534 IEEE ANTENNAS AND WIRELESS PROPAGATION LETTERS, VOL.

5, 2006

The Role of Polarization Diversity for MIMO

Systems Under Rayleigh-Fading Environments
Juan F. Valenzuela-Valdés, Miguel A. García-Fernández, Antonio M. Martínez-González, and
David Sánchez-Hernández, Senior Member, IEEE

Abstract—Polarization diversity techniques have not received polarization states of the EM field through colocation instead of
as much attention as others due to the significant difference more voluminous spatially separated arrays can provide equiv-
in mean signal level between copolarized and cross-polarized alent channel capacity increase [3]. The slanted polariza-
branches when one polarization is transmitted. However, mul-
tiple-input–multiple-output (MIMO) systems where the multipath tion diversity combination was found to perform just as good as
fading is only partially correlated could use polarization diversity the spatial polarization diversity with two elements, and com-
to provide a high diversity gain. Hence there is a need to fully mercial systems like GSM and UMTS soon switched to this new
understand the role of true polarization diversity in such systems. technique. Some combinations of two-branch polarization and
In this letter, progressive (true) polarization diversity performance spatial diversity have also been reported [4]. Recently, a triaxial
for 3 3 MIMO systems under Rayleigh-fading environments is
evaluated through simulations and measurements. True polariza- combination of polarization and pattern diversity has also been
tion diversity was found to be as significant as spatial diversity at proposed [5], with some contradicting results.
improving diversity gain, and hence MIMO system capacity. In mobile communications scenarios, however, multiple scat-
Index Terms—Diversity gain, multiple-input–multiple-output tering may not be sufficient for a given polarization to decouple
(MIMO) systems, polarization diversity. half its power into the orthogonal polarization [6]. Channel
behavior is therefore different for different polarization states
[6] and fading cross-correlations may even increase the ergodic
I. INTRODUCTION capacity beyond the case of independent channels [7]. Hence,
diversity research results based upon orthogonal polarization
P OLARIZATION diversity, early applied to HF, radar, and
imaging systems, has demonstrated its potential for im-
proving the capacity of wireless communications systems, de-
states cannot be generalized to arbitrary (true) polarization
states for more general TxR MIMO systems. On the other hand,
spite some disappointing premature predictions [1]. The im- MIMO systems do not require perfectly uncorrelated signals
provement is typically granted by an additional decorrelated to perform near full capacity, and more than two orthogonal
channel provided by a polarization state made orthogonal to polarizations could be used efficiently [8]. Yet, the full potential
the existing one, usually at the transmitting end .A of true polarization diversity where any polarization state could
randomly orientated linearly-polarized antenna is also typically be employed is not fully addressed in the literature and requires
used at the receiver for evaluating polarization diver- more research [9].
sity. Consequently, the cross-polarization discrimination (XPD) In this letter, the role of true polarization diversity for
factor is the usual evaluating parameter, with low correlation MIMO systems under Rayleigh-fading scenarios is investigated
coefficients being achieved even in NLOS situations [2]. Due to through simulations and measurements in a reverberation
the significant difference in mean received signal level between chamber. Since true polarization diversity can be successfully
copolarized and cross-polarized branches when one polariza- combined to other diversity schemes, which could be limited
tion is transmitted, considerably more attention has been paid due to spatial or coherent bandwidth restrictions, the results
to spatial diversity. The (de)coupling effect between different presented in this paper are of particular importance for future
polarizations is a complex mechanism to be simulated, which diversity-combined MIMO systems.
has also limited true polarization diversity research. Yet, since
at least horizontal and vertical separation distances are
required for efficient spatial outdoor diversity in practice, polar- II. SIMULATION AND MEASUREMENTS
ization diversity has recently gained attraction. The use of vector
antennas which can respond to more than one component and/or The multipath environment can be generated artificially in
a reverberation chamber fed by wall-mounted antennas, which
thereby provides a statistically repeatable laboratory-produced
Manuscript received September 5, 2006; revised November 15, 2006. This environment for characterizing mobile terminals and antennas,
work was supported in part by Fundación Séneca, the R&D coordinating well described in the literature. The processed S-parameters rep-
unit of the Autonomous Region of Murcia (Spain) under project reference
2I05SU0033. The authors wish to thank Bluetest AB for its generous educa- resent estimates of the matrix H of multipath communication
tional discounting. channels set up between the wall antennas and the MIMO array
The authors are with the Departamento de Tecnologías de la Información inside the chamber. Apparent diversity gain for the selection
y Comunicaciones, Universidad Politécnica de Cartagena, Cartagena E-30202
Spain (e-mail: [email protected]; http://www.gimre.upct.es). combining (SC) or maximal ratio combining (MRC) techniques
Digital Object Identifier 10.1109/LAWP.2006.889552 is obtained from the processed S-parameters by evaluating the
1536-1225/$20.00 © 2006 IEEE

Authorized licensed use limited to: University of Electronic Science and Tech of China. Downloaded on September 03,2020 at 07:14:43 UTC from IEEE Xplore. Restrictions apply.
 VALENZUELA-VALDÉS et al.: POLARIZATION DIVERSITY FOR MIMO SYSTEMS 535

Fig. 1. Geometries of analyzed MIMO scenarios.

