LARGE SCALE MILLIMETER WAVE CHANNEL MODELING FOR 5G

LARGE SCALE MILLIMETER WAVE CHANNEL MODELING FOR 5G 1 ARCADE NSHIMIYIMANA, 2 DEEPAK AGRAWAL, 3 WASIM ARIF 1, 2,3 Electronics and Communication Engineering, Department of NIT Silchar. National Institute of Technology Silchar. Assam, India. E-mail: 1 narce655@gmail.com Abstract- The Fifth Generation (5G) is a promising technology to improve the data rate and bandwidth requirements for wireless and mobile communication. It is supposed to be a complementary network which consists of a scalable Radio Access Technologies (RATs). The use of higher frequencies in the range of 30GHz to 300GHz is one of attractive research towards 5G implementation. The corresponding wavelength will be in the range of millimeters (one millimeter to ten millimeters), termed as millimeter wave (mmw) band. In this work we study the channel model for mmw and tackle especially the path loss which is one of the most critical parameter while designing any communication link. The band of mmw will suffer from obstruction to penetrate in many materials and shadowing will be another major problem due to short wavelength. In this work, we present the path loss and large scale model of the channel; we consider the impact of higher frequencies and conduct our analysis at few hundred meters of distance. Attenuation is higher for some specific frequencies as an addition loss of power. Key words- 5G, mmw, channel model, FSPL. I. INTRODUCTION The Fifth Generation (5G) is a promising improvement for wireless and mobile communication in many aspects. It is supposed to be a complementary network which consists of a scalable Radio Access Technologies (RATs). To achieve the awaited high data rate [1], the use of higher frequencies will be adopted. To this end, the corresponding wavelength will be in the range of millimeters (one millimeter to ten millimeters) [1], this technology is termed as millimeter wave (mmw) and it will use mmw band ranging from 30GHz to 300GHz [2],[3],[4]. Some critical frequencies are evaluated in terms of attenuation due to various factors such as rain or water vapor, oxygen and many more [5], fig. 1 shows a scenario of losses and attenuation experienced in water vapor as well as oxygen [4]. As the wavelength will be short [4], the distance between transmitter and receiver will reduce considerably as compared to existing communication and interference should be carefully dealt with [4],[6],[7]. Highly directional steerable antennas will be used [4],[7]. Since the bandwidth will increase, the channel capacity will be improved and as a result the data rate will be high [4],[7]. Additionally, the loss due to the medium will have a big impact and need much attention, the free space contribute much in the loss of the signal which always require a good study in order to design the link budget ensuring enough power at the receiver end. This work discusses the path loss due to the distance constraint as well as frequencies; Log Normal Shadowing (LNS) is used to model Large Scale Channel Fading (LSCF) for mmw band by only considering frequency and distance as main basic contribution parameters [1],[2],[3],[6],[8],[9],[10]. Some literatures have suggested and measured different values which are related to this channel model by considering the Path Loss Exponent (PLE) [1],[2],[3],[8]. Fig.1. Atmospheric attenuation for mmw band. In this study, we simulate the Free Space Path Loss (FSPL) by using the well-known Friis equation [7],[8],[11] by considering mmw band and take records for some specific frequencies which are much cited in the literature due to the countries [3],[8] using them or various other reasons such as attenuation [1],[5],[7] we analyze the contribution of different parameters like distance, frequency as well as PLE. We further simulate the log normal shadowing channel model and use the same approach as before but with the consideration of different data taken from the literature for PLE and deviation factor. We analyze the results by comparing them and provide some suggestion and recommendations on the contribution to the overall path loss due the channel. 78

The rest of this paper is organized as follows: section 2 discusses FSPL as a function of frequency as well as distance with the standard PLE which is 2, section 3 presents the channel model based on LNS and analyzes the impact of distance as well as PLE to the overall contribution of FSPL, section 4 states a concluding remark and presents some suggestion for future directions of mmw. II. FSPL FOR MMW BAND In microwave band which is currently used for many services for communication, the FSPL is proportional to the carrier frequency as well the distance between transmitter and receiver. It is expressed as follows [7]: 4 fd FSPL 10n log 10 ( ) c (1) Where n is the PLE, f is the carrier frequency, d is the distance between transmitter and receiver and c is the speed of the light. For microwave system n for free space is averaged to 2, π is 3.14 and c is 3*10^8m/s. By replacing known quantities in eqn.1 the FSPL expression will be rewritten as: FSPL 32.43 20 log ( f ) 20 log ( d) 10 10 (2) The frequency f is in GHz and the distance d is in m. The following table shows the results obtained by calculating the FSPL using eqn.2 at reference distance taken as 1m [1],[2]. Furthermore, some additional attenuation are remarkable for specific frequencies as shown in table 1 and corresponding remarks are highlighted for clear information. Fig. 2 and fig. 3 show the simulation results for the overall FSPL buy using the well-known formula also given in eqn.2 at reference distance as well as considering distance respectively. It is observed that the frequency contributes to the FSPL at a higher percentage than the distance between transmitter and receiver. Table 1: Some critical frequencies and Path Loss (PL) due to frequencies and corresponding attenuation Path Loss (PL) at reference distance 85 Loss due to frequency PL in 75 70 65 79 60 0 50 150 200 250 300 Frequency in GHz Fig.2. Path loss due to frequency at reference distance d=1

