels. IEEE Transactions on Information
Theory, 49(5):1073–1096.
OPTICS 2011 - International Conference on Optical Communication Systems
132