Inter national J our nal of Electrical and Computer Engineering (IJECE) V ol. 15, No. 2, April 2025, pp. 1774 ∼ 1782 ISSN: 2088-8708, DOI: 10.11591/ijece.v15i2.pp1774-1782 ❒ 1774 Buffers balancing of b uffer -aided r elays in 5G non-orthogonal multiple access transmission inter net of things netw orks Mohammad Alkhwatrah, Nidal Qasem Communications and Computer Engineering Department, Al-Ahliyya Amman Uni v ersity , Amman, Jordan Article Inf o Article history: Recei v ed Jun 13, 2024 Re vised Dec 4, 2024 Accepted Dec 14, 2024 K eyw ords: Balancing Buf fers Internet of things Non-orthogonal multiple access Relays ABSTRA CT Buf fer -aided cooperati v e non-orthogonal multiple access (NOMA) enhances the ef cienc y of utilizing the spectral by allo wing more users to share the same re- sources to establish massi v e connecti vity . This is remarkably attracti v e in the fth generation (5G) and be yond systems, where a massi v e number of links is essential lik e in the internet of things (IoT). Ho we v er , the capabilit y of b uf fer co- operation in reducing the outage is limited due to empty and full b uf fers, where empty b uf fers can not transmit and full b uf fers can not recei v e data pack ets. Therefore, in this paper , we propose balancing the b uf fer content of the inter - connected relays, so the b uf fers that are more full send pack ets to the emptier b uf fers, hence all b uf fers are more balanced and f arther from being empty or full. The simulations sho w that the proposed balancing technique has impro v ed the netw ork outage probability . The results sho w that the im pact of the balancing is more ef fecti v e as the number of relays in the netw ork is increased. Further - more, utilizing the balancing with a lo wer number of relays may lead to better performance than that of more relays without balancing. In addition, gi ving the balancing dif ferent le v els of priorities gi v es dif ferent le v els of enhancement. This is an open access article under the CC BY -SA license . Corresponding A uthor: Mohammad Alkha w atrah Department of Communications and Computer Engineering, Al-Ahliyya Amman Uni v ersity Amman, Jordan Email: m.alkha w atrah@ammanu.edu.jo 1. INTR ODUCTION W ireless communicati o n is currently one of the most crucial forms of communication. As a result, the unprecedented gro wth in the number of online de vices is making wireless communication in future applications increasingly comple x. F or instance, the internet of things (IoT), which relies on fth-generation (5G) and be yond technologies, demands e xtensi v e wireless connecti vity while ensuring a lo w probabil ity of outages. Ho we v er , these requirements cannot be fullled by the e xisting infrastructure [1], [2]. In con v entional wireless communication, orthogonal transmission is typically used, where each link between transmitters and recei v ers is assigned a unique frequenc y band, time-slot, or code to pre v ent interfer - ence between dif ferent links. Ho we v er , this method reduces spectral ef cienc y and is i nadequate for the future of communication systems [3]. As a result, non-orthogonal multiple access (NOMA) has been introduced to enable simultaneous use of the same resources. Unlik e orthogonal schemes, NOMA allo ws multiple users to transmit using the same code, time, and frequenc y , b ut with v arying po wer le v els. Specica lly , NOMA dedi- cates less po wer to users with good channel conditions, who are referred to as strong users, as the y can decode J ournal homepage: http://ijece .iaescor e .com Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Elec & Comp Eng ISSN: 2088-8708 ❒ 1775 their mes sages with less po wer . W ith adv ances in signal processing, NOMA becomes feasible with sophisti- cated recei v ers, which enable strong users to eliminate interference from sharing the same resources through successi v e interference cancellation (SIC) [4]. Since higher po wer is allocated to weak er users (with poorer channels), the y can treat the strong user’ s signal as interference and successfully decode their o wn. As a result, NOMA enables more users to share the same resource, such as frequenc y or time slots, with dif ferent po wer le v els, thereby enhancing spectral ef cienc y [5]. Hence, NOMA is an attracti v e solution to achie v e the massi v e connecti vity required for 5G applications such as the IoT [6]-[8]. At present, NOMA is being considered for inclusion in the 3GPP Release 16 standards for 5G systems [9]. Relay techniques are another crucial method for enhancing netw