1 Introduction
NonOrthogonal Multiple Access (NOMA) facilitates the simultaneous data transmission to multiple users over the same radio resource by allocating power domain multiplexing [1]. NOMA achieved a great interest in industrial and academic level due to high spectral efficiency and serve a significant amount of users as well [2]. Power domain NOMA and Code domain NOMA are two basic types of NOMA [3]. Since most of the existing researches are based on power based NOMA hence power domain NOMA is considered here [45].
In downlink CNOMA, the transmitter transmits a superposed signal to the intended users and each of the users receives that composite signal. Afterward, the cell center user (CCU) performs successive interference cancellation (SIC) to decode its own signal from the received signal [6]. According to the theory of NOMA, the signal power for CEU is significantly higher than CCU hence this is directly decoded at CEU and considering the signal of CCU as noise. Moreover, Cooperative nonorthogonal multiple access (CNOMA) improves reliability and coverage area of a wireless communication system by performing user as a relay or utilizes dedicated relay [78]. In this paper, CCU is considered as a relay to enhance the reliability and coverage area of the cell edge user (CEU)[9].
The main challenge is to perform the relaying operation by CCU without draining out the battery of the user equipment [6,9,14]. This situation can turn off CCU equipment and causes the degradation of the performance of the network due to the absence of the relay link. Hybrid Simultaneous Wireless Information and Power Transfer (HS) can empower the relay operation of CCU [1011] for CNOMA system. In addition, the complex transmits antenna selection (TAS) technique is proposed for Multiple Input Single Output (MISO) based system SWIPT system for CNOMA [1012]. Due to simplicity and energy efficiency, Single Input Single Output (SISO) is considered here. Moreover, complex TAS is not required for SISO based system. In [10], BS to CEU link was idle for the existing HS protocol. Which degrades the performance of the network. So to overcome this efficiency loss, we have proposed data transmission by utilizing this idle link in the proposed enhanced HS (EHS) protocol. In addition, the outage probability of a hybrid SWIPT based CNOMA protocol is investigated only in [1011]. However, the ergodic sum capacity (ESC) and energy efficiency (EE) of the system is not explored extensively here and the mathematical analysis is not performed to validate the simulation results for ESC [10]. To enhance the ESC of hybrid SWIPT protocol, an enhanced hybrid SWIPT protocol (EHS) is proposed in this paper with an efficient CNOMA strategy. In this strategy, the idle link from the base station (BS) to CEU can be utilized to transfer additional symbol to the CEU. As a result, the ESC is enhanced for the DL transmission of the SWIPT protocol. Moreover, EE is also analyzed in this paper for the proposed EHS and conventional HS protocol for CNOMA as well which was not explored in previous literature.
In [10], the selection combining (SC) technique is used at CEU as a combining technique. Maximal ratio combining (MRC) is superior to SC as an effective combining technique for information decoding (ID) [15]. That’s why MRC is considered here at CEU as combining technique. The contributions of this letter are enumerated below:

EHS protocol is proposed and integrated with CNOMA downlink (DL) transmission for two users. MRC is considered at CEU to combine direct and relay link and perform ID effectively.

The ergodic sum capacity (ESC) and outage probabilities (OP) of each user for EHS with CNOMA are investigated and analytical derivations are performed for the Rayleigh fading channel.

The EE is also analyzed for the EHSCNOMA and HSCNOMA to evaluate the performance of the proposed EHS protocol over conventional HS protocol.

By using simulation and analysis, the performance improvement of the proposed EHS protocol with considered SISO based CNOMA DL transmission (EHSCNOMA) over existing HS protocol with SISO based CNOMA DL transmission with SC (HSCNOMA) are analyzed explicitly.
The rest of the paper is organized as follows: Section 2 describes the proposed protocol with system model. Moreover, the system architecture and CNOMA strategy are also described elaborately in this section. Section 3 mathematically analyzes the ergodic sum capacity. Section 4 analyzes the OP for the proposed EHS protocol with the considered model. Section 5 analyzes the ESC for the proposed EHS protocol with the considered model. Section 6 analyzes the EE for the proposed EHS protocol with the considered model. Section 7 exhibits the numerical result analysis. This paper is concluded in Section 8.
