This article has been accepted and published on J-STAGE in advance of copyediting. Content is final as presented. IEICE Communications Express, Vol.1, 1–6 Performance Enhancement by Beam Tilting in SD Transmission Utilizing Two-Ray Fading
Tomohiro Seki 1 a) Ken Hiraga2, Kazumitsu Sakamoto2, Maki Arai2 1 College of Industrial Technology, Department of Electrical and Electronic Engineering Nihon University Izumicho1-2-1, Narashino, 275-8575 Japan 2NTT Network Innovation Laboratories, NTT Corporation Hikarinooka 1-1, Yokosuka, 239-0847 Japan a) [email protected]
Abstract: A method is proposed for enhancing the transmission performance in a spatial division transmission system that utilizes the fading characteristics of two-ray ground reflection propagation. The method is tilting the elevation angle of antenna beams purposely towards out of the communicating peer. Using ray-tracing simulation, it is shown that the performance of the system is significantly improved when high-gain antennas with narrow beamwidth are used. Keywords: Two-ray fading, parallel transmission Classification: Transmission systems and transmission equipment for communications
References
[1] K. Hiraga et al., "Spatial Division Transmission without Signal Processing for MIMO Detection Utilizing Two-Ray Fading," IEICE Transactions on Communications, vol. E97.B, no. 11, pp. 2491-2501, 2014. [2] C. Cordeiro, "IEEE doc.:802.11-09/1153r2," p. 4, 2009. [3] Wilocity. Wil6200 Chipset Datasheet. [Online]. http://wilocity.com/resources/Wil6200-Brief.pdf [4] W.L. Stutzman, "Estimating directivity and gain of antennas," IEEE Antennas and Propagation Magazine, vol. 40, no. 4, pp. 7-11, Aug. 1998. [5] D. Parsons, "Propagation over a reflecting surface," in The Mobile Radio Propagation Channel.: Pentech Press, 1992, ch. 2.3, pp. 24-26. [6] K. Uehara, T. Seki, and K. Kagoshima, "Indoor Propagation Calculation Considering Antenna Patterns Using Geometrical Optics Method," IEICE Transactions , vol. J78-B-2, no. 9, pp. 593-601, Sep. 1995.
©IEICE 2017 DOI: 10.1587/comex.2016XBL0213 Received December 11, 2016 Accepted February 27, 2017 Publicized March 28, 2017
IEICE Communications Express, Vol.1, 1–6
1 Introduction To meet increasing demand for wireless transmission rates, some of the hot areas identified in the last decade include applying multiple-input and multiple-output (MIMO). The use of MIMO, however, inevitably results in the need for high computing cost for signal processing. Our solution for this issue was to propose a new spatial division (SD) transmission method that utilizes the fading characteristics of two-ray ground reflection propagation [1]. This method utilizes two-ray fading, taking advantage of the characteristic that high attenuation locations change along with the antenna heights. Multi-stream transmission performance of the SD method relies on the high attenuation in the unwanted path. To obtain these attenuations, reception power level through the direct (or LOS, line of sight) ray and one through the ground reflection (GR) ray should be close to each other, ideally equal. When antennas have directivity, the difference in antenna gains in the LOS ray and GR ray directions makes reception power levels in both rays different. Hence, multi-stream transmission performance is reduced by that directivity. In general, to obtain sufficient signal-to-noise ratio (SNR), high-frequency transceivers have high-gain antennas with high directivity (e.g., 14 dBi in [2]). In fact, some commercial 60 GHz transceivers for WLANs employ narrow-beam antenna arrays with 16 or even 32 elements [3]. To make antenna gains in LOS and GR rays closer to each other and consequently improve transmission performance, we here propose, to adopt depression of the beam in each antenna element.
2 Description of the proposed method In the system analyses that have been presented so far [1], the antenna beam direction has been fixed to be horizontal. Beam tilting methods themselves have been already used in various systems to enhance channel gain. For example, the cellular base stations use beam tilting to increase antenna gain in the direction of the mobile stations that are mostly located lower than the base station’s antenna. However the concept of the beam tilting method introduced this paper is in totally different from classical ones. This method tilts beams purposely towards out of the direction of the communicating peer. Figure 1 illustrates the proposed method. Transmitting and receiving arrays with M elements (here M is three) are placed on the ground facing each other. The distance between the arrays is D. The i-th transmitting element Tx#i and the i-th receiving element Rx#i have the same height from ground hi and directly face each other, where i = 1, 2, 3. LOS rays and GR rays cancel each other in paths except the one between Tx#i and Rx#i. The elevation angles of beams are θi for Tx#i and Rx#i.