ork performance [10]. Relays help transmit pack ets from the source to users by pro viding an alternati v e path between the source and the destina- tion. As a result, long-term e v olution (L TE) Release 10 has ackno wledged the role of relay node cooperation as a k e y component in enabling modern wireless communications [11]. There are tw o primary types of relay- ing: amplify-and-forw ard (AF) and decode-and-forw ard (DF). In the AF method, the relay node amplies the recei v ed signal and forw ards it to the users, which simplies its implementation. Ho we v er , this approach also amplies an y noise present in the signal. In contrast, the DF method allo ws the relay node to decode the re- cei v ed signal, re-encode it, and then send the decoded signal to the users. While DF resolv es the issue of noise amplication, it demands higher channel g ains to achie v e acceptable quality of service (QoS), making it more resource-intensi v e than AF [12]. The adv antage of relay sel ection in netw orks with multiple relays lies in its competiti v eness with multiple-input, multiple-output (MIMO) systems, while remaining simpler to implement. This is because relay selection does not require comple x ph ysical layer techniques lik e synchronization, which are necessary in MIMO systems [13], [14]. Gi v en the ef fecti v eness of relays in mitig ating communication link losses, the inte gration of cooper - ati v e relaying with adv anced techniques has been e xtensi v ely e xplored in the literature [15]. Non-orthogonal multiple access (NOMA) has been successfully implemented in cooperati v e relay netw orks, with se v eral stud- ies suggesting the use of con v entional (non-b uf fer) relay selection for cooperati v e NOMA [16]. In [17], a tw o-stage relay selection strate gy is proposed to maximize the data rate for users. The analysis and simulation results demonstrate that this approach outperforms non-cooperati v e NOMA. A recent inno v ation in cooperati v e netw orks is the use of b uf fer -aided relays [18], which allo ws for better alignment of tr ansmissions with stronger links compared to traditional non-b uf fer relay selection methods [19]. As a result, b uf fer -aided techniques ha v e become the state-of-the-art in cooperati v e NOMA netw orks. Additionally , [20] introduced an adapti v e link selection strate gy for a single-relay NOMA netw ork, assuming an innite b uf fer size. The analysis re v eals that this system achie v es lo wer outage rates and hi gher throughput compared to con v entional relaying schemes in NOMA. In real-w orld scenarios with limited b uf fer capaciti es, b uf fers often e xperience frequent states of be ing either full or empty . The performance of b uf fer -aided cooperati v e relay netw orks is hea vily dependent on the number of pack ets stored in the b uf fers, as this directly dictates the b uf fer’ s state. When a relay b uf fer is either full or empty , the corresponding source-to-relay or relay-to-user link becomes una v ailable for pack et transmission or reception, respecti v ely [21]. In [22], the outage probability is dened as the lik elihood that either the source-to-relay link cannot support the NOMA data rate or the relay-to-user links are unable to transmit the NOMA data. As noted in [23], nding an optimal protocol that minimizes outage probability while adhering to a specic del ay constraint remains an unresolv ed challenge, e v en in the simplest form of b uf fer - aided relay netw orks. Consequently , de v eloping t he ideal selection scheme for relays with nite, practical b uf fer si zes remains an open research problem. Based on the achie v ed outage probability , the best a v ailable relay selection schemes consider the b uf fer contents (states) in addition to the links states as well. Such selection schemes prioritize relays based on tar get b uf fer length t o minimize the occurrence of full and empty b uf fers. Ho we v er , to the best of the author’ s kno wledge, none of the a v ailable studies has considered transferring data pack ets between relays so the b uf fers that are closer to be full help other b uf fers to a v oid being empty , we call this process the balancing. Accordingly , moti v ated to ll this g ap in the literature and to minimize the outage probability by using b uf fer -aided relays with NOMA and get closer to realizing the IoT netw orks. T o summarize, the k e y no v elty of this article is to apply the balancing to b uf fer -aided cooperati v e NOMA netw ork to reduce the outage probability . The main contrib utions of this article are summarized as follo ws: i ) proposing a no v el relay content balancing to reduce the occurrence of empty and full b uf fers; ii) Studying the impact of the balancing with v arious numbers of relays on the NOMA netw ork outage probability; and iii) studying the outage probability of the netw ork under prioritizing recei ving pack ets o v er the balancing. Buf fer s balancing of b uf fer -aided r elays in ... (Mohammad Alkhwatr ah) Evaluation Warning : The document was created with Spire.PDF for Python.
