Performance Evaluation of Cooperative NOMA IRS Network Using Particle Swarm Optimization Performance evaluation of cooperative NOMA IRS network using particle swarm optimization 1st Anh Le Thi Nationa[.]
2021 8th NAFOSTED Conference on Information and Computer Science (NICS) Performance evaluation of cooperative NOMA-IRS network using particle swarm optimization 1st Anh Le Thi 2nd Hong Nguyen Thi National Laboratory for Securing Information Hanoi, Vietnam leanh41@gmail.com Telecommunication Technical College Hanoi, Vietnam hongntcnc@gmail.com 3rd Trung Pham Viet 4th Vo Nguyen Quoc Bao Vietnam Information Security Association Hanoi, Vietnam trungpv.vie@gmail.com Posts and Telecommunications Institute of Technology HoChiMinh, Vietnam baovnq@ptithcm.edu.vn Abstract—In this paper, we investigate the performance of intelligent reflecting surface (IRS) aided cooperative nonorthogonal multiple access (NOMA) system in an energyharvesting (EH) relay network Particularly, in our proposed system, a base station communicates with two NOMA users with a help of the best EH relay employing a power-splitting (PS) architecture and IRS component To boost the system performance, an important object is the maximization of sum data rate (SDR) which is proposed by using particle swarm optimization (PSO) based on power allocation (PA) for each destination user and phase shift of IRS To validate the performance of PSO, another swarm intelligence technique as Genetic Algorithm (GA) and Exhaustive Search (ES) methods are considered Finally, the outstanding performance of NOMAsupported IRS in comparison with the system without IRS is also shown in this paper Index Terms—Multiple non-orthogonal multiple access, intelligent reflecting surface, particle swarm optimization, sum data rate I I NTRODUCTION The exponentially growing demand for mobile connectivity directed at the new era of the internet of things (IoT) has pushed wireless networks to make revolutionary strides in both technology and architecture Wireless technologies that assist communication networks to support high efficiency, data rate, reduced latencies, huge numbers of connections, and intelligence are interesting topics in the research community Among these technologies, intelligence reflecting surfaces (IRS) has been considered as a promising solution to achieve a smart and programmable wireless transmission environment for the future wireless network, such as beyond 5G and 6G [1] IRS consists of a large number of low-cost re-configurable passive elements that are capable of reflecting electromagnetic waves leading to the amplitude and phase of the signals are influenced [2], [3] Since then, IRS has been studied in [4], to address some new challenges to be efficiently integrated into wireless communication systems such as reflection optimization, channel estimation, and deployment from communication design perspectives Also has the author, [5] investigated a symbiotic unmanned aerial vehicle (UAV) –assisted IRS radio 978-1-6654-1001-4/21/$31.00 ©2021 IEEE system, where the UAV is leveraged to help the IRS reflect its signals to the base station, and meanwhile enhance the UAV transmission by passive beamforming at the IRS The nonorthogonal multiple access (NOMA) technique, which is another multiple access method, has been investigated as a promising solution for the incoming wireless networks [6] Since the principle of NOMA allows multiple users to be superimposed on the same time-frequency resource by modulating signals based on different power levels (on power domain) and multiplexing based on different codes (on code domain) The protrusive advantages of the NOMA approach compared to the previous orthogonal multiple-access (OMA) techniques with high spectral efficiency, massive connectivity, and enhanced user fairness in [7], [8] In particular, in PDNOMA (power domain NOMA), users under low-powered channel conditions will be allocated more transmission power, and successive interference cancellation (SIC) at the receiver is applied to detect the desired signal Owing to its promising features, NOMA has been highlighted in many study papers on the effectiveness of NOMA, which has been combined this technique with many earlier well-known others, produced by industrial and academic bodies For instance, [9] investigated the system performance of a downlink cooperative NOMA scheme with an AF relay and based on the comparison of NOMA with OMA, it was concluded that NOMA offers better spectral efficiency and user fairness Additionally, the applications of multiple-input multiple-output (MIMO) in NOMA systems have been studied in [10], [11] Particularly, authors in [11] demonstrated the rate region in which two popular multiple antennas aided NOMA architectures or when multiple users are grouped in a cluster, exploring both the sum channel capacity and ergodic sum capacity and the results were compared with the results obtained using MIMO/OMA In the context of NOMA, it was considered that IRS can tune the propagation environment to guarantee specific users; order, and hence, enabling the efficient deployment of NOMA wireless systems Furthermore, by efficiently applying 296 2021 8th NAFOSTED Conference on Information and Computer Science (NICS) IRS, it becomes more feasible to change users’ orders to satisfy a particular system requirement according to certain user’s priorities, rather than relying on the random channel of wireless systems As a result that, motivated by the envisioned potentials of IRS, the integration of IRS into NOMA systems has been appealing from the research community [12]–[14] In a network system, which has combined those complicated techniques together, the problem of optimization of system parameters, particularly of power allocation (PA) to each user, is one of the most important problems that must be addressed to improve the system performance [15] Because of the very simple calculation, PSO is considered one of the SI (Swarm Intelligence) algorithms that can be used to find approximate solutions to extremely difficult and complex problems This provides advanced capabilities in terms of robustness, search ability, and improves performance.e [16], [17] To the best of our knowledge, there is limited literate on the consideration of system performance of NOMA-IRS using the PSO technique to maximize the system capacity Thus, in our proposed system, we first investigate the NOMA-IRS system with a multiple-EH relaying network where PSO is applied to address the maximization data rate problem based on power allocation with constraint conditions II S YSTEM MODEL AND SIGNAL TRANSMISSION A System Model Figure presents the proposed system model of an IRSsupported cooperative-NOMA system in a multiple-EH relaying network where IRS is composed of M reflecting elements In this model, EH relays use PSR architecture and employ AF protocol to forward the received signals to users via two paths: direct links to users and reflected links to users with IRS To improve the performance, only one selected relay participates in transmission between the base station and destination users The base station broadcasts its superimposed signal comprised of multiple destination user signals, and each destination user detects its signal by using SIC Moreover, because of NOMA principles, the user that suffers from a weak channel condition will be allocated a higher power whereas a user with a good channel condition will be assigned a lower power In a practical wireless communication network, we can assume that the longer the transmission distance, the poorer the condition of the channel Thus, in this paper, we set up positions of destination users in the proposed NOMA system including the further location as User (U1 ), the closer location as User (U2 ) In addition, in the proposed system, BS communicates with two destination users In addition, in the proposed system, N transmitters communicate with N destination users via a selected relay that is chosen from a relaying cluster including K nodes (Rk , k = K), and these relays employ the AF protocol to forward the received signals to all destination users Two time slots are set up for information processing over the entire communication: (i) in the first time slot, BS transmits its superimposed signals to the selected AF relay, (ii) during the second time slot, the selected relay sends the amplified signal to two NOMA users Without loss of generality, we denote hk , k = 1, 2, , K, gkm , m = 1, 2, , M , gmj , j = 1, 2, and hkj are the channel coefficients of BS − Rk , Rk -IRS, IRS − Uj , and Rk-Uj links respectively The distances of Rk -Uj and IRS − Uj links are dj and dmj respectively In addition, let denote as the reflection amplitude coefficient and as phase shift are designed for IRS-aided our proposed system, where ≤ ηm ≤ and θm ∈ (−π, π] Throughout this paper, because of assumption for positions of destination users for proposed NOMA system, the effective channel gains from the selected relayselected M relay-users are ordered MIRS – users, and the P P iθm iθm gkm ηm e gm2 gkm ηm e gm1 ≤ as and m=1 2 m=1 ∥hm1 ∥ ≤ ∥hm2 ∥ , respectively Moreover, we assume that (1) the channel links of BS − Rk , Rk − Uj are identically and independently distributed Rayleigh fading; (2) the