2 Proposed Protocol and System Model
The two userbased system model is shown in Fig.1. According to the concept of DL CNOMA [4,1014], the lower power () is assigned for CCU and higher power is assigned () for CEU based on their channel conditions from BS [3]. and are the normalized distances from BS to CCU and CEU respectively. The downlink transmission from BS to the users is performed by two consecutive phases. The transmitted symbols and are dedicated for CEU and is dedicated for CCU. In (), direct signals are transmitted from BS to CCU and CEU respectively. In second phase (), a relay link is transmitted from CCU to CEU for . Relaying link is required here for at CEU to increase the reliability for at CEU. If the high data rate is required for at CEU, higher signal to noise ratio (SNR) and multiple links can provide the facilities without interruption. That is why a suitable combing technique is considered at CEU for . However, relaying link is not considered for or . Because is transmitted to CEU with full power so higher reliability can be ensured for . Moreover, the additional relaying link for consumes more energy for relaying. In addition, is transmitted to CCU which user has comparatively good channel condition to meet the reliability. SISO based transmitter and receiver is considered in Fig.1.
All wireless channels are considering as Rayleigh block flat fading channel in this model. Suppose denotes the fading coefficient of a channel from BS to a user K, where . So
can be modeled as independent and identically distributed (i.i.d.) complex Gaussian random variables with zero mean and variance
. Moreover, denotes additive white Gaussian noise (AWGN) at user K. The variance of the receiver noise is for the AWGN model.To energize the relay operation of CCU and information decoding, improved hybrid SWIPT is integrated with the considered model. In addition, hybrid SWIPT is the combination of time switching (TS) and power splitting (PS) methods as well [1011,18]. Where T is total time duration to perform the DL transmission for CNOMA. So for the TS part, a fraction of time block for energy harvesting is . Firstly by time duration, BS transmits with full power (). So, CCU harvested energy from the received observation and CEU receives and decodes at this stage. PS is applied in the ((1)T/2) duration of time at , the power splitting ratio for PS technique is and the fraction is utilized for energy harvesting (EH) by PS. Moreover, the fraction is used for information decoding (ID) as well at CCU. Furthermore, and symbols are received and decode at CCU and CEU simultaneously from BS. Afterward, CCU exploits all of the harvested energy for information relaying operation by decode and forward (DF) technique [1011]. The DF is performed at the second phase () and ((1)T/2) duration of time. The EHS protocol for CNOMA is illustrated in Fig.2.
2.1 Direct Transmission and Energy Harvesting of Proposed Hybrid SWIPT based CNOMA
According to the concept of DL NOMA, the multiplexed signal is directly transmitted to CCU and CEU which illustrates in Fig.1. Where, ) and . Here, is total transmit power. Moreover, is considered here and is the SNR [10,18]. The multiplexed signal is transmitted to CCU and CEU simultaneously. Assume, , where and are the channel coefficients for CCU and CEU respectively. The channel fading coefficients between BS and CCU is with zero mean and variance . Moreover, the Additive White Gaussian noise is at CCU. Whereas, is the variance of the noise. Similarly, the channel fading coefficients between BS and CEU is with zero mean and variance . Moreover, the Additive White Gaussian noise is at the receiver end. is the variance of the noise. In this paper, assume and total time duration is [1011,18]. At the first phase, energy harvesting is performed by TS at CCU. BS utilizes full power () to transmit to CEU. So, the SNR to receive at CEU is expressed as below:
(1) 
By using the concept of DL CNOMA, and are transmitted to CCU and CEU as well at . Successive Interference Cancellation (SIC) receiving technique is considered at CCU in Fig.1. Hence should be decoded first to decode at CCU according to the theory of SIC receiver [6,1011]. So, the required signal to interference ratio (SINR) to decode at CCU can be expressed as below[10,18]:
(2) 
At the required SNR to retrieve at CCU is expressed as below according to the concept of DL CNOMA [1011,18]:
(3) 
Moreover, the power of is significantly higher than , so can be directly received and decoded at CEU and is treated as noise at CEU and cancel out [11]. In addition, the SNR requirement to receive and decode at CEU is derived as below at [1011]:
(4) 
2.2 Relay Operation by CCU
CCU utilized the harvested energy to perform relay operation for CEU. So, should be satisfied to perform the DF relay operation by CCU. In addition, and denotes the channel coefficients between CCU and CEU for relay operation. The channel for the relay is modeled as an independent complex Gaussian random variable with zero mean and variance is . Hence the received signal at CEU from CCU is expressed as below [1011]:
(5) 
Here, is the reencoded version of . So for , the required SNR for relay link to receive by utilizing the harvested energy is expressed as following equation [7]:
(6) 
Where, . is the transmitted power from CCU to CEU and is energy conversion efficiency. Moreover, plays a vital role to convert the energy from radio frequency signal [78].