1776 ❒ ISSN: 2088-8708 The rest of the article is or g anized as follo ws: the system model for the suggested balanced b uf fer - aided cooperati v e NOMA netw orks is in s ection 2. The performance analysis of the balanced b uf fer -aided relay is presented in section 3. Simulation trials of the proposed system along with comparison with other a v ailable solutions are discussed in detail in section 4. Finally , the conclusion is presented in section 5. 2. SYSTEM MODEL The system model of the proposed balanced b uf fer -aided cooperati v e NOMA netw ork is sho wn in Figure 1. Figure 1 illustrates a source node S , k half-duple x DF b uf fer -aided relays, k = 1 , 2 , 3 ...., K , with R 1 is the selected relay with enough pack ets for transmitting via NOMA to the tw o users U 1 and U 2 simultaneously . The system model can be e xtended to an y number of users as in [24]. The interconnection between relays in Figure 1, assures the balancing of data pack ets to a v oid full and empty b uf fers. So, the relay wit h a longer queue can transfer pack ets to other relays as sho wn for R 1 and R 2 . It is w orth noting that R 2 with an empty b uf fer is not connected to the users and the source is not connected to R 1 as it has a full b uf fer . Figure 1. Balanced b uf fer -aided cooperati v e NOMA netw ork The relay R k has a L -size b uf fer to store the pack ets. The source-to-relay S − R k , source-to-users S − U m ( m denotes the user number) and relay-to-users R k − U m links ha v e the channel coef cients h sr k , h su m and h r k u m , respecti v ely . The channels are modeled with at Rayleigh f ading coef cients, which remain constant during each time-slot b ut v ary randomly across dif ferent time-slots. F or simplicity , P t denotes the transmit po wer at all transmitting nodes (whether source or relay), and σ 2 represents the noise v ariance at all recei ving points. The tar get rate for data transmission is assumed to be constant, denoted by ϵ . If the capacity of a link meets or e xceeds ϵ , the link i s considered acti v e and capable of supporting the transmission. If the capacity is lo wer than ϵ , the link is inacti v e and transmission cannot occur , meaning the link is in outage. Due to shado wing ef fects, the direct link between the source S and user U m , denoted S − U m , is assumed to be block ed. All nodes are assumed to ha v e information about the states of all links. 2.1. OMA transmission At time-slot t , channel capacities are calculated as (1): C sr k ( t ) = log 2 (1 + γ sr k ( t )) , C r k u m ( t ) = log 2 (1 + γ r k u m ( t )) , (1) Int J Elec & Comp Eng, V ol. 15, No. 2, April 2025: 1774-1782 Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Elec & Comp Eng ISSN: 2088-8708 ❒ 1777 where γ sr k ( t ) = P t σ 2 | h sr k ( t ) | 2 and γ r k u m ( t ) = P t σ 2 | h r k u m ( t ) | 2 . The channel g ains | h sr k ( t ) | 2 and | h r k u m ( t ) | 2 are assumed to follo w an e xponential distrib ution, with mean v alues θ sr k = E [ | h sr k ( t ) | 2 ] and θ r k u m = E [ | h r k u m ( t ) | 2 ] , where E [ . ] denotes the e xpectation operator . Both γ sr k ( t ) and γ r k u m ( t ) also follo w e xpo- nential distrib utions, with means ¯ γ sr k = P t σ 2 θ sr k and ¯ γ r k u m = P t σ 2 θ r k u m , respecti v ely . Thus, γ sr k ( t ) , γ su m ( t ) , and γ r k u m ( t ) represent the instantaneous signal-to-noise ratios (SNRs), while ¯ γ sr k and ¯ γ r k u m are the a v erage SNRs for the channels h sr k ( t ) and h r k u m ( t ) , respecti v ely . As mentioned earlier , when the link capacity f alls belo w the tar get data rate, an outage occurs. This outage is calculated using the f act that both γ sr k ( t ) and γ r k u m ( t ) follo w e xponential distrib utions. Their cumulati v e distrib ution functions (CDFs) are gi v en by (2): P { log 2 (1 + γ sr k ( t )) < ϵ } = 1 − exp − 2 ϵ − 1 ¯ γ sr k P { log 2 (1 + γ r k u m ( t )) < ϵ } = 1 − exp − 2 ϵ − 1 ¯ γ r k u m (2) 2.2. NOMA transmission The outage analysis for the cooperati v e NOMA netw ork is detailed