Indonesian J our nal of Electrical Engineering and Computer Science V ol. 24, No. 2, No v ember 2021, pp. 904 909 ISSN: 2502-4752, DOI: 10.11591/ijeecs.v24.i2.pp904-909 904 Exploiting outage perf ormance in de vice-to-de vice f or user gr ouping Dinh-Thuan Do, Chi-Bao Le F aculty of Electronics T echnology , Industrial Uni v ersity of Ho Chi Minh City (IUH), Ho Chi Minh City , V ietnam Article Inf o Article history: Recei v ed Feb 22, 2021 Re vised Sep 7, 2021 Accepted Sep 14, 2021 K eyw ords: De vice-to-de vice Non-orthogonal multiple access Outage probability ABSTRA CT The spectrum ef cienc y and massi v e connections are joint designed in ne w form of de vice-to-de vice for user grouping. A pair of users is implemented with non- orthogonal multiple access (NOMA) systems. Although NOMA benets to such sys- tem in term of the serving users, de vice to de vice (D2D) f aces the interference from normal cellular users (CUE). In particular , we deri v e e xact formulas of outage prob- ability to sho w system performance. In this article, we compare tw o schemes to nd rele v ant schem e to implement in practice. The frame structure is designed with tw o timeslot related to uplink and do wnlink between the base station and D2D users. W e conrm the better scheme in numerical result by considering the impacts of man y pa- rameters on outage performance. This is an open access article under the CC BY -SA license . Corresponding A uthor: Dinh-Thuan Do F aculty of Electronics T echnology Industrial Uni v ersity of Ho Chi Minh City (IUH) Ho Chi Minh City , V ietnam Email: dodinhthuan@iuh.edu.vn 1. INTR ODUCTION T o implement multiple access schemes in cellular netw orks for current wireless systems, orthogonal multiple access (OMA) systems is deplo yed most of syst ems, e.g., frequenc y di vision multiple access (FDMA), time di vision multiple access (TDMA), and orthogonal frequenc y di vision multiple access (OFDMA). In OMA, e xclusi v e resources are allocated to users. Although OMA has adv antages such as lo w comple xity recei v ers and no intracell interference, it suf fers from tw o main disadv antages such as limited number of users and lo w spectral ef cienc y . In the perspecti v e of demand of massi v e connections, non-orthogonal multiple access (NOMA) is promising candidate for multiple access scheme. In the principle of NOMA, it emplo ys dif ferent po wer le v els to multiple x multiple users at the same frequenc y , time and code resources [1]-[7]. The w ork [8] studied do wnlink NOMA by e xamining joint optimization of po wer f actors assigned to users and secure performance w as also e v aluated. The current communication sys tems may also get benets by enabling NOMA for multiple access. The NOMA massi v e MIMO w as presented in [9], [10] to e xplore antennas di v ersity . The w ork in [11] compared NOMA transmissions between multiple antennas case and single antenna case. The near optimal sum-r ate (SR) performance w as e xplored in MIMO-NOMA system and a high-comple xity beamforming w as recommended in such multiple antennas NOMA approach [12]. It is further necessary to study de vice-to-de vice (D2D) communications in the heterogeneous nature of 5G cellular NOMA -aided systems. The base station (BS) normally w ants to e xchange their controlling signal while D2D communications enable proximate cellular users [ 13 ] , [14]. Especially , the spectrum band is reused for pair of D2D users in the cellular systems [15], [16]. Ho we v er , as the main disadv antage, D2D users J ournal homepage: http://ijeecs.iaescor e .com Evaluation Warning : The document was created with Spire.PDF for Python.
Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752 905 meet mutual interference among D2D and normal cellular links. The authors in [17] deplo yed D2D for the applications of NOMA netw orks by allo wing one D2D transmitter send signals to multiple D2D recei v ers with assistance of NOMA. The other promising applications of D2D can be seen in [18]-[24]. As main benet, D2D enables users to communicate ef fecti v ely at close distance, especially processing multiple access with NOMA scheme. Moti v ated by recent studies [18]-[24] and [25], we de v elop tw o practical schemes of D2D-NOMA system along with system performance analysis. 2. SYSTEM MODEL Consider uplink and do wnlink transmission in D2D-NOMA system which consists of groups of paired user under the impacts of BS and con v entional user (CUE), as sho wn in Figure 1 [25]. It is assumed that the BS, and all users are equipped with single antenna. In each group, tw o users need tw o time slots for signal processing. The relay is assumed to decode signal perfectly , which e xhibit tw o schemes. R e l a y U E C U E D 1 U E D 2 B S F i r s t   t i m e   s l o t ,   d e s i r e d   s i g n a l S e c o n d   t i m e   s l o t ,   d e s i r e d   s i g n a l F i r s t   t i m e   s l o t ,   i n t e r f e r e n c e   s i g n a l S e c o n d   t i m e   s l o t ,   i n t e r f e r e n c e     s i g n a l Figure 1. Enabling D2D in NOMA First, P iT is transmit po wer at the source, here we denote ( i = 1 , 2 , c, b, r ) . W e characterize h ij as Raleigh f ading channels to reect g ains of i j link ( j = 1 , 2 , c, b, r ) . It is assumed that h ij = h j i . The additi v e comple x Gaussian noise is assumed fro noise n i ( i = 1 , 2 , c, b, r ) , i.e. n i C N (0 , N 0 ) : P ij = P iT d α ij (1) where α is the path-loss e xponent. In the rst phase, tw o users send their signals ( s 1 and s 2 ) to the relay . The relay needs the second phase to send back its signal s r to the tw o destinations (D1 and D2). W e treat the recei v ed signal as collection of three components of signals as belo w: y D F r = p a 1 P 1 r h 1 r s 1 + p a 2 P 2 r h 2 r s 2 + p P cT h cr s c + n r (2) In the second phase, the relay sends signal s r to destinations. By treating signal s b from the base station. The recei v ed signal at user D1 and D2 are gi v en by [25]: y D F r 1 = P r T h r 1 s r + P bT h b 1 s b + n 1 (3) and y D F r 2 = P r T h r 2 s r + P bT h b 2 s b + n 2 (4) In Scheme I, it is assumed that to decode D1’ s signal the rel ay can eliminate interference from D2 perfectly . F or the rst phase, the signal to interference plus noise ratio (SINR) at relay to detect signal s 1 is gi v en by: γ 1 r = ¯ P 1 r | h 1 r | 2 P cr | h cr | 2 + N 0 (5) Exploiting outa g e performance in de vice-to-de vice for user gr ouping (Dinh-Thuan Do) Evaluation Warning : The document was created with Spire.PDF for Python.
906 ISSN: 2502-4752 where ¯ P 1 r = a 1 P 1 r . Similarly , to help the relay decode D2’ s signal, SINR is gi v en by: γ 2 r = ¯ P 2 r | h 2 r | 2 P cr | h cr | 2 + N 0 (6) where ¯ P 2 r = a 2 P 2 r . In the second phase of Scheme I, the SINRs are computed at tw o destinations D1, D2 respecti v ely: γ r 1 = P r 1 | h r 1 | 2 P cr | h b 1 | 2 + N 0 (7) and γ r 2 = P r 2 | h r 2 | 2 P cr | h b 2 | 2 + N 0 (8) In Scheme II, by treating interference from the CUE, SINR in the rst phase is gi v en by [25]: γ up = ¯ P 1 r | h 1 r | 2 + ¯ P 2 r | h 2 r | 2 P cr | h cr | 2 + N 0 (9) The other computations of SINRs in the second phase at Scheme II are similar as one in Scheme I. 3. AN AL YSIS OF OUT A GE PR OB ABILITY The main system performance metric, namely outage probability , which is dened as probability