IEICE Communications Express, Vol.1, 1–6
Tx Array Radiation pattern Rx Array Tx#3 Rx#3 θ3 θ3 Tx#2 Rx#2
θ ... 2 θ2 Tx#1 Rx#1 θ1 θ1 h3 h2 h1 Ground D
Figure 1. An illustration of the proposed method
3 Performance simulation procedures
Figure 1 shows the simulation model. Antenna heights, key parameters for the SD method, are optimized for each transmission distance using the equations below. These are derived with the same procedure shown in [1], setting the reflection coefficient on the ground surface to negative real number, because when the polarization is parallel to the ground and the grazing angle is sufficiently low, the ground reflection coefficient is negative whether the ground is conductive or not. When M = 2, the optimum antenna heights are h 5 1 D h 1 D 1 , 2 , (1) 0 2 0 0 5 1 0 where λ0 is the wavelength. When M = 3, the optimum antenna heights are expressed as Eq.(2), where N12, N23, and N31 are unequal odd numbers excluding 1
[1]. In this paper these parameters are set to be the smallest combination, N12 = 3,
N23 = 5, and N31 = 7. h N N D h N N D h N N D 1 12 31 , 2 12 23 , 3 23 31 . (2) 0 2N 23 0 0 2N31 0 0 2N12 0 In this model it is assumed there is antenna directivity in the plane on which all antenna elements are located. The directivity of Tx#i or Rx#i is denoted by
Gi(θ), where θ is the elevation angle. Here θ of depression (below horizon) is set to be negative. The antenna’s radiation pattern without tilt G0(θ) is set to have no side lobes and is defined by using the simplified cosine model shown in eqn. (15) and
(18) in [6]. Using the half-power beamwidth, θ3dB, we express the model’s radiation pattern without tilting G0(θ) as a function of the elevation angle θ.
q log2 2 G0 () cos , where q (3) log2 cos3dB Here, using eqn. (12) in [6], we assumed the beam was rectangular and was equivalent to a beam of half-wave dipole array. The antenna gain g is defined as
g 41253 783dB , (4) where η is the radiation efficiency of the antenna. Each antenna element in the arrays on both sides has equal directivity. Each antenna is assumed to have the same gain, g. The radiation pattern of Tx#i or Rx#i with beam tilting, Gi(θ), is
IEICE Communications Express, Vol.1, 1–6
G () G ( ) i 0 i , (5) where θi is the elevation angle of the beam of Tx#i or Rx#i, as described in the previous section.
From the calculation shown in [7], E0_ij, the complex valued electric field at
Rx#j, which is produced by the LOS ray r0_ij transmitted from element Tx#i, and
Er_ij, which is produced by the GR ray rr_ij transmitted from the same element, are expressed as 2 2 j r j r 30Pt gG j ( 0 _ ij ) 0_ij 30Pt gG j ( r _ ij ) r _ij 0 0 . E0 _ ij e , Er _ ij Rij e (6) r1_ ij rr _ ij
The elevation angle toward the LOS ray is denoted by θ0_ij and that toward the GR ray by θr_ij. Rij is the reflection coefficient for the reflection from the ground surface in the GR ray rr_ij and Pt is the transmission power. The path length of the direct wave is r0_ij and that of the reflected wave is rr_ij; these lengths are calculated from the numbers given in Figure 1. The actual reception power at an antenna Rx#j is calculated, taking the phase rotations due to the path lengths into account, by using the formulated calculations shown in [8]. 2 2 2 2 r r g G j (0_ij )Gi (0_ij ) 0 _ ij G j (r _ij )Gi (r _ij ) r _ ij 0 0 0 . Pij Pt e Rij e (7) 4 r0_ij rr _ij Since signal-to-interference-plus-noise ratio (SINR) at Rx#i is P SINR(i) ii , i 1,2,,M M , (8) Pij PN _ i j1, ji where the noise level at each receiving antenna is PN_i. To calculate the channel capacity, the equation below is used. M
C log 2[1 SINR(i)] (9) i1 4 Effectiveness of proposed method and optimum beam tilt angles In this section, using the ray-tracing whose procedure is described above, we show the effectiveness of the proposed method. It is assumed that the ground is a conductor and that the polarization is parallel to the ground. Each antenna height is set using eqn. (1) and (2). Taking the transmission of modulated signal into account, we can clarify this type of SD transmission system’s performance by evaluating the SIR (signal-to-interference ratio) in each transmission path. SIR(8). Here, SIR(i) denotes an SIR at Rx#i. Figure 2(a) shows the SIR enhancement obtained for M = 2 in the 10° - 90° half-power beamwidth range by using the proposed beam tilting at the transmission distance of 2000 wavelengths. Figure 2(b) shows the same for M = 3. It can be seen that beam tilting significantly improves the SIR values when half-power beamwidth is less than 60°. This indicates that the proposed method improves the performance. On the other hand, it cannot be expected to enhance the performance of SD transmission systems with large beamwidth because relatively little degradation occurs in them even without beam tilting.