belo w . In netw orks util izing orthogonal transmission, an outage occurs when the link capacity drops belo w the required data rate. In the case of NOMA, where transmission occurs from R k to users, it is necessary for the relay R k to ha v e recei v ed the pack ets from both users before NOMA can be implemented. As stated in [19], NOMA can only be utilized if the link between S and R k is capable of simultaneously trans mitting both pack ets. F or the S → R k link, if it meets, C sr k ( t ) ≥ 2 ϵ, (3) this mak es the outage of the S − R k link P { log 2 (1 + γ sr k ( t )) < 2 ϵ } = 1 − exp − 2 ϵ − 1 ¯ γ sr k . Con v ersely , for the R k → U m link, where m = 1 or 2 , NOMA enables the simultaneous transmission of pack ets to both U 1 and U 2 . The combined NOMA symbol at R k is e xpressed as (4), x r k ( t ) = √ α x r k , 1 ( t ) + √ 1 − α x r k , 2 ( t ) , (4) where x r k , 1 ( t ) and x r k , 2 ( t ) are data for users U 1 and U 2 respecti v ely , and 0 ≤ α ≤ 1 is the po wer allocation f actor . Then the recei v ed signal at U m is gi v en by (5), y m ( t ) = p α P t h r k u m ( t ) x r k , 1 ( t ) + p (1 − α ) P t h r k u m ( t ) x r k , 2 ( t ) + n m ( t ) , m = 1 , 2 , (5) where n m ( t ) is the noise at us er U m . When NOMA is applied, when γ r k u 1 ( t ) > γ r k u 2 ( t ) , the SNR to dec od e x r k , 2 ( t ) at U 2 is gi v en by (6), S I N R ( x r k , 2 ( t )) = (1 − α ) γ r k u 2 ( t ) α γ r k u 2 ( t ) + 1 . (6) Because γ r k u 1 ( t ) > γ r k u 2 ( t ) , x r k , 2 ( t ) can also be decoded at U 1 if it can be decoded at U 2 . Dropping x r k , 2 ( t ) from the arri v ed signal at U 1 by SIC, the suf cient SNR to decode x r k , 1 ( t ) at U 1 is gi v en by (7), S N R ( x r k , 1 ( t )) = α γ r k u 1 ( t ) . (7) F ollo wing similar procedures as those in [21], the condition that there e xists an α to support NOMA transmis- sion to both U 1 and U 2 (i.e. log 2 (1 + S I N R ( x r k , 2 ( t )) ≥ η and log 2 (1 + S N R ( x r k , 1 ( t )) ≥ η ) is gi v en by (8), (9), (1 − α ) γ r k u 2 ( t ) α γ r k u 2 ( t ) + 1 ≥ 2 η − 1 , (8) α γ r k u 1 ( t ) ≥ 2 η − 1 , (9) from (8) and (9), 2 η − 1 γ r k u 1 ( t ) ≤ α ≤ 1 2 η (1 − 2 η − 1 γ r k u 2 ( t ) ) , (10) Buf fer s balancing of b uf fer -aided r elays in ... (Mohammad Alkhwatr ah) Evaluation Warning : The document was created with Spire.PDF for Python.
1778 ❒ ISSN: 2088-8708 γ r k u 2 ( t ) ≥ (2 η − 1) γ r k u 1 ( t ) γ r k u 1 ( t ) − 2 η (2 η − 1) , if γ r k u 1 ( t ) > γ r k u 2 ( t ) . (11) Similarly , if γ r k u 1 ( t ) < γ r k u 2 ( t ) , NOMA condition becomes, γ r k u 1 ( t ) ≥ (2 η − 1) γ r k u 2 ( t ) γ r k u 2 ( t ) − 2 η (2 η − 1) . (12) If the signal-to-noise ratio (SNR) for the R k → U m links ( m = 1 or 2 ) is insuf cient to meet the conditions in (11) or (12), then NOMA transmission either becomes unfeasible or inef cient. In such cases, if C r k u m ( t ) > η , OMA can be uti lized to transmit a single pack et to U m . By follo wing procedures similar to those outlined in [22], we deri v e, P k , ( m,n ) = 1 ¯ γ r k u m e − (2 η − 1) ¯ γ r k u m +(2 2 η − 2 η ) ¯ γ r k u n ¯ γ r k u m ¯ γ r k u n Z ∞ 2 η − 1 e − x ¯ γ r k u m − 2 η (2 η − 1) 2 ¯ γ r k u n x d x − ¯ γ r k u n ¯ γ r k u m + ¯ γ r k u n e − (2 η − 1)( ¯ γ r k u m + ¯ γ r k u n )(2 2 η +2 η ) ¯ γ r k u m ¯ γ r k u n , (13) where ( m, n ) ∈ { (1 , 2) , (2 , 1) } . Although NOMA transmission brings benets to the netw ork by i n c reasing the throughput, it imposes some dif culties by raising the required channel g ain to a v oid outage (abo v e 2 ϵ ). Adding to this outage po- tential, the outage caused by empty and full b uf fers raises the outage to unacceptable le v els. T o reduce the occurrence of empty and full b uf fers, this paper proposes the balancing of the b uf fers by mo ving pack ets from more full b uf fers to emptier b uf fers. The balancing can be done on dif ferent le v els. F or instance, the balancing is performed when the netw ork is in outage so no more b urden is added to the netw ork by balancing. This le v el of balancing is preferable when bad channel g ains are the dominant cause of the outage. On the other hand, it is desirable to gi v e higher priority for the balancing in good channels g ains where the fullness and emptiness of b uf fers is the dominant cause for outage, more details on this in section 4. 