to SINR less than the required thresholds γ th . Such outage probability corresponding SINR γ is gi v en by: P out = P r ( γ γ th ) (10) 3.1. Scheme I: ideal NOMA Proposition 1: The outage probability at relay R in phase I or at de vice at phase II to detect signal from each de vice is gi v en by [25]: P out,ij = 1 Q ij P k j γ th + Q ij e N 0 Q ij γ th , Q ij = ¯ P ij , i { 1 , 2 } ; j { r } P ij , other w ise (11) Pr oof: W e denote X = P ij | h ij | 2 , Y = P k j | h k j | 2 + N 0 and Z = X / Y . The PDFs of these denotations are represented as f X ( x ) = 1 /P ij e x/P ij and f Y ( x ) = 1 /P k j e y /P k j e N 0 /P k j . W e ha v e PDF f Z ( x ) : f Z ( x ) = Z N 0 y f ( z y , y ) dy = N 0 P k j z + P ij e N 0 P ij z + P ij P k j ( P k j z + P ij ) 2 e N 0 P ij z (12) In particular , f Z ( γ th ) is gi v en by [25]: f Z ( γ th ) = Z γ th 0 N 0 e N 0 P ij z ( P k j z + P ij ) 2 | {z } F Z 1 dz + Z γ th 0 N 0 e N 0 P ij z P k j z + P ij | {z } F Z 2 dz (13) F Z 1 = 1 P ij P k j γ th + P ij e N 0 P ij γ th Z γ th 0 N 0 e N 0 P ij z P k j z + P ij dz + Z γ th 0 N 0 e N 0 P ij z P k j z + P ij dz (14) Then, we ha v e outage probability for uplink: F z ( γ th ) = 1 P ij P k j γ th + P ij e N 0 P ij γ th (15) Indonesian J Elec Eng & Comp Sci, V ol. 24, No. 2, No v ember 2021 : 904 909 Evaluation Warning : The document was created with Spire.PDF for Python.
Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752 907 Similarly , we ha v e outage probability for do wnlink, then we achie v e nal P out . This completes the proof. Therefore, we e xamine outage performance of whole system D2D which is formulated by: P D 2 D ,out,I = 1 (1 P out, 1 r ) (1 P out, 2 r ) (1 P out,r 1 ) (1 P out,r 2 ) (16) 3.2. Scheme II Proposition 2: In Scheme 2, considering uplink from D2D users to the relay , the outage probability at relay in phase I is gi v en by: P out,up = F ( γ up γ thm ) =1 + ¯ P 2 1 r ¯ P 2 r ¯ P 1 r P cr γ thm + ¯ P 1 r e N 0 ¯ P 1 r γ thm P 2 2 r ¯ P 2 r ¯ P 1 r P cr γ thm + ¯ P 2 r e N 0 ¯ P 2 r γ thm (17) Pr oof: W e denote S = ¯ P 1 r | h 1 r | 2 + ¯ P 2 r | h 2 r | 2 , T = P cr | h cr | 2 + N 0 and U = S /T . W e ha v e PDFs ad belo w: f S ( s ) = 1 / ¯ P 2 r ¯ P 1 r e s/ ¯ P 2 r 1 / ¯ P 2 r ¯ P 1 r e s/ ¯ P 1 r (18) and f T ( t ) = 1 /P C T e t/P cr + N 0 /P cr (19) It is noted that f asy U ( u ) is computed by [25]: f asy U ( u ) = Z N 0 tf ( ut, t ) dt = e N 0 P cr P cr ¯ P 2 r ¯ P 1 r ¯ P 2 r u + ¯ P 2 r P cr ne u ¯ P 2 r N 0 P cr + e N 0 P cr P cr ¯ P 2 r ¯ P 1 r   ¯ P 2 r u + ¯ P 2 r P cr ! 2 e u ¯ P 2 r N 0 P cr e N 0 P cr P cr ¯ P 2 r ¯ P 1 r ¯ P 1 r u + ¯ P 1 r P cr N 0 e u ¯ P 1 r N 0 P cr e N 0 P cr P cr ¯ P 2 r ¯ P 1 r   ¯ P 1 r u + ¯ P 1 r P cr ! 2 e u ¯ P 1 r N 0 P cr (20) Then, f asy U ( u ) is re written by: f asy U ( u ) = ¯ P 2 r N 0 P cr ¯ P 2 r ¯ P 1 r   1 u + ¯ P 2 r P cr ! e N 0 ¯ P 2 r u + ¯ P 2 2 r P cr ¯ P 2 r ¯ P 1 r   1 u + ¯ P 2 r P cr ! 2 e N 0 ¯ P 2 r u ¯ P 1 r N 0 P cr ¯ P 2 r ¯ P 1 r   1 u + ¯ P 1 r P cr ! e N 0 ¯ P 1 r u ¯ P 2 1 r P cr ¯ P 2 r ¯ P 1 r   1 u + ¯ P 1 r P cr ! 2 e N 0 ¯ P 1 r u (21) F or uplink, we ha v e P out,up as: P out,up = F ( γ up γ thm ) = 1 + ¯ P 2 1 r ¯ P 2 r ¯ P 1 r P cr γ thm + ¯ P 1 r e N 0 ¯ P 1 r γ thm ¯ P 2 2 r ¯ P 2 r ¯ P 1 r P cr γ thm + ¯ P 2 r e N 0 ¯ P 2 r γ thm (22) This completes the proof. By combining both uplink and do wnlink between the relay and D2D users, we can achie v e the outage probability as belo w for Scheme II: P D 2 D ,out,I I = 1 (1 P out,up ) 2 (1 P out,r 1 ) 2 (1 P out,r 2 ) 2 (23) Exploiting outa g e performance in de vice-to-de vice for user gr ouping (Dinh-Thuan Do) Evaluation Warning : The document was created with Spire.PDF for Python.