2.3 Combining Method at Cell Edge User
Maximal ratio combining (MRC) is considered here at CEU to combine from direct and relay link. Moreover, MRC is used to decode effectively at CEU and increase reliability for the respective symbol [1718]. Hence, the signal to interference and noise ratio (SINR) for the combined signal is given as below:
(7) 
2.4 Channel Capacities
So from (1) the channel capacity for at CEU is expressed as below [10,18]:
(8) 
Consequently, the channel capacity for at CCU from (2) can be derived as following way [10,18]:
(9) 
Based on the MRC technique at CEU for , the channel capacity at CEU is expressed as following equation[1011,18]:
(10) 
So the sum channel capacity of the considered system with proposed EHS protocol can be expressed as below[10,18]:
(11) 
Where E[.] is the mean or expectation operation.
3 Mathematical Analysis of Ergodic Sum Capacity
The ESC for EHSCNOMA with MRC is derived and analyzed in this section. So, from (8) the channel capacity for from BS to CEU for the considered system is written as below [1011,18]:
(12) 
Let, . The CDF of is written as below:
(13) 
So, the ergodic capacity of at CEU can be achieved by using [13,16] as below:
(14) 
Where, and is the exponential integral. Since is the main concern for CCU. So the channel capacity for at CCU for the considered system is expressed as below [10]:
(15) 
Let . The CDF of can be written as below:
(16) 
So by using [13,16], the ergodic channel capacity of at CCU can be achieved as below:
(17) 
Where, and is the exponential integral. Furthermore, from (10) the channel capacity at CEU to receive by using MRC technique is given as below:
(18) 
So by using [1718], the ergodic channel capacity of (18) can be derived as below:
(19) 
Where, , , and is the exponential integral. So the ESC can be achieved by following equation for EHSCNOMA with MRC:
(20) 
Where is the mean or expectation operation.
4 Outage Analysis
Let is the predefined data rate of CCU. Moreover, and are the predefined data rate of CEU. The OP of each user is derived as below:
4.1 Outage Probability of CCU
The outage event of CCU is occur when CCU cannot decode symbol since SIC is performed at CCU. Moreover, if CCU can decode successfully but cannot decode symbol that can cause outage event as well. Hence OP of CCU can be derived in following way[1920]:
(21) 
Where, and .Now we can put, and into (21) to get the desired OP for CCU as below:
(22) 
4.2 Outage Probability of CEU
The OP of CEU can be occur for two different symbols such as and . Since and are transmitted from BS to CEU. At , if CEU cannot decode then outage event is occur at CEU for symbol . So the OP of CEU for can be derived as below:
(23) 
Where, and now we can put in above equation to get the following equation:
(24) 
Moreover, the outage event at CEU for symbol can be happen if CCU or CEU cannot decode successfully. If CCU failed to receive and decode successfully, the outage event is occur at CEU for symbol. In contrast, if CCU can received and decode successfully but cannot decoded successfully by MRC at CEU. This situation also causes outage event at CEU for symbol. So, the outage event occurs at CEU for can be derived as below [1920]:
(25) 
Where, and put and in (25). So the derived equation after putting the values is like as below:
(26) 
5 Energy Efficiency
The harvested energy at CCU by EHS protocol can be derived as below [11],
(27) 
In addition, the transmitted power from CCU to CEU can be expressed as below based on the harvested energy [21],
(28) 
Moreover, the EE can be derived as below for the proposed EHSCNOMA scheme [21],
(29) 
So Eq.29 shows that EE is related to the and .
6 Numerical Results and Discussion
In the result analysis, the lines represent the simulation result and the markers represent analytical results respectively. For simulation purpose and are considered [1011,18]. Moreover, , path loss exponent, and are considered here for simulation purpose as well [24]. Additionally, , , and are also considered for simulation purpose. The collinear outcomes of simulation and analytical results verify the appropriateness of the analysis for both ESC and OP.
6.1 Ergodic sum capacity
Fig.3 illustrates that EHSCNOMA with MRC outplayed HS protocol in case of ESC by transmitting for the EH at CCU and information transfer for CEU. Higher SNR provides better ESC than lower SNR for all cases. Moreover, the analytical result also validates the simulation result as well for the EHSCNOMA with MRC.
Fig.4 demonstrates the effect of the fraction of time block for () on ESC. Moreover, is considered in this case. So the EHSCNOMA with MRC provides linearly increased ESC for . However, the provided ESC for EHSCNOMA with MRC is saturated for . Because for a longer duration of , duration is shorter. So CCU and CEU cannot properly receive and decode the symbols properly for EHSCNOMA with MRC. Though is transmitted successfully from BS to CEU for the proposed protocol hence it provides significantly higher ESC than others in case of considered CNOMA model. Moreover, the ESC of the HSCNOMA with SC [1011] is degraded due to the idle link is not utilized for . In addition, CCU and CEU cannot receive and decode symbols () properly due to very short duration time for decoding. In this paper, ESC is compared with respect to only because this is the dominating factor for HS protocol according to [18].