3. PERFORMANCE AN AL YSIS Buf fer -aided relays play a k e y role in reducing the outage probabili ty , which in turn impro v es system throughput. Ho we v er , when the b uf fers are either completely full or empty , the system performance suf fers, leading to a higher probability of an outage. This section presents an analytical comparison highlighting the benets of a bal anced b uf fer -aided relay system o v er an unbalanced one. In the conte xt of each b uf fer -aided relay , the number of stored data pack ets determines its state. Assuming there are K relays, each with a b uf fer size of L , there are ( L + 1) K distinct possible states. Each of these states af fects the a v ailability of the S − R k and R k − U m links. Specically , the S − R k link is a v ailable when the recei ving b uf fer is not full, while the R k − U m link is a v ailable when the transmitting b uf fer is not empty . The state v ector for the l -th state is dened as (14), s ( l ) = [ s ( l ) 1 , s ( l ) 2 , · · · , s ( l ) k ] , l = 1 , · · · , ( L + 1) k , (14) where s ( l ) k is the length of the b uf fer at R k at state s ( l ) . Considering all possible states, the outage probability is dened as the lik elihood that the system remains in the same s tate, implying that no communication (either transmission or reception) tak es place during the current time-slot. Hence, the outage probability for the b uf fer -aided system can be e xpressed as (15), P out = ( L +1) K X i =1 P s ( i ) out π i . (15) where π i denotes the stationar y probability of state s ( i ) , and P s ( i ) out represents the outage probability at state s ( i ) . Int J Elec & Comp Eng, V ol. 15, No. 2, April 2025: 1774-1782 Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Elec & Comp Eng ISSN: 2088-8708 ❒ 1779 In b uf fer -aided relays, the b uf fer states are modeled as a discrete-time Mark o v chain, with the transi- tion matrix A capturing the state transitions within a space of ( L + 1) M × ( L + 1) M . The entry A ij represents the probability of transitioning from state s ( j ) at time t to state s ( i ) at time t + 1 : A ij = P ( X t +1 = s ( i ) | X t = s ( j ) ) (16) The Mark o v chain is both irreducible and aperiodic. A chain is irreduci ble if all states are reachable from an y other state, and aperiodic if there is a nonzero probability of remaining in an y gi v en state, as discussed in [25]-[26]. As sho wn in [27], for an irreducible and aperiodic Mark o v chain, the stationary state probability v ector can be calculated as (17), π = ( A − I + B ) − 1 b, (17) where π = [ π 1 , π 2 , · · · , π ( L +1) ] is the stationary probability v ector , π i is the probability of being in state s i , b = [1 , · · · , 1] T , I is the identity matrix, and B is an ( L + 1) × ( L + 1) matrix lled with ones. Using this, we can determine the outage probability of the b uf fer -aided relay system wh e n the Mark o v chain remains in the same state as (18): P out = ( L +1) M X i =1 π i A ii (18) where A ii are the diagonal elements of A . Assuming that all links are independent and identically distrib uted (i.i.d.), if balancing is successfully applied to pre v ent b uf fer undero w or o v ero w , the outage probability of an y gi v en state is, (1 − e − 2 2 ϵ − 1 ¯ γ S R k ) O S R k s ( l ) × ( P k , ( m,n ) ) O R k U m s ( l ) (19) where O S R k s ( l ) denotes the number of a v ailable S − R k links at state s ( l ) , and O R k U m s ( l ) is the number of a v ailable R k − U m links at state s ( l ) . On the other hand, if the balancing is not applied, for each full b uf fer the number O S R k s ( l ) is reduce d by one. Similarly , for each empty b uf fers, the number O R k U m s ( l ) is reduced by one. Since we are dealing with probabilities the numbers (1 − exp ( − 2 2 ϵ − 1 ¯ γ S R k )) and ( P k , ( m,n ) ) are less than one, so decreasing there po wers ( O S R k s ( l ) and O R k U m s ( l ) respecti v ely) increases t he outcome of (19), which increases the outage probabil ity of the netw ork. This sho ws the benets of a v oiding empty or full b uf fers via balancing. 4. SIMULA TION RESUL TS This section discusses the results of the e xperimental simulations conducted to v alidate the analysis presented earlier . W e e v aluate the ef fecti v eness of the proposed interconnection between b uf fer -aided relays in a cooperati v e NOMA netw ork. F or the simulations, we assume that the noise v ariance, σ 2 , is normalized to 1, and we adopt the data rate ϵ = 2 bps/Hz, as suggested in [22]. Additionally , the b uf fer size is set to L = 5 . Firstly , we sho w the ef fect of the balancing of the b uf fer content on the outage probability . The pos iti v e impact of the balancing on the outage probability is ob vious in Figure 2. Figure 2 s ho ws the comparison between balanced