908 ISSN: 2502-4752 4. NUMERICAL RESUL TS W e conduct 10 6 iterations for realizing independent channels. The path loss e xponent setting to be α = 4 . W e set the distances d 12 = 1 , d 1 r = 0 . 7 and d 2 r = 0 . 3 , γ th = γ thm = { 3 , 5 , 7 } . The po wer allocation coef cients for NOMA scheme a 1 = 0 . 2 and a 2 = 0 . 8 . Figure 2 and Figure 3 demonstrate the trends of outage probability of Scheme I and Scheme II v ersus transmit SNR respecti v ely . It can be seen clearly that better outage probability occurs at high SNR re gion. The lo wer required threshold γ th = 3 is reported as the better case. as sho wn in Figure 2. Main precise result is recognized when Monte-Carlo and analytical curv es are matched v ery well, which conrm the e xactness of deri v ations. Figure 4 and Figure 5 compare performance of tw o schemes in terms of outage probability and throughput respecti v ely . It is noted that throughput at the x ed rate R is computed by T = R (1 P out ) . 0 5 10 15 20 25 30 35 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 th  = 3, 5, 7 (dB) Figure 2. Scheme I: D2D-NOMA s outage performance when v arying transmit SNR, with d 1 r = 0 . 7 and d 2 r = 0 . 3 -5 0 5 10 15 20 25 10 -3 10 -2 10 -1 10 0 thm  = 3, 5, 7 (dB) Figure 3. Scheme II: D2D-NOMA s outage performance when v arying transmit SNR, with d 1 r = d 2 r = 0 . 5 0 5 10 15 20 25 30 35 10 -1 10 0 th  =  thm   = 10 (dB) th  =  thm  = 5 (dB) th  =  thm  = 0 (dB) Figure 4. Comparison of outage probability between Scheme I and Scheme II, with d 1 r = 0 . 7 and d 2 r = 0 . 3 -5 0 5 10 15 20 25 10 -2 10 -1 10 0 SNR = 10 (dB) SNR = 25 (dB) Figure 5. Comparison of throughput between Scheme I and Scheme II, with d 1 r = 0 . 7 and d 2 r = 0 . 3 5. CONCLUSION In this paper , we ha v e studied a D2D based NOMA transmission scheme in the e xistence of traditi onal cellular user . T o e v aluate the proposed schemes, we computed SINRs and then e xpressions of outage proba- bility are presented. F or the tw o scenarios, we pro vided comprehensi v e analysis of the system performances metrics, and deri v e the closed-form e xpressions of the outage probability . In the follo wing, we concluded that system performance of D2D-NOMA system relying on Scheme I is better than that using Scheme II. More paired users are emplo yed in D2D-NOMA systems in the future w ork. Indonesian J Elec Eng & Comp Sci, V ol. 24, No. 2, No v ember 2021 : 904 909 Evaluation Warning : The document was created with Spire.PDF for Python.
Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752 909 REFERENCES [1] A. Benjebbo vu, A. Li, Y . Saito, Y . Kishiyama, A. Harada and T . Nakamura, “System-le v el performa nce of do wnlink NOMA for future L TE enha ncements, In 2013 IEEE Globecom W orkshops (GC Wkshps) , 2013, pp. 66-70, doi: 10.1109/GLOCOMW .2013.6824963. [2] Z. Ding, M. Peng and H. V . Poor , “Cooperati v e Non-Orthogonal Multiple Access in 5G Systems, IEEE Communi- cations Letter s , v ol. 19, no. 8, pp. 1462-1465, Aug. 2015, doi: 10.1109/LCOMM.2015.2441064. [3] Dinh-Thuan Do and T .-T . Thi Nguyen, “Exact Outage Performance Analysis of Am plify-and F orw ard-A w are Co- operati v e NOMA, T elecommunication, Computing , Electr onics and Contr ol (TELK OMNIKA) , v ol. 16, no. 5, pp. 1966-1973, 2018, doi: 10.12928/telk omnika.v16i5.9823. [4] D. -T . Do, M. -S. V . Nguyen, F . Jameel, R. J ¨ antti and I. S. Ans ari, “Performance Ev aluation of Relay-Aided CR- NOMA for Be yond 5G Communications, IEEE Access , v ol. 8, 2020, doi: 10.1109/A CCESS.2020.3010842. [5] D.-T . Do, T .-L. Nguyen, K. M. Rabie, X. Li and B. M. Lee, “Throughput Analysis of Multipair T w o-W ay Replaying Netw orks W ith NOMA and Imperfect CSI, IEEE Access , v ol. 8, 2020, doi: 10.1109/A CCESS.2020.3008674. [6] D-T . Do et al ., “W ireless po wer transfer enabled NOMA relay systems: tw o SIC modes and performance e v aluation, T elecommunication, Computing , Electr onics and Contr ol (TELK OMNIKA) , v ol. 17, no.6, pp. 2697-2703, 2019, doi: 10.12928/telk omnika.v17i6.12218. [7] Dinh-Thuan Do, Chi-Bao Le and A.-T . Le, “Cooperati v e underlay cogniti v e radio assisted NOMA: secondary net- w ork impro v ement and outage performance, T elecommunication, Computing , Electr onics and Contr ol ( TELK OM- NIKA) , v ol. 17, no. 5, pp. 2147-2154, 2019, doi: 10.12928/telk omnika.v17i5.12451. [8] S. M. R. Islam, M. Zeng, O. A. Dobre, and K.-S. Kw ak, “Resource allocation for do wnlink NOMA sys- tems: K e y techniques and open issues, IEEE W ir eless Commun. , v ol. 25, no. 2, pp. 40–47, Apr . 2018, doi: 10.1109/MWC.2018.1700099. [9] Z. Ding and H. V . Poor , “Design of massi v e-MIMO-NOMA with limited feedback, IEEE Signal Pr ocess. Lett ., v ol. 23, no. 5, pp. 629–633, May 2016, doi: 10.1109/LSP .2016.2543025. [10] B. W ang, F . Gao, S. Jin, H. Lin, and G. Y . Li, “Spatial- and frequenc y wideband ef fects in millimeter -w a v e massi v e MIMO systems, IEEE T r ans. Signal Pr ocess. , v ol. 66, no. 13, Jul. 2018, doi: 10.1109/TSP .2018.2831628. [11] Q. Y ang, H.-M. W ang, D. W . K. Ng, and M. H. Lee, “NOMA in do wnlink SDMA with limited feedback: Performance analysis and optimization, IEEE J . Sel. Ar eas Commun. , v ol. 35, no. 10, 2017, doi: 10.1109/JSA C.2017.2725107. [12] C. Chen, W . Cai, X. Cheng, L. Y ang, and Y . Jin, “Lo w comple xity beamforming and user selection schemes for 5G MIMO-NOMA systems, IEEE J . Sel. Ar eas Commun. , v ol. 35, no. 12, 2017, doi: 10.1109/JSA C.2017.2727229. [13] L. Lei, Z. Zhong, C. Lin, and X. Shen, “Operator controlled de viceto-de vice communications in L TE-adv anced netw orks, IEEE W ir