Fig.3. Path loss due to frequency at different distances with n=2 In fig. 2, the FSPL is calculated at reference d=1m and n=2, the frequencies ranging from 28GHz to 300GHz are considered and the results corresponding on the plot are also reported in table 1. It can be seen that the frequency has a big impact for the loss of a communication system. In fig. 3, it can be observed that the change of loss in for both the distance and frequency; is less that 20 at m for example. In fig. 4, there is big change in loss, 40 for examples at 60GHz for a range of 10m to 500m. Space Path Loss (FSPL) in mmw FSPL in 140 130 120 110 90 0 200 300 400 500 Distance in m FSPL @28GHZ FSPL @30GHZ FSPL @34GHZ FSPL @38GHZ FSPL @54GHZ FSPL @57GHZ FSPL @60GHZ FSPL @64GHZ FSPL @73GHZ FSPL @90GHZ FSPL @99GHZ FSPL @164GHZ FSPL @200GHZ FSPL @300GHZ Fig.4. Path loss due to distance at different frequencies with n=2 the contribution of the above mentioned parameter and uses LSCF using LNS. Additionally the path loss exponent is a crucial parameter which can cause an impact to the overall path loss for different distances in different environments. The following discussion will focus on the evaluation of FSPL for fading channel at large scale. The large scale fading or LSCF is modeled by logarithmic function basing on the obstructing objects or shadowing, thus LNS and its expression for FSP is as follows [1],[2],[4],[5],[6], [7], [11]: FSPL FSPL( f, d, env.) (3) Eqn.3 can be detailed as: c FSPL FSPL ( f, d ) 10 n log ( d) X c 0 env., fc 10 env., fc The reference distance d 0 is 1m [1],[6] and the final expression is given as: FSPL 32.43 20log ( f ) 10n log d X 10 10 (4) Where f is the carrier frequency in GHz, n is the best fit minimum mean square error PLE, d is the distance between transmitter and receiver and X σ is a zero mean Gaussian random variable with standard variation σ. The first two quantities give similar results as in table 1 and the next two quantities are to be evaluated. We assume that X σ will have almost same values and its study is left behind. Therefore for different values on n and d, eqn.4 will be written as a function of distance as follows: FSPL( d, n) 10n log ( d) 10 (5) For a range of n between 1.77 to 4.7 [1] without specific consideration of environment (indoor or outdoor) and the distance ranging from 10m to 1km, we have simulated the FSPL due to both the parameters (d, n) and the results are shown in fig. 5 and fig. 6. Loss Loss due to distance and PLE in 150 50 Loss @10m Loss @20m Loss @50m Loss @m Loss @500m Loss @0m III. FSPL BASED CHANNEL MODEL FOR MMW Generally the FSPL is caused by the carrier frequency of the signal as well as the distance between receiver and transmitter. Modeling mmw channel considers 0 1.5 2 2.5 3 3.5 4 4.5 5 Path Loss Exponent (PLE) Fig.5. Path loss due to PLE at different distances In fig. 5, the loss due to PLE is greater than 90 at d=0m for the simulated values of n. for exapmle

