Progress In Electromagnetics Research, PIER 55, 175–188, 2005 THE PERFORMANCE OF POLARIZATION DIVERSITY SCHEMES IN OUTDOOR MICRO CELLS T.-C. Tu, C.-M. Li, and C.-C. Chiu Electrical Engineering Department Tamkang University Tamsui, Taipei, R.O.C. Abstract—The application of polarization diversity reception at the mobile terminal in micro cells at 2 GHz is presentedin this paper. Ray- tracing tool is usedto studyeffects of electric fieldpolarization on the received power in outdoor environments. The performance of diversity schemes with vertical/ horizontal polarization and+45 ◦/−45◦ slanted polarization are comparedin different line-of-sight (LOS) andnonline- of-sight (NLOS) environments. Basedon the evaluation of cross polarization discrimination (XPD) parameters, it is clarifiedthat different environments will affect XPD values in micro cells. Then, the vertical/horizontal polarization diversity and +45◦/ − 45◦ slantedpolarization diversity are chosen to compare with space diversity. Several different combining techniques of polarization and space diversity schemes are also compared in different environments. It is foundthat dual-polarizedantennas for mobile terminal are a promising alternative for two spacedantennas. 1 Introduction 2 Simulation Description 3 Simulation Results and Discussions 3.1 Cross Polarization Discriminations (XPD) 3.2 Diversity Gain 3.3 Diversity Combining 4 Conclusions References 176 Tu, Li, and Chiu 1. INTRODUCTION Multipath propagation resultedin Rayleigh fadingin nonline-of-sight (NLOS) and Rician fading in line-of-sight (LOS) paths in a radio propagation channel [1]. Space diversity is traditionally used to reduce fading problems at the base-station (BS) end in mobile networks. However, two separate receiving antennas are requiredwhen this scheme is appliedandantenna implementation is spatially large. Unfortunately, large antenna spacing increases both size andcost of BS and renders the use of multiple antennas in handsets very difficult. The use of dual-polarized antennas for the mobile terminal is promising cost andspace-effective alternative, where two spatially separated uni-polarizedantennas are replacedby a single antenna structure employing two orthogonal polarizations [2]. Polarization diversity is one of the most promising techniques to reduce fading with a compact antenna configuration requiring only one antenna location for the mobile terminal. The applicability of polarization diversity can partly be evaluated to analyze signal cross correlation andcross polarization discrimination(XPD) values. Further, the effectiveness of a diversity system is measured by a quantity known as diversity gain. The first aim of this paper is to clarify the influence of an environment on polarization diversity scheme in micro cells. This is basedon the evaluation of XPD parameter. The secondtarget is to study the system performance of the polarization diversity scheme and compare it with horizontal space diversity schemes for different environments. 2. SIMULATION DESCRIPTION The characteristics of the micro-cellular channel in outdoor environ- ments are at 2 GHz. Multiple reflections, transmissions, diffractions are taken into account. We simulate two different environments as shown in Figure 1 (urban) andFigure 2 (semi-urban). Figure 1 shows a propagation environment consisting of 12 buildings with different heights between 20 m∼50 m. The main street is 30 m wide. The testing routes are labeledas R1 (LOS) andR2 (NLOS), respectively. Figure 2 shows a propagation environment consisting of 6 buildings with different heights between 20 m∼45 m. The main street is 30 m wide. Above the two routes in different environments, the transmitting antenna is locatedat Tx on the main street near the crossroad. Progress In Electromagnetics Research, PIER 55, 2005 177 150 H=45m 100 H=35m R1 50 H=50m H=30m H=25m 0 Tx Y position(m) H=30m H=20m H=29m H=20m R2 -50 H=40m H=30m H=35m -100 -150 -100 -80 -60 -40 -20 0 20 40 60 80 100 X position(m) Figure 1. Layout of urban. 150 H=45m 100 H=35m 50 R1 0 Tx Y position(m) H=20m H=30m H=20m -50 R2 H=30m -100 -150 -100 -80 -60 -40 -20 0 20 40 60 80 100 X position(m) Figure 2. Layout of semi-urban. 178 Tu, Li, and Chiu Transmitting andreceiving antennas are both half- wavelength dipole. The heights of the transmitting andreceiving antennas are 20 m and 2 m, respectively. Transmitting power is 3.6 W, andthe operating frequency is 2 GHz. In Figures 3 and4, we show the comparisons of propagation loss in the different routes in the urban area. Figures 5 and6, show the comparison of propagation loss in the different routes in the semi-urban area. Each of T x/Rx antennas uses four different polarization types: 1) Tx: vertical polarization/Rx: vertical polarization. 