eless Commun. , v ol. 19, no. 3, pp. 96–104, Jun. 2012, doi: 10.1109/MWC.2012.6231164. [14] T . F ang, D. W u, M. W ang and J. Chen, “Multi-stage hierarchical channel allocation in U A V -assisted D2D netw orks: A stack elber g g ame approach, China Communications , v ol. 18, no. 2, 2021, doi: 10.23919/JCC.2021.02.002. [15] A. Asadi, Q. W ang and V . Mancuso, A Surv e y on De vice-to-De vice Communication in Cellular Netw orks, IEEE Communications Surve ys and T utorials , v ol. 16, no. 4, pp. 1801-1819, 2014, doi: 10.1109/COMST .2014.2319555. [16] Y . Liu and E. Erkip, “Capacity and Rate Re gions of a Class of Broadcast Interference Channels, IEEE T r ansactions on Information Theory , v ol. 62, no. 10, pp. 5556-5572, Oct. 2016, doi: 10.1109/TIT .2016.2519029. [17] J. Zhao, Y . Liu, K. K. Chai, Y . Chen and M. Elkashlan, “Joint Subchannel and Po wer All ocation for NOMA Enhanced D2D Communications, IEEE T r ansactions on Communications , v ol. 65, no. 11, pp. 5081-5094, No v . 2017, doi: 10.1109/TCOMM.2017.2741941. [18] Y . Ji et al. , “Spectral Ef cienc y Enhanced Cooperati v e De vice-to-De vice Systems W ith NOMA, in IEEE T ransac- tions on Intelligent T ransportation Systems, v ol. 22, no. 7, Jul. 2021, doi: 10.1109/TITS.2020.3006857. [19] N. Madani and S. Sodag ari, “Performance Analysis of Non-Orthogonal Multiple Access W ith U nderlaid De vice-to- De vice Communications, IEEE Access , v ol. 6, pp. 39820-39826, 2018, doi: 10.1109/A CCESS.2018.2855753. [20] I. Budhiraja, N. K umar and S. T yagi, “Cross-Layer Interference Management Scheme for D2D Mobile Users Using NOMA, IEEE Systems J ournal , v ol. 15, no. 2, Jun. 2021, doi: 10.1109/JSYST .2020.2997731. [21] H. Sun, D. Zhai, Z. Zhang, J. Du and Z. Ding, “Channel Allocation and Po wer Control for De vice-to-De vice Com- munications Underlaying Cellular Netw orks Incorporated W ith Non-Orthogonal Multiple Acce ss, IEEE Access , v ol. 7, pp. 168593-168605, 2019, doi: 10.1109/A CCESS.2019.2954467. [22] J. Liu et al. , “Joint Po wer Allocation and User Scheduling for De vice-to-De vice-Enabled Heterogeneous Netw orks W ith Non-Orthogonal Multiple Access, IEEE Access , v ol. 7, 2019, doi: 10.1109/A CCESS.2019.2916921. [23] B. W ang et al. , “Graph-Based File Dispatching Protocol W ith D2D-Enhanced U A V -NOMA Communications in Lar ge-Scale Netw orks, IEEE Internet of Things J ournal , v ol. 7, no. 9, Sep. 2020, doi: 10.1109/JIO T .2020.2994549. [24] I. Budhiraja, N. K umar and S. T yagi, “ISHU: Interference Reduction Scheme for D2D Mobile Groups Using Uplink NOMA, IEEE T r ansactions on Mobile Computing , pp. 1-16, 2021, doi: 10.1109/TMC.2021.3051670. [25] Y . Ni, S. Jin, Ran T ian, K. W ong, H. Zhu and S. Shao, “Outage analysis for de vice-to-de vice communication as- sisted by tw o-w ay decode-and-forw ard relaying, In 2013 Int ernational Confer ence on W ir eless Communications and Signal Pr ocessing , 2013, pp. 1-6, doi: 10.1109/WCSP .2013.6677242. Exploiting outa g e performance in de vice-to-de vice for user gr ouping (Dinh-Thuan Do) Evaluation Warning : The document was created with Spire.PDF for Python.