independent links of Rk − IRS and IRS − Uj undergo Rice fading channels; (3) all relays are located in a cluster meaning that the distances between relays are negligible compared to the distance from Relay to other nodes; (4) there are no direct links between BS and Uj ; (5) the local channel state information (CSI) is assumed at the relay, and the global CSI is assumed at BS and the users The selected relay Rk is selected as k th best relay selection by the following strategy: {R∗ } = k th arg max ∥hk ∥ (1) 1≤k≤K B Signal Transmission In the BS-EH relay hop, BS broadcasts the superimposed signal of users as p p (2) x = P x1 + P x2 where xj (j = 1, 2) denotes the signal to user j, and Pj denotes the power allocation coefficient such that P1 + P2 = PS , P1 ≥ P2 , PS is transmitted power from BS The superimposed information x in Eq (2) is transmitted to the selected relay Rk and the received signal at Rk is given by p p [a] yRk = P1 x1 hk + P2 x2 hk + nRk (3) where hk is the channel coefficient of the link BS – Rk , [a] is antenna white Gaussian noise and nRk ∼ CN 0, σaR (AWGN) With PSR architecture, at the selected relay Rk , the received signal, yRk , in (3) is split into two parts, including for energy harvesting and information processing, which can be respectively expressed as p p p [a] βyRk = β1 xhk + βnRk (4) 297 p p p [a] − βyRk = (1 − β)xhk + − βnRk (5) 2021 8th NAFOSTED Conference on Information and Computer Science (NICS) Fig The system model A RF-to-baseband conversion unit samples the received RF signal Hence, the input signal of the information receiver at the selected Rk in the PSR protocol is given by p p [a] [c] c yR = (1 − β)xhk + − βnRk + nRk (6) k | {z } and SINR21 = µβP1 χ2 |h0k | σU 2 µβP2 χ2 |h0k | σU nRk In the EH relay – users hop, the selected relay will forward the received signal to IRS and destination users via reflection links and direct links, respectively Due to the AF protocol is employed at Rk , the remaining signal is transmitted to multiple NOMA users and IRS before amplifying signal by factor G = −1 2 (1 − β) PS ∥hk ∥ + σR k Moreover, the selected relay Rk uses all harvested energy for forwarding the received signal to users, and the transmit power of Rk can be described as PR = µβ (P1 + P2 ) ∥h0k ∥ (7) The selected relay transmits the superposed signals to destination users by two paths: reflection links of an IRS and direct links Thus, the received signal at Uj , j = 1, 2, composed of signal that reflected by IRS at the j-th user and directed on Rk − Uj link is given by ! M X p iθm c yUj = gkm ηm e gmj + hkj GPR yR + nUj (8) k m=1 SINR22 = SINR11 = µβP2 χ1 |h0k | σU + SINR12 = µβσR k (1−β)σU k (1−β)σU χ1 + 1 µβP2 χ1 |h0k | σU µβσR (9) χ1 + (10) µβσR (11) χ2 + k (1−β)σU 2 µβP2 χ2 |h0k | σU 2 µβσR k (1−β)σU (12) χ2 + M P jθm here, χ1 = gkm ηm e hm1 + h1 and M m=1 P jθm χ2 = gkm ηm e hm2 + h2 m=1 Because of NOMA principle, U2 decodes x1 before decoding its signal x2 , the achievable rate for x1 must guarantees accurate signal-decoding for x1 at both users Thus, achievable rates at U1 , U2 can be expressed as follows R1 = log2 (1 + (SIN R11 , SIN R21 )) (13) R2 = log2 (1 + SIN R22 ) (14) and Then, the sum data rate for proposed NOMA system is given as By applying SIC at the NOMA receiver of each user; hence, SINRs at U1 and U2 for information of x1 , x2 can be expressed respectively as µβP1 χ1 |h0k | σU + Rsum = R1 + R2 (15) III P ERFORMANCE A NALYSIS In this section, we analyze the system performance of the proposed system model via the main performance metric including Sum Data Rate (SDR).To evaluate the performance, we consider the problem of how to achieve the highest data rate for the proposed system Especially, in the NOMA system, the power allocation (PA) for each user is an important factor that affects significantly system performance Although, many publications considered the fixed PA for each user 298 2021 8th NAFOSTED Conference on Information and Computer Science (NICS) depending on the position of each user, in fact, the channel condition in wireless medium always changes Because of this, dynamic PA is more significant than fixed PA Thus, in this paper, we propose the maximization of SDR using the swarm intelligence technique based on PA The maximization problem of SDR under multipleconstraint conditions for our IRS-aided transmission can be formulated as follows M aximization {Rsum } = R1 + R2 | {z } (16) Let w, c1 , and c2 denote the inertia weight parameter and cognitive and social parameters, respectively The velocity and position are updated as Vkt+1 = ϖVkt + C1 ∗ r1t LBtk − Xtk + C2 ∗ r2t GBk − Xtk (20) Xt+1 = Xtk + Vkt+1 k Here, r1t and r2t (21) are random numbers between (0,1) at time t P1 ,P2 θm ∈(−π,π] s.t SIN Rjl ≥ γ0 j,l={1,2} θm =[−π,π) P1 + P2 ≤ PSmax P1 ≥ P2 PR = µβ (P1 + P2 ) ∥hk ∥ ≥ P0 (17) Here, P0 is the minimum transmit power that required to forward Relay’s signal to users; γ0 is threshold at which information of signal can be decoded successfully; PSmax is the maximum transmitted power from BS IV G LOBAL BEST PSO ALGORITHM The overall algorithm of the global best PSO applied for our proposed system model [18] is presented in flowchart in Figure Here, each individual particle, k = [1; 2: : : ; Q], where Q is the total number of particles, has a current position Xk = [P1 , P2 ], a current velocity Vk , a local best position LBk , and a global best position GBk that corresponds the position where particle k had the highest value as determined by the objective function f for maximization problem, and a global best position GBk that is the biggest value among all the local best LBk In this algorithm, a penalty function is applied to deal with the above constraint conditions at which we define bi as the i-th constraint (i = 1; 2; 3; 4), αi is the penalty value And φ (bi ) is the satisfaction of the constraint where φ (bi ) = is not satisfied and φ (bi ) = is satisfied Then, based on the objective function, the fitness function can be expressed as X f (Xk ) = Rsum − penalty (18) where P penalty = P Fig The flowchart of the global PSO V N UMERICAL R ESULTS AND D ISCUSSION αi φ (bi ) i=1 The following equations define how the local best position is updated at the next time step t + and the global best position at time step t, where t = [0, , Imax], and Imax is the maximum number of interactions, respectively: t LBtk iff Xt+1 t+1 k < LBkt LBk = (19) t+1 t+1 Xk iff Xk ≥ LBk Then GBtk = max LBtk , f (Xtk ) ≥ GBk , GBk = Xtk In this section, we investigate the secrecy performance of our proposed NOMA-IRS model in the EH relaying network based on the power allocation using the PSO algorithm To illustrate the benefit of the PSO for our proposed model, comparisons with the Exhaustive Search Method and Genetic Algorithm are presented The system parameters are given as the distances of BS-Relay link, Relay-IRS, and IRS-U1 , IRSU2 are 25, 11, 8, 12 (meters) respectively The value of the noise variance at the relays and users is set as -60dBm 299 2021 8th NAFOSTED Conference on Information and Computer Science (NICS) Figure shows the convergence as a function of iteration of the PSO algorithm with three different particle sizes This figure shows a clear trend of the objective function SDR that is improved when increasing the number of iterations The convergence reaches a stable value after approximately 30 iterations; hence the fast convergence of our proposed PSO algorithm for our proposed model system Fig Effects of relay number, IRS element number on Average SDR Fig Convergence behavior of PSO on SDR Figure shows the effects of relay number and IRS element number on SDR First, we can see that the IRS-assisted relaying network offers a higher average data rate significantly compared to the system without IRS Then, the performance is enhanced by increasing the number of Relay nodes and IRS elements Moreover, in this figure, the comparison among three methods as Exhaustive Search (Ex), GA, and PSO is shown And, the average SDR is reached in three cases is approximate However, PSO has an outstanding advantage in time computation when comparing with others This is illustrated in table In this table the iterations for time computation is set up as 500 and the number of IRS elements as M=5 and M=3 Fig Effects of IRS element number on Average SDR TABLE I T IME C OMPUTATION Methods PSO GA Ex M=5 2.781858 (s) 8.4272055 (s) 14324.3372 (s) M=3 0.109041 (s) 0.986498 (s) 1641.04558 (s) Figure presents the benefits of relay selection methods for both case of system with IRS and without IRS It can be seen that NOMA system with IRS case archives the higher performance than the case of no-IRS Moreover, the best relay selection method offers the better performance in comparison with the second relay selection and random relay selection methods Figure and illustrate the effects of distance IRSrelay lind and BS-U1 link on the average SDR with three cases of relay selection strategies 300 Fig Effects of distance of IRS-Relay link on the average SDR 2021 8th NAFOSTED Conference on Information and Computer Science (NICS) Fig Effects of distance of BS-farthest user on the average SDR VI C ONCLUSION In this paper, we first considered and analyzed the average data rate of IRS-assisted cooperative NOMA system using the PSO algorithm based on power allocation In our proposed model, a source communicates to two users with a help of multiple EH relaying nodes and an IRS including M elements Simulation results showed that (i) our proposed NOMA-IRS system outperformed the system without IRS; (ii) the benefits of increasing the number of IRS elements and relay nodes; (iii) the choice of the relay selection method affects the performance of the system; (iii) the position of relay network and IRS influence to the average SDR of the system; (iv) using PSO-based PA will improve the value of average data rate of proposed NOMA-IRS [7] M Liaqat, K A Noordin, T Abdul Latef and K Dimyati, ”Powerdomain non orthogonal multiple access (PD-NOMA) in cooperative networks: An overview”, Wireless Netw., vol 26, no 1, pp 181-203, Jan 2020 [8] D Wan, M Wen, F Ji, H Yu and F Chen, ”Non-orthogonal multiple access for cooperative communications: Challenges opportunities and trends”, IEEE Wireless Commun., vol 25, no 2, pp 109-117, Apr 2018 [9] L Lv, J Chen, Q Ni, Z Ding, H Jiang and S Sharing, ”Cognitive nonorthogonal multiple access with cooperative relaying: A new wireless frontier for 5g spectrum sharing”, IEEE Commun Mag., vol 56, no 4, pp 188-195, Apr 2018 [10] Le, T.A., Kong, H.Y Energy harvesting relay-antenna selection in cooperative MIMO/NOMA network over Rayleigh fading Wireless Netw 26, 2075–2087 (2020) [11] Liu, Yang, Gaofeng Pan, Hongtao Zhang, and Mei Song ”On the capacity comparison between MIMO-NOMA and MIMO-OMA.” IEEE Access (2016): 2123-2129 [12] B Zheng, Q Wu and R Zhang, ”Intelligent reflecting surface-assisted multiple access with user pairing: NOMA or OMA?”, IEEE Commun Lett., vol 24, no 4, pp 753-757, Apr 2020 [13] Z Ding and H V Poor, ”A simple design of IRS-NOMA transmission”, IEEE Commun Lett., vol 24, no 5, pp 1119-1123, May 2020 [14] M Fu, Y Zhou and Y Shi, ”Intelligent reflecting surface for downlink non-orthogonal multiple access networks”, Proc IEEE Globecom Workshops (GC Wkshps), pp 1-6, Mar 2020 [15] T Le Anh and I P Hong, ”Secrecy Performance of a Multi-NOMAMIMO System in the UEH Relaying Network Using the PSO Algorithm,” in IEEE Access, vol 9, pp 2317-2331, 2021 [16] Z.-L Gaing, ”Particle swarm optimization to solving the economic dispatch considering the generator constraints”, IEEE Trans Power Syst., vol 18, no 3, pp 1187-1195, Aug 2003 [17] Y Shi and R C Eberhart, ”Comparison between genetic algorithms and particle swarm optimization”, Proc 7th Int Conf Evol Program., pp 611-616, Mar 1998 [18] Pathak K, Vahinde G Comparison of particle swarm optimization and genetic algorithm for load balancing in cloud computing environment Int J of Research in Computer and Information 2015;1(1) ACKNOWLEDGMENT This work was supported by the 2021 Research Fund of National Laboratory for Securing Information and Telecommunication Technical College, Vietnam R EFERENCES [1] Wu, Qingqing, and Rui Zhang ”Towards smart and reconfigurable environment: Intelligent reflecting surface aided wireless network.” IEEE Communications Magazine 58, no (2019): 106-112 [2] Abeywickrama, Zhang, Wu and Yuen ”Intelligent reflecting surface: Practical phase shift model and beamforming optimization” IEEE Transactions on Communications, 68(9), pp.5849-5863 (2019) [3] Huang, Zappone, Alexandropoulos, Debbah and Yuen Reconfigurable intelligent surfaces for energy efficiency in wireless communication IEEE Transactions on Wireless Communications, 18(8), pp.4157-4170 (2019) [4] Wu, Zhang, Zheng, C.You, and Zhang ”Intelligent reflecting surface aided wireless communications: A tutorial” IEEE Transactions on Communications (2021) [5] Hua, Meng, Luxi Yang, Qingqing Wu, Cunhua Pan, Chunguo Li, and A Lee Swindlehurst ”UAV-assisted intelligent reflecting surface symbiotic radio system.” IEEE Transactions on Wireless Communications (2021) [6] Z Ding, Y Liu, J Choi, Q Sun, M Elkashlan, I Chih-Lin, et al., ”Application of non-orthogonal multiple access in LTE and 5G networks”, IEEE Commun Mag., vol 55, no 2, pp 185-191, Feb 2017 301 ... capabilities in terms of robustness, search ability, and improves performance. e [16], [17] To the best of our knowledge, there is limited literate on the consideration of system performance of NOMA-IRS... III P ERFORMANCE A NALYSIS In this section, we analyze the system performance of the proposed system model via the main performance metric including Sum Data Rate (SDR).To evaluate the performance, ... problem of optimization of system parameters, particularly of power allocation (PA) to each user, is one of the most important problems that must be addressed to improve the system performance