and non-balanced cases for tw o relays and three relays netw orks. The balancing enhances the netw ork performance in the tw o cases. It is w orth noting that with more relays the importance of the balancing increases, this can be observ ed by noticing a higher impact of the balancing on the three relays case. This is true because a v oiding empty or full b uf fer increases the de gree of freedom, hence the di v ersity g ain is increased as well. T aking into account a higher number of a v ailable b uf fers leads to a higher de gree of freedom. F or instance, to get a 0.1 outage probability in the tw o relay cases, the required SNR is about 13.5 dB and 14.5 dB for non-balancing and balancing cases respecti v ely . So the reduction in the required SNR is 1 dB. If we do the same comparison in the case of three relays, the reduction is abo v e 3 dB which is higher than that of the tw o relays case. Buf fer s balancing of b uf fer -aided r elays in ... (Mohammad Alkhwatr ah) Evaluation Warning : The document was created with Spire.PDF for Python.
1780 ❒ ISSN: 2088-8708 Figure 3 stresses the importance of the balancing as it sho ws the impact of the balancing on the system throughput. F or the 3 relay case the throughput impro v ement can be higher than 0.5 pack et per time-slot at 10 dB SNR. Based on the proportionality between outage probability and throughput, we can infer that the impro v ement of the balancing becomes more ef fecti v e by adding more relays to the netw ork, similar to what happened with the outage probability in Figure 2. Figure 2. The impact of the b uf fer content balancing on the outage probability in tw o and three relays netw orks Figure 3. Throughput comparison of balancing and non-balancing in three relays netw ork Another signicant benet of appl ying balancing in b uf fer -aided relays is achie ving the performance of a lar ge number of relays with a minimal number o f relays. This reduction in number of relays reduces the netw ork comple xity and cost as well. The decline in outage probability can be achie v ed by increasing the number of relays, see Figure 2. Ho we v er , the same ef fect can be realized with the balancing as sho wn in Figure 4. The non-bal ancing three relays outperform the balancing tw o relays. But, as mentioned abo v e, the ef fecti v eness of the balancing increases with more relays, this is e xactly the case in Figure 4, the balancing three relays outperform the non-balancing four relays netw ork. After studying the benets of considering the balancing of the b uf fer content, it is important to study the impact of applying the balancing in dif ferent approaches. T w o mechanisms are sho wn in Figure 5. In the rst one, the balancing is prioritized o v er the rece i ving, which means no recei ving (relays recei v ed from the source) can tak e place before the content of the b uf fer is balanced. As in [28], gi ving higher priority to the transmission from relays to the users and lo wer priority to the recei ving (relays recei v e from source) is im- portant to enhance the outage probabili ty of the netw ork. Therefore, in prioritize balancing, the transmission has the highest priority then the balancing is gi v en a higher priority than the recei ving. It is w orth noting that the balancing itself causes the netw ork to be outage according to the outage denition where the netw ork is outage in a specic time slot if no data pack ets are recei v ed or transmitted by relays during the time slot. An- other mechanism is to gi v e higher priority to transmission and recei ving and perform balancing only when no transmission or recei ving is possible. Figure 5 sho ws that prioritizing recei ving and performing balancing only when the netw ork is in an outage is better than gi ving higher priority to balancing. This is true as prioritize bal- ancing causes outages and gi ving balancing lo wer priority to happen only whe n the netw ork is in outage which causes no more outages. If it is feasible to dedicate a separate channel for communication between relays, then balancing can be al w ays performed (not in outage only) without causing more outages. Int J Elec & Comp Eng, V ol. 15, No. 2, April 2025: 1774-1782 Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Elec & Comp Eng ISSN: 2088-8708 ❒ 1781 Figure 4. Outage probability comparison between the non-balancing 4 relays netw ork and the balancing 3 relays netw ork Figure 5. The impact of prioritizing reception or balancing on the outage probability of the netw ork 5. CONCLUSION This study proposes emplo ying