cumulative probability distributions of the measured channel

samples received at each MIMO array antenna

Fig. 2. Spatial separation D (d=) and angular separation d ( ) versus

(1) measured diversity gain with T = 3, R = 3. 5Angular separation
( ) with D = 0;  Spatial separation (d=) with d = 0.
where is the index for the transmitting antenna, is the index
for the receiving antenna, is the net chamber transfer func-
tion for the reference antenna and is its radiation efficiency.
The effective diversity gain is the increase from the reference
level to the combined signal that is observed at the 1% proba-
bility level.
All measurements illustrated in this paper have been per-
formed with the RC800 reverberation chamber by Bluetest AB,
with dimensions of , three wall-mounted
transmitting antennas, 25 platform stirring positions, two me-
chanical stirrers with 15 different positions for each platform
position and 100 MHz frequency stirring. Measurements were
performed at 900 MHz and 3 dipoles were used as receiving
MIMO array antennas following the setup depicted in Fig. 1.
SC diversity gain increments were evaluated by rotating a cer-
tain angle an antenna with regard to a contiguous one, namely,
by applying a progressive angular separation between elements, Fig. 3. Measured diversity gain versus Dipole separation D(d=) with angular
and compared to measured SC spatial diversity performance and separation d ( ) as a parameter and T = 3, R = 3. 5 d = 0 ;
their simulated MIMO Rayleigh-fading channels counterparts. d = 9 ; 2 d = 18;  d = 27; 
Polarization diversity was applied at both transmitting and re- d = 36.
ceiving ends.
Fig. 2 illustrates a comparison between spatial and polariza-
good combination of the two diversity techniques with both low
tion diversity for 20 different measured 3 3 MIMO systems,
spatial and angular separation values represents the most effi-
showing wavelength-normalized spatial separation D and
cient method for optimum diversity performance for the same
angular separation versus measured diversity gain with
reduced available volume.
and . From this figure it is clearly observed that
an equivalence scenario is obtained for the two diversity tech-
niques. As an example, a spatial separation of is equiv- III. CONCLUSION
alent to an angular separation of for the 3 3 MIMO The full potential of spatial and true polarization diversity ob-
systems under evaluation. tained in a set of rotational linearly polarized elements in both
In order to evaluate the full potential of combined-diversity transmitting and receiving ends of 3 MIMO systems has been
techniques, SC diversity gain measurements were performed for investigated. Measured results have demonstrated that spatial
three receiving dipoles separated by both an angular and a spa- diversity can provide considerable gain to a MIMO system. We
tial separation. Fig. 3 shows the measured diversity gain versus have also demonstrated that polarization diversity is equally im-
angular separation with spatial separation D as a parameter. portant and particularly interesting for volume-limited schemes
As expected, the combination of both spatial and polarization where the full potential of spatial diversity techniques cannot be
diversity provided increased diversity gain with only three el- exploited.
ements in the array. When the spatial separation is large, the Future research should concentrate on more general TxR
angular separation can hardly improve the diversity gain. Sim- MIMO systems, system capacity and diverse fading environ-
ilarly, when the angular separation is large, the spatial separa- ments to investigate the full capabilities of rotational polariza-
tion can barely improve the diversity gain. This suggests that a tion diversity schemes and combined-diversity methods.

Authorized licensed use limited to: University of Electronic Science and Tech of China. Downloaded on September 03,2020 at 07:14:43 UTC from IEEE Xplore. Restrictions apply.
 536 IEEE ANTENNAS AND WIRELESS PROPAGATION LETTERS, VOL. 5, 2006

REFERENCES [5] M. C. Tumbuka and D. J. Edwards, “Investigation of tri-polarized

MIMO technique,” Electron. Lett., vol. 41, no. 3, pp. 137–138, Feb.
[1] M. S. Smith and L. Neal, “A comparison of polarization and space di-
2005.
versity for indoor propagation at 900 MHz,” in Proc. 2nd Int. Conf. Uni- [6] R. G. Vaughan, “Polarization diversity in mobile communications,”
versal Personal Communications: Gateway to 21st Century, vol. 1, Oct.
IEEE Trans. Veh. Technol., vol. 39, pp. 177–186, Aug. 1990.
1993, pp. 74–78. [7] C. Oestges and A. J. Paulraj, “Beneficial impact of channel correlations
[2] J. J. A. Lempiainen and J. K. Laiho-Steffens, “The performance of po- on MIMO capacity,” Electron. Lett., vol. 40, no. 10, pp. 606–608, May
larization diversity schemes at a base station in small/micro cells at 1800
2004.
MHz,” IEEE Trans. Veh. Technol., vol. 47, no. 3, pp. 1087–1092, Aug. [8] C. Waldschmidt, C. Kuhnert, T. Fügen, and W. Wiesbeck, “Measure-
1998.
ments and simulations of compact MIMO-systems based on polariza-
[3] A. S. Konanur, K. Gosalia, S. H. Krishnamurthy, B. Hughes, and G. tion diversity,” in Proc. IEEE Topical Conf. Wireless Commun. Technol.,
Lazzi, “Increasing wireless channel capacity through MIMO systems
2003, pp. 284–285.
employing co-located antennas,” IEEE Trans. Microw. Theory Techn., [9] R. M. Narayanan, K. Atanassov, V. Stoiljkovic, and G. R. Kadambi,
vol. 53, no. 6, pp. 1837–1844, Jun. 2005.
“Polarization diversity measurements and analysis for antenna config-
[4] K. Cho, T. Hori, and K. Kagoshima, “Effectiveness of four-branch urations at 1800 MHz,” IEEE Trans. Antennas Propagat., vol. 52, no. 7,
height and polarization diversity configuration for street microcell,” pp. 1795–1810, Jul. 2004.
IEEE Trans. Antennas Propagat., vol. 6, no. 6, pp. 776–781, Jun. 1998.

Authorized licensed use limited to: University of Electronic Science and Tech of China. Downloaded on September 03,2020 at 07:14:43 UTC from IEEE Xplore. Restrictions apply.