Fig. 5 shows the comparisons between three cases for ESC with respect to in this case. Moreover, is considered in this case as well. So this is clearly visible that for different values of , EHSCNOMA with MRC provides significantly higher ESC than other cases. The enhancement of ESC for the proposed EHS protocol is provided due to the utilization of unused link for and effective decoding by utilizing MRC at CEU for . However, for the increasing distance between BS and CCU causes degradation of channel condition among them hence the ESC is decreased for all cases.
6.2 Outage Probability
Fig. 6 illustrates that the EHSCNOMA with MRC provides the same OP than HSCNOMA with SC for at CCU. In case of the OP of for CEU, is only transferred with full power to CEU for the EHS protocol. So it is only compared with the analytical result of itself. Moreover, the link for is idle for the existing HS protocol, so the OP among two different cases cannot be compared. The OP of at CEU, EHS protocol with CNOMA provides lower OP than HSCNOMA with SC for lower SNR due to MRC. The analytical results also validate the simulation results as well for the proposed EHS protocol with the considered model. All OP for CCU and CEU are decreased linearly along with the increasing SNR.
Fig.7 demonstrates the effect of the fraction of time block for () on OP of CCU and CEU for different symbols. Moreover, is considered in this case. Fig 8. depicted that for increasing values of decreased the time duration. That is why CCU and CEU cannot receive and decode their corresponding symbols appropriately. As a result, OP is increased linearly for higher values of . Fig.8 also shows that the OP of for CCU increased linearly for an increasing number of for all protocols. According to the EHS protocol, is transmitted directly to CEU from BS with full power. But cannot receive and decode properly at CEU due to the extremely higher time duration of . Moreover, higher values of also reduce the other time duration of EHS protocol significantly. So the OP of at CEU is increased linearly for higher values of . In case of the OP of at CEU, EHSCNOMA with MRC provides less OP than HSCNOMA with SC due to MRC. Though the OP for both cases increases to a great extent for . Because lower time duration of decoding and relaying caused the huge outage event for at CEU.
Fig. 8 shows the effect of on OP of CCU and CEU for different symbols. Moreover, is considered in this case as well. So this is clearly visible that for different values of , EHSCNOMA with MRC provides the same OP as HSCNOMA with SC for at CCU. Though the OP is increased for at CCU because the distance between BS and CCU is increased. Hence the channel condition among them has degraded accordingly. So it cannot decode successfully at CCU. For the OP of at CEU, the EHSCNOMA with MRC provides lower OP than HSCNOMA with SC for MRC at CEU. However, for higher , the OP of at CEU increased linearly due to the channel condition between BS and CCU. The EHSCNOMA with MRC provides the constant OP for at CEU. Because is transmitted directly from BS to CEU with full power. Only is relaying from CCU to CEU for all cases.
6.3 Energy Efficiency
Fig.9 illustrates the EE comparisons between the proposed EHSCNOMA scheme and HSCNOMA scheme as well. As EE is related to the ESC and . The EHS protocol provides higher ESC than conventional HS protocol which is illustrated in Fig. 3. Hence, the EE is significantly higher for the proposed EHSCNOMA compared to HSCNOMA in Fig.9. However, the EE is comparatively much higher for the less SNR but EE is decreased for higher SNR in case of both cases. Because due to the increasing values of ESC with respect to SNR, the is also increased for both cases based on the harvested energy. Moreover, the difference between the achieved EE between EHSCNOMA and HSCNOMA is significantly higher for lower SNR compared to higher SNR.
7 Conclusion
In this paper, the performance of EHSCNOMA has been analyzed, where CCU is used as a relay for CEU. The ESC and OP of the proposed EHSCNOMA with MRC technique are also investigated with their analytical derivations. The analytic results have been validated by the simulation results. Moreover, the advantages of the proposed EHS protocol have been demonstrated by numerical result analysis. To evaluate the system performance, the impact of different parameters on ESC and OP has been investigated and compared with HSCNOMA. Moreover, the EE of the EHSCNOMA is analyzed and compared with HSCNOMA as well. The proposed protocol provides a significantly higher capacity than other protocols. Moreover, the considered CNOMA strategy also improved the OP of CCU and CEU for different symbols. Furthermore, EE for the proposed EHSCNOMA is significantly higher than HSCNOMA as well. However, the decoding complexity and SIC complexity in the considered model is same as existing CNOMA [26]. In the future, the work can be extended by integrating the EHS with amplify and forward relay assisted CNOMA [3].