2) Tx: vertical polarization/Rx: horizontal polarization. 3) Tx: +45◦ slantedpolarization/ Rx: +45◦ slantedpolarization. 4) Tx: +45◦ slantedpolarization/ Rx: −45◦ slantedpolarization. 3. SIMULATION RESULTS AND DISCUSSIONS 3.1. Cross Polarization Discriminations (XPD) Two methods are considered in the following. In the first method, the primary polarization is set to be vertical. Therefore, co-polarization for this case means Vertical/Vertical, andcross polarization means Vertical/Horizontal [4]. We assume vertical andhorizontal components of the receiving fieldto have uncorrelatedsmall-scale fadingbecause of different propagation paths. The value of XPD in this method can be determined as: | | | | Pvv 1/Pvv(loss) Pvh(loss) XPDv/h = = | | = | | (1) Pvh 1/Pvh(loss) Pvv(loss) The primary polarization is set to be 45◦ in the secondmethodwhere primary polarization is +45◦. Therefore, co-polarization for this case means +45◦/ +45◦, andcross polarization means +45 ◦/ − 45◦.We assume +45◦ and −45◦ component of the receiving filedto have uncorrelated small-scale fading because of different propagation paths. The value of XPD in this methodcan be determinedas: | ◦ ◦ | | ◦ ◦ | P+45◦+45◦ 1/P+45 +45 (loss) P+45 −45 (loss) XPD+45◦/−45◦ = = = ◦ ◦ | ◦ ◦ | | ◦ ◦ | P+45 −45 1/P+45 −45 (loss) P+45 +45 (loss) (2) We used two methods to compare XPD in different environments (urban & semi-urban areas). In Figures 7 and8, we illustrate the comparison of XPD in the different routes. It is shown that the simulatedXPD value of the vertical/ horizontal polarization diversity scheme was larger than +45◦/ − 45◦ Progress In Electromagnetics Research, PIER 55, 2005 179 150 Tx(V)-Rx(V)---urban Tx(V)-Rx(H)---urban 140 Tx(+45)-Rx(+45)---urban Tx(+45)-Rx(-45)---urban 130 ) 120 B d ( s s o l 11 0 n o i t a 100 g a p o r 90 P 80 70 60 0 10 20 30 40 50 60 70 80 90 100 receiver position(m) Figure 3. Propagation losses in route R1 (urban). 150 140 130 ) 120 B d ( s s o l 11 0 n o i t a 100 g a p o r 90 P 80 Tx(V)-Rx(V)---urban Tx(V)-Rx(H)---urban Tx(+45)-Rx(+45)---urban 70 Tx(+45)-Rx(-45)---urban 60 0 10 20 30 40 50 60 70 80 90 100 receiver position(m) Figure 4. Propagation losses in route R2 (urban). 180 Tu, Li, and Chiu 150 Tx(V)-Rx(V)---semi-urban Tx(V)-Rx(H)---semi-urban 140 Tx(+45)-Rx(+45)---semi-urban Tx(+45)-Rx(-45)---semi-urban 130 ) 120 B d ( s s o l 11 0 n o i t a 100 g a p o r 90 P 80 70 60 0 10 20 30 40 50 60 70 80 90 100 receiver position(m) Figure 5. Propagation losses in route R1 (semi-urban). 150 Tx(V)-Rx(V)---semi-urban Tx(V)-Rx(H)---semi-urban 140 Tx(+45)-Rx(+45)---semi-urban Tx(+45)-Rx(-45)---semi-urban 130 ) 120 B d ( s s o l 11 0 n o i t a 100 g a p o r 90 P 80 70 60 0 10 20 30 40 50 60 70 80 90 100 receiver position(m) Figure 6. Propagation losses in route R2 (semi-urban). Progress In Electromagnetics Research, PIER 55, 2005 181 70 Tx(V)-Rx(V/H)---urban Tx(+45)-Rx(+45/-45)---urban 60 Tx(V)-Rx(V/H)---semi-urban Tx(+45)-Rx(+45/-45)---semi-urban 50 40 ) B d ( 30 D P X 20 10 0 -10 0 10 20 30 40 50 60 70 80 90 100 receiver position(m) Figure 7. XPD in route R1. 70 60 50 40 ) B Tx(V)-Rx(V/H)---urban d ( 30 Tx(+45)-Rx(+45/-45)---urban D P Tx(V)-Rx(V/H)---semi-urban X Tx(+45)-Rx(+45/-45)---semi-urban 20 10 0 -10 0 10 20 30 40 50 60 70 80 90 100 receiver position(m) Figure 8. XPD in route R2. 182 Tu, Li, and Chiu slantedpolarization diversity scheme in each route andenvironment. It was foundthe simulatedXPD value is usually greater in LOS than NLOS paths, because the direct ray is a great influence for the reception, dominating over the other multipath contributions. It was also foundthat XPD value were higher in the semi-urban than urban environment, this may be due to the fact that semi-urban is a more open scenario andthere is fewer obstacles near the antennas andthus the signal receivedwith horizontal ( −45◦) polarization is not sufficiently depolarized to be an important coupling of energy in the vertical (+45◦) polarization. 3.2. Diversity Gain Diversity gain is defined as the ratio of output SNR after combining (γout) to the input SNR on the strongest branch (γin), andis calculated basedon cumulative probabilities. For given cumulative of X, the diversity gain is [5]: γout(X) Gdiv(X)= (3) γin(X) For a cumulative probability X, the SNR after combining is: S2 (X) γ (X)= out (4) out 2σ2 Nout 2 where σNout is the noise power in the combinedsignal and Sout(X)is the envelope in the combinedsignal.