balancing in b uf fer -aided relays in cooperati v e NOMA netw orks. This is ur ged due to the performance limitations of b uf fer -aided relays when relays cannot recei v e or transmit with full or empty b uf fers respecti v ely . The proposed balancing technique impro v es the netw ork performance by making full or empty b uf fers less lik ely to happen. As the number of b uf fer -aided relays is increased, the outage probability is decreased. The im pact of the balancing on the outage probability increases with more relays. In particular , the balancing has an impact similar to and better (in some scenarios) than adding more non-balancing netw orks. In addition, adding more relays could be costly , while the balancing can be performed in the a v ailable resources without added cost. Finally , prioritizing the balancing can be done at dif ferent le v els with dif ferent outcomes. F or instance, gi ving the balancing the lo west priority by performing it only when the netw ork cannot transmit or recei v e is capable of reducing t he outage probability . Allo wing t he balancing to be performed all the time by dedicating a channel for the balancing can achie v e better results. REFERENCES [1] D. Nguyen et al. , “6G internet of things: a comprehensi v e surv e y , ” IEEE Internet of Things Journal , v ol. 9, no. 1, pp. 359-383, 2022, doi: 10.1109/JIO T .2021.3103320. [2] M. Z. Iskandarani, “V ehicular connecti vity analysis using enhanced qual ity slotted ALOHA (EQS-ALOHA), ” Science and Infor - mation Conference , 2024, v ol 1019, pp. 484-509, doi: 10.1007/978-3-031-62273-1-31. [3] M. Alkha w atrah, ”Cooperati v e NOMA based on O AM transmission for be yond 5G applications, ” Computer Systems Science and Engineering , v ol. 45, no. 2, pp. 1187–1197, 2023, doi: 10.32604/csse.2023.030699. [4] Z. Ding, M. Peng, and H. V . Poor , “Cooperati v e non-orthogonal multiple acces s in 5G systems, ” IEEE Communications Letters , v ol. 19, no. 8, pp. 1462–1465, 2015, doi: 10.1109/LCOMM.2015.2441064. [5] L. Lv , J. Chen, and Q. Ni, “Cooperati v e non-orthogonal multiple access in cogniti v e radio, ” IEEE Communications Letters , v ol. 20, no. 10, pp. 2059–2062, Oct 2016, doi: 10.1109/LCOMM.2016.2596763. [6] M. Alkha w atrah, “Buf fer -aided cooperati v e millimeter w a v es for IoT netw orks, ” Journal of Electrical and Computer Engineering , v ol. 1, 2022, doi: 10.1155/2022/7474679. [7] M. Alkha w atrah, “Buf fer -aided cooperati v e relays in orbital angular momentum based IoT netw orks, ” Journal of Communications , v ol. 19, no. 4, 2024. [8] M. Alkha w atrah, “Ener gy-harv esting cooperati v e NOMA in IoT netw orks, ” Modelling and Simulation in Enginee ring , 2024, doi: 10.1155/2024/1043973. [9] H. S. Ghazi and K. W . W esoło wski, “Impro v ed detection in successi v e interference cancellat ion NOMA OFDM recei v er , ” IEEE Access , v ol. 7, pp. 103325-103335, 2019, doi: 10.1109/A CCESS.2019.2931809. [10] J. G. Andre ws et al. , “What W ill 5G Be?, ” IEEE Journal on Selected Areas in Communications , v ol. 32, no. 6, pp. 1065-1082, June 2014, doi: 10.1109/JSA C.2014.2328098. Buf fer s balancing of b uf fer -aided r elays in ... (Mohammad Alkhwatr ah) Evaluation Warning : The document was created with Spire.PDF for Python.
1782 ❒ ISSN: 2088-8708 [11] A. Nosratini a, T . Hunter , and A. Hedayat, “Cooperati v e communication in wireless netw orks, ” IEEE Communications Mag azine , v ol. 42, no. 10, pp. 74–80, 2004, doi: 10.1109/MCOM.2004.1341264. [12] J . Laneman, G. W ornell and D. Tse, “ An ef cient protocol for realizing cooperat i v e di v ersity in wireless netw orks, ” in Proceedings. 2001 IEEE International Symposium on Information Theory , 2001, doi: 10.1109/ISIT .2001.936157. [13] A. Bletsas, H. Shin and M. W in, “Cooperati v e communications with outage-optimal opportunistic relaying, ” IEEE T ransactions on W ireless Communications , v ol. 6, no. 9, pp. 3450–3460, 2007, doi: 10.1109/TWC.2007.06020050. [14] M . Z. Iskandarani, “In v estig ation of ener gy consumption in WSNs withi n enclosed spaces using beamforming and LMS (BF-LMS), ” in IEEE Access , v ol. 12, pp. 63932-63941, 2024, doi: 10.1109/A CCESS.2024.3395932. [15] C . T atino, N. P appas, I. Malanchini, L. Ewe and D. Y uan, ”On the benets of netw ork-le v el cooperation in millimeter - w a v e communications, ” in IEEE T ransactions on W ireless Communications , v ol. 18, no. 9, pp. 4408-4424, Sept. 2019, doi: 10.1109/TWC.2019.2924432. [16] Z. Y ang, Z. Ding, Y . W u and P . F an, ”No v el relay selection strate gies for cooperati v e NOMA, ” in IEEE T ransactions on V ehicular T echnology , v ol. 66, no. 11, pp. 10114-10123, No v . 