IEICE Communications Express, Vol.1, 1–6
70 40 60 30 50 40 SIR(1) w/o Tilt 20 SIR(2) w/o Tilt
30 SIR(1) w/o Tilt SIR(3) w/o Tilt
SIR [dB]SIR SIR [dB]SIR 20 SIR(2) w/o Tilt 10 SIR(1) w/ Tilt SIR(1) w/ Tilt SIR(2) w/ Tilt 10 M = 2 SIR(2) w/ Tilt M = 3 SIR(3) w/ Tilt 0 0 0 30 60 90 0 30 60 90 Half-power beamwidth, θ 3dB Half-power beamwidth, θ3dB
Figure 2. SIR enhancements.(a)M = 2, (b) M = 3. D = 2000 λ0.
We can also see that SIR of M = 3 is lower than that of M = 2 in a whole range. When M = 2, the optimum antenna heights and beam tilt bring complete canceling of LOS ray and GR ray simultaneously in both diagonal paths (the path between Tx#1 and Rx#2 and the path between Tx#2 and Rx#1). On the other hand, when M = 3 there is no combination of θi (i = 1,2,3) and antenna heights that brings complete canceling in all diagonal paths, because, for example, the tilt angles are not fully optimum for both path between Tx#1 and Rx#3 and path between Tx#1 and Rx#2. There is a trade-off for these two paths when we optimize the beam tilts. Figure 3 shows the optimum beam tilt angles as the function of half-power beamwidth on the same conditions as Figure 2. Here θ1 is fixed to zero for M = 2 because either θ1 or θ2 has to be tuned when M = 2, as explained above. When the beam is narrow, the path gains of the LOS ray and GR ray can be equalized with small beam tilt. So the optimum tilt is small for narrow beamwidth.
On the other hand, when the beam is wide (e.g., θ3dB > 60° for M = 2), SIR degradation is very small even without beam tilting, hence there is no need for tilting and the optimum tilt gets smaller along with the increase of the beamwidth. Accordingly the curves shown in this figure have maximum values. When M = 3, the optimum tilt angles increase along with θ3dB in the range of up to 80°, which is larger value than M =2 system. This is because, in M = 3 system, antenna heights are higher than those in M = 2 system with the same transmission distance, so larger tilt angles are still required even with wider beam in order to minimize the difference between the received powers of LOS and GR rays. Accordingly the
maximum value of the curve is at larger beamwidth. ]
-3
degrees [
-2 angle θ2θ 2(M=2)for M = 2 -1 θ3θ 3(M=3)for M = 3 θ2θ 2(M=3)for M = 3 θ1θ 1(M=3)for M = 3
-0 Optium tilt Optium 0 30 60 90
Half-power beamwidth, θ3dB
Figure 3. Optimum tilt angles. D = 2000 λ0. For M = 2, θ1 = 0°.
IEICE Communications Express, Vol.1, 1–6
5 Conclusions This paper proposed a method to enhance the transmission performance in a spatial division transmission system without any MIMO signal processing by exploiting two-ray fading. Using ray-tracing simulation, we clarified the method’s effectiveness.