8 Acknowledgments
This work was supported by Priority Research Centers Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2018R1A6A1A03024003).
References
 [1] [[1]] D. Zou, D. Deng, Y. Rao, X. Li and K. Yu, Relay selection for cooperative NOMA system over correlated fading channel, Physical Communication, 35, 2019, pp. 100702, https://10.1016/j.phycom.2019.04.016.
 [2]
 [3] [[2]] L. Dai, B. Wang, Z. Ding, Z. Wang, S. Chen and L. Hanzo, A Survey of NonOrthogonal Multiple Access for 5G, in IEEE Communications Surveys and Tutorials, 20 (3), 2018, pp. 22942323,https://doi:10.1109/COMST.2018.2835558.
 [4]
 [5] [[3]] Y. Li, Y. Li, Y. Chen, Y. Ye and H. Zhang, Performance analysis of cooperative NOMA with a shared AF relay,IET Communications, 2018, 12(19), 2018, pp. 2438  2447,https://doi:10.1049/ietcom.2018.5415.
 [6]
 [7] [[4]] S. M. R. Islam, N. Avazov, O. A. Dobre and K. Kwak, Power domain nonorthogonal multiple access (NOMA) in 5G systems: Potentials and challenges, IEEE Commun. Surveys Tuts., 2017,19(2), pp. 721742, https://doi:10.1109/COMST.2016.2621116.
 [8]
 [9] [[5]] Z. Ding, X. Lei, G. K. Karagiannidis, R. Schober, J. Yuan and V. K. Bhargava, A survey on nonorthogonal multiple access for 5G networks: Research challenges and future trends, IEEE J. Sel. Areas Commun., 2017, 35(10), pp. 21812195, https://doi:10.1109/JSAC.2017.2725519.
 [10]
 [11] [[6]] G. Song and X. Wang, Comparison of interference cancellation schemes for nonorthogonal multiple access system, Proc. 2016 IEEE 83rd Veh. Technol. Conf., May 2016, pp. 15, https://doi:10.1109/VTCSpring.2016.7504172.
 [12]
 [13] [[7]] J. N. Laneman, D. N. C. Tse and G. W. Wornell, Cooperative diversity in wireless networks: Efficient protocols and outage behavior, IEEE Trans. Inf. Theory, 2004, 50(12), pp. 30623080, https://doi:10.1109/TIT.2004.838089.
 [14]
 [15] [[8]] Md. Fazlul Kader, Mohammed Belal Uddin, S.M.Riazul Islam and Soo Young Shin, Capacity and outage analysis of a dualhop decodeandforward relayaided NOMA scheme, Digital Signal Processing, 2019, 88, pp. 138148, https://doi.org/10.1016/j.dsp.2019.02.014.
 [16]
 [17] [[9]] Y. Liu, Z. Ding, M. Elkashlan and H. V. Poor, Cooperative Nonorthogonal Multiple Access With Simultaneous Wireless Information and Power Transfer, IEEE Journal on Selected Areas in Communications, 2016, 34(4), pp. 938953, https://doi:10.1109/JSAC.2016.2549378.
 [18]
 [19] [[10]] T. N. Do, D. B. da Costa, T. Q. Duong and B. An, Improving the Performance of CellEdge Users in MISONOMA Systems Using TAS and SWIPTBased Cooperative Transmissions, IEEE Transactions on Green Communications and Networking, 2018, 2(1), pp. 4962, https://doi:10.1109/TGCN.2017.2777510.
 [20]
 [21] [[11]] N. T. Do, D. Benevides da Costa, T. Q. Duong and B. An, Transmit antenna selection schemes for MISONOMA cooperative downlink transmissions with hybrid SWIPT protocol, IEEE International Conference on Communications (ICC), Paris, France, 2017, pp. 16, https://doi:10.1109/ICC.2017.7997005.
 [22]
 [23] [[12]] W. Han, J. Ge and J. Men, Performance analysis for NOMA energy harvesting relaying networks with transmit antenna selection and maximalratio combining over Nakagamim fading, IET Communications, 2016, 10(18), pp. 26872693, https://doi:10.1049/ietcom.2016.0630.
 [24]
 [25] [[13]] M. F. Kader and S. Y. Shin, Coordinated Direct and Relay Transmission Using