Enabling full-duplex in multiple access technique for 5G wireless networks over Rician fading channels

Nowadays, unmanned aerial vehicle (UAV) relays’ assisted internet of things (IoT) systems provide facility in order to overcome the large scale fading between source and sink. The full-duplex scheme enables wireless network to provide higher spectrum efficient technology. This paper analyses performance of two users which are served by new emerging non-orthogonal multiple access (NOMA) technique. Exact outage probability of such two users are derived and checked via Monte-Carlo simulation. These analytical results provide guideline to design UAV in real application. This paper provides a comprehensive study to examine impact of interference, fixed power allocation factors to system performance.

INTRODUCTION State of the art, wireless network provides ability to serve massive connections and such requirement satisfied by the application of non-orthogonal multiple access (NOMA) in fifth generation (5G) networks. NOMA has drawn wide attention due to its potential to improve spectral efficiency [1] and more reliable improvement with relaying scheme [2]. Different from conventional orthogonal multiple access (OMA), NOMA benefits from relaying design, far user can be served by the base station under help of the relay. NOMA enables various applications to serve multiple users to be served at the same time and frequency by superimposing multiple users in the power domain at the transmitter and using successive interference cancellation (SIC) at the receiver [3,4]. In NOMA system, device-to-device transmission mode is activated based on the users divided into different kinds of categories according to their channel conditions, i.e., the near user and the far user [5].
In the context of NOMA, cooperative mode is mainly divided into two cases. These cases exhibit implementation of NOMA to imrpove performance of users located edge of cell [6][7][8][9]. In cooperative NOMA, the near users with strong channel conditions needs to act as relays. The relay provides improved performance of the far users who have poor channel conditions [9,10]. However. However, half-duplex (HD) mode is studied in the cooperative NOMA relay in the early works [7][8][9][10][11]. The authors in [6]  and enhancing end to end transmission quality. The authors in [12] investigated advantage achieved by FD mode, which proved enhancing performance gain. While the authors in [13] maximized energy efficiency for full-duplex cooperative NOMA with power allocation. However, open problem still exists related to fading model. This paper fulfills a gap in [13][14][15][16][17][18][19][20][21][22], in which UAV-based relay is not considered.

SYSTEM MODEL
This paper considers a two-user NOMA architecture, where UE-1 directly exchanges data with the base station (BS) as depicted in Figure 1, while UE-2 receives signals from the BS via unmanned aerial vehicle (UAV) relay. The link BS to UE-2 is supported by a dedicated relay. Note that each node is equipped single antenna except for relay which requires two antenna to provide ability of FD. The probabilistic LoS and non-LoS (NLoS) model for UAV is adopted to indicate the large scale fading. We use such model for the channel between the UAV and terrestrial user due to impact of the density of buildings and the distance between the UAV and users. The probability of user which has benefit of a LoS link is expressed as [24] in which we denote p and q are constant values depending on the surrounding environment (sub-urban, urban, dense-urban). Therefore, θ k in (1) can be expressed as in which H denotes the height of UAV, d k = r 2 k + H 2 is the distance between user k and the UAV, and r k is the distance between users and UAV. Obviously, the probability of NLoS is P N LoS,k = 1 − P LoS,k . The received signal at the relay is expressed by

Signal link Interference link
where κ, 0 ≤ κ ≤ 1 denotes as level of self-interference (LI), ϕ k , k = 1, 2 are power allocation factors to two NOMA users who need receive signal s k , P S is transmit power of the BS,s LI is self-interference signal due to FD mode, d S = r 2 S + H 2 is the distance between BS and UAV, and α is the path loss exponent from the BS to UAV. The signal to interference plus noise (SINR) to detect signal s 2 is given by Ì ISSN: 1693-6930 where ρ = P S /N 0 = P R /N 0 is the transmit signal-to-noise radio (SNR). Employing SIC, interference s 2 is deleted to detect s 1 corresponding SINR as Then, the received signal at user UE-1 is given as whered 1 is the distance between BS to UE-1, P L k = (P LoS,k + υP N LoS,k ), k ∈ {1, 2} and υ denotes the additional attenuation factor of NLoS transmission. The SINR to detect s 2 and then s 1 at UE-1 are respectively expressed by The received signal and SNR at user UE-2 is formulated respectively as 3. OUTAGE PROBABILITY ANALYSIS Case 1: 0 < κ < 1, the outage probability without interference link of UE-1 is given by where ∆ = ρP L 1 d −α 1 κ,θ = max ε2 ρ(ϕ2−ϕ1ε2) ,ε 1 ϕ1ρ ,ε 1 = 2 2R1 − 1 with R 1 is denoted as the target rate at UE-1 to detects 1 andε 2 = 2 2R2 − 1 with R 2 being the target rate at UE-2 to detects 2 . Thus, the probability distribution function (PDF) of the unordered squared channel gain X, X ∈ {h 0 , h 1 , h 2 , g 0 , g 1 }, is formulated by a non-central chi-square distribution with two degrees of freedom as [25] where I 0 (.) is the zeroth-order modified Bessel function of the first kind, K X = |µ X | 2 2σ 2 is the Rician factor and Ω X = E |X| 2 = 1 is the normalized average fading power. The corresponding cumulative distribution function (CDF) is known as where Q (ν, ι) where Λ g1 = (1+Kg 1 ) (14) Case 2: κ = 0, the outage probability without interference link of UE-1 is given by In particular, the outage probability with impact of interference at UE-2 is given by . Similarly with solving (13), it can be achieved φ 1 as where Λ g0 = (1+Kg 0 ) Next, φ 2 is calculated as   (18) and (16) into (15), OP U E−2 is given by

NUMERICAL RESULTS
To perform simulations, we set K = K h0 = K h1 = K h2 = K g0 = K g1 = 2 and Ω = Ω h0 = Ω h1 = Ω h2 = Ω g0 = Ω g1 = 1; power allocation factors are ϕ 1 = 0.2 and ϕ 2 = 0.8; target rates are R 1 = 1 and R 2 = 0.5; coefficient related to SI from FD is κ = 0.01. Path loss exponent is α = 2, the height of UAV H = 30m, additional attenuation factor is υ = 20 (dB), environment parameter is p = 4.8860, environment parameter is q = 0.4290. The times of Monte Carlo simulation 10 6 ,d 1 = 0.7. In Figure 2, outage performance of user UE-2 is better than that of UE-1 at numerous case of Rician fading parameters. It is valuable as wellmatching between Monte-Carlo and analytical simulations. At higher SNR, improved outage performance can be seen. As illustration in Figure 3, it is existence of optimal outage performance of user UE-1 as varying a 2 from 0.5 to 1. It can be further seen that lower SI leads to better outage performance at two users. While Figure 4 indicates that improvement outage performance happens at higher value of K related to Rician fading channel.

CONCLUSION
In this paper, we have discussed how impacts of Rician fading and full-duplex mode have been implemented in the relaying network. We have explored several exact expressions of outage probability to show performance of each user. This findings benefit to deployment of full-duplex to provide higher spectrum efficiency for cellular network in practice.