2017, doi: 10.1109/TVT .2017.2752264. [17] Z. Ding, H. Dai, and H. V . Poor , “Relay selection for cooperati v e NOMA, ” IEEE W ireless Communications Letters , v ol. 5, no. 4, pp. 416–419, Aug 2016, doi: 10.1109/L WC.2016.2574709. [18] N. Nomik os et al. , “ A surv e y on b uf fer -aided relay selection, ” IEEE Communications Surv e ys T utorials , v ol. 18, no. 2, pp. 1073–1097, 2016, doi: 10.1109/COMST .2015.2512441. [19] M. Alkha w atrah, Y . Gong, G. Chen, S. Lambotharan, and J. A. Chambers, “Buf fer -aided relay selection for cooperati v e NOMA in the internet of things, ” IEEE Internet of Things Journal , v ol. 6, no. 3, pp. 5722–5731, 2019, doi: 10.1109/JIO T .2019.2905169. [20] S. Luo and K. C. T eh, “ Adapti v e transmission for cooperati v e NOMA system with b uf fer -aided relaying, ” IEEE Communications Letters , v ol. 21, no. 4, pp. 937–940, April 2017, doi: 10.1109/LCOMM.2016.2647250. [21] S. Luo and K. C. T eh, “Buf fer state based relay selection for b uf fer -aided cooperati v e relaying systems, ” IEEE T ransactions on W ireless Communications , v ol. 14, no. 10, pp. 5430–5439, No v . 2015, doi: 10.1109/TWC.2015.2438296. [22] Q. Zhang, Z. Liang, Q. Li, and J. Qi n, “Buf fer -aided non-orthogonal multiple access relaying systems in rayleigh f ading channels, ” IEEE T ransactions on Communications , v ol. 65, no. 1, pp. 95–106, Jan 2017, doi: 10.1109/TCOMM.2016.2630050. [23] V . Jamali, N. Zlatano v , and R. Schober , “Bidir ectional b uf fer -aided relay netw orks with x ed rate transmission—part II: Delay-constrained case, ” IEEE T ransactions on W ireless Communications , v ol. 14, no. 3, pp. 1339–1355, March 2015, doi: 10.1109/TWC.2014.2365806. [24] Z. W ei, J. Guo, D. W . K. Ng and J. Y uan, ”F airness comparison of Uplink NOMA and OMA, ” in 2017 IEEE 85th V ehicular T echnology Conference (VTC Spring) , 2017, pp. 1-6, doi: 10.1109/VTCSpring.2017.8108680. [25] Norris and R. James, “Mark o v chains, ” Cambridge Uni v ersity Press., 1998. [26] A. Berman and R. Plemmons, “Nonne g ati v e matrices in the mathematical sciences, ” Society for Industrial and Applied Mathematics , 1994. [27] I. Krikidis, T . Charalambous and J. Thompson, “Buf fer -aided relay selection for cooperati v e di v ersity systems with- out delay constraints, ” IEEE T ransactions on W ireless Communications , v ol. 11, no. 5, pp. 1957-1967, May 2012, doi: 10.1109/TWC.2012.032712.111970. [28] Z. T ian, Y . Gong, G. Chen, and J. A. Chambers, “Buf fer -aided relay selection with reduced pack et del ay in cooperati v e netw orks, ” IEEE T ransactions on V ehicular T echnology , v ol. 66, no. 3, pp. 2567–2575, Mar . 2017, doi: 10.1109/TVT .2016.2573378. BIOGRAPHIES OF A UTHORS Mohammad Alkhawatrah recei v ed the B.S. and M.S. de grees in communication en- gineering from Al-Ahliyya Amman Uni v ersity (AA U), Amman, Jordan, in 2008 and 2016, respec- ti v ely . He recei v ed the Ph.D. de gree from the signal processing and netw orks resear ch Group in 2020 from W olfson School of Mechanical, Electrical and Manuf acturing Engineering at Loughborough Uni v ersity , Loughborough, U.K. He is currently an associate profes sor in Electronic and Commu- nication Department in Al-Ahliyya Amman Uni v ersity . His research interests include b uf fer -aided relays, non-orthogonal multiple access, relay selection, machine learning, AI, cooperati v e netw orks and signal processing. He can be contacted at email: M.alkha w atrah@ammanu.edu.jo. Nidal Qasem recei v ed his B.Sc. de gree in electronics and communications engineering (Honours) from Al-Ahliyya Amman Uni v ersity , Amman, Jordan, in 2004. He obtained his M.Sc. de gree in digital communication systems for netw orks mobile applications (DSC) in 2006, fol- lo wed by a Ph.D. in wireless and digital comm unication systems, both from Loughborough Uni- v ersity , Loughborough, United Kingdom. He currently holds the position of full professor in the Department of Communications and Computer Engineering at Al-Ahliyya Amman Uni v ersity . His research interests include propag ation control in b uildings, specically impro ving the recei v ed po wer , FSS measurement s and designs, antennas, ultra-wide band, orbital angular momentum, and wireless system performance analyses. He is a senior member of the IEEE. He can be contacted at email: Ne.qasem@ammanu.edu.jo. Int J Elec & Comp Eng, V ol. 15, No. 2, April 2025: 1774-1782 Evaluation Warning : The document was created with Spire.PDF for Python.