n=3 has a loss between 30 to 90 for the full range of 10m t0 0m which makes it 60 which is really a big value. In fig. 6, for diffeerent values of n, in a range of 10m to 1Km, the PL is shown and also the loss is very much and increases considerably from one value of n toanother. Loss Loss due to distance and PLE in 150 50 Loss @n=1.77 Loss @n=2 Loss @n=2.3 Loss @n=3 Loss @n=3.3 Loss @n=4.5 Loss @n=4.7 Pr 40m 47i 47 i (9 68 35.4) Pr 21.6m, which equivalent to: 21.6 10 2.16 Pr 10 10 144.54mW For the non-line of sight (NLOS) scenario n=3.85 [4], the total power at m is Pr 40m 47m 47 m (9 68 77) Pr 20m, which is equivalent to: 20 10 2 Pr 10 10 0.01mW It can be seen that for a change in PLE, the power will considerably change and the loss will be high resulting in a bad signal reception. To overcome this problem different literature have proposed that mmw band will consist of many small antennas to create an environment where transmitter and receiver will be in LOS since the analysis shown that the high change is present when the environment changes from LOS to NLOS resulting in the change of PLE and overall power received. 0 0 200 300 400 500 600 700 0 900 0 Distance in m Fig.6. Path loss due to distance for different values of PLE The purpose of modeling the system is to set necessary conditions in order to design a good communication system. For mmw as a promising technology for 5G implementation especially for its high bandwidth, the received power depends on the power transmitted and gains for both transmitter and receiver antenna but limited by the loss due to medium or channel as well as other sources of signal attenuation. The SNR is a factor to evaluate the quality of overall signal at the receiver, SNR and channel capacity has been studied in [8]. The total received power is calculated in terms of the transmitted power, the antenna gains, the distance between transmitter and receiver with the PLE as well as the free space path loss. The overall expression of the received power is given as [8]: Pr m Pt m Gr i Gt i Pl (6) Where Pr and Pt are received and transmitted power respectively, Gr and Gt being receiver and transmitter antenna gain respectively and Pl is the FSPL. The expression of the FSPL is given in eqn.4, assuming X σ as a average value which is 9 as reported in [1] where it ranges from 7 to 12. We assume a PLE n=1.77 [8] in the line of sight (LOS), we take a fixed frequency of 60GHz from a range of frequency in the unlicensed band 59GHz to 66GHz [8] with a bandwidth of 7GHz, we calculated the FSPL for a distance of m. For a maximum allowed power of 40m [8] and a maximum antenna gain of 47i, the total received power is given by: CONCLUSION In this paper, we presented the channel model based on LSCF considering the effect of FSPL where the important parameters such as distance, carrier frequency as well as PLE were evaluated and some discussion were provided for their role in the path loss. We have simulated the FSPL for different scenario and derived the total power from different literature. We have observed that the environment also is another factor which can cause the received power to change considerably when the same distance and frequency are used for communication. We suggested that for a better signal propagation many small steerable directional antennas can be used to increase the LOS effect in a communication system in order to maintain a good signal reception. The reason is that in mmw, the wavelength is short which results in reduction of antenna height as well as transmission distances, we cannot forget that short wavelength are obstructed by many objects and easily which is also a challenge to the link designers. All the proposals and challenges are the key parameters which should be utilized in the implementation of the next generation of wireless and mobile technology 5G. the mmw is an opportunistic feature for 5G to improve the high data rate as the band width will be high and the capacity will be increased manifold for example in the unlicensed band (59GHz to 66GHz) where wireless communication will gain much benefit. REFERENCES [1] T. S. Rappaport, G. R. Maccartney, S. Member, M. K. Samimi, S. Member, S. Sun, and S. Member, 81

Wideband Millimeter-Wave Propagation Measurements and Channel Models for Future Wireless Communication System Design, vol. 63, no. 9, pp. 3029 3056, 2015. [2] M. R. Akdeniz, S. Member, Y. Liu, M. K. Samimi, S. Member, S. Sun, S. Member, S. Rangan, and S. Member, Millimeter Wave Channel Modeling and Cellular Capacity Evaluation, vol. 32, no. 6, pp. 1164 1179, 2014. [3] A. Al-hourani, S. Chandrasekharan, and S. Kandeepan, Path Loss Study for Millimeter Wave Device-to- Device Communications In Urban Environment Path Loss Study for Millimeter Wave Device-to-Device Communications In Urban Environment, no. JUNE, 2014. [4] A. I. Sulyman, A. T. Nassar, M. K. Samimi, and G. R. M. Jr, Radio Propagation Path Loss Models for 5G Cellular Networks in the 28 GHz and 38 GHz Millimeter-Wave Bands, vol. 52, no. 9, pp. 78 86, 2014. [5] S. Rangan, S. Member, T. S. Rappaport, and E. Erkip, Millimeter Wave Cellular Wireless Networks : Potentials and Challenges, pp. 1 17. [6] S. Sun, G. R. Maccartney, and T. S. Rappaport, Millimeter-Wave Distance-Dependent Large-Scale Propagation Measurements and Path Loss Models for Outdoor and Indoor 5G Systems, no. EuCAP, 2016. [7] Y. Azar, G. N. Wong, K. Wang, R. Mayzus, J. K. Schulz, H. Zhao, F. Gutierrez, D. Hwang, and T. S. Rappaport, 28 GHz Propagation Measurements for Outdoor Cellular Communications Using Steerable Beam Antennas in New York City, no. Icc, 2013. [8] J. Wang, Capacity of 60 GHz Wireless Communication Systems over Fading Channels, vol. 7, no. 1, pp. 203 209, 2012. [9] V. S. Abhayawardhana, I. J. Wassell, D. Crosby, M. P. Sellars, and M. G. Brown, Comparison of Empirical Propagation Path Loss Models for Fixed Wireless Access Systems. [10] C. Pérez-vega and J. L. G. G, Frequency behavior of a power-law path loss model, no. 1, pp. 1 4. [11] O. N. Anthony and O. Obikwelu, Web Site : www.ijettcs.org Email : editor@ijettcs.org Characterization of Signal Attenuation using Pathloss Exponent in South-South Nigeria, vol. 3, no. 3, pp. 104, 2014. [12] S. Janakiraman and P. Marichamy, Propagation Characteristics of Millimeter-Wave Band for 5G Mobile Communications, vol. 8, no. October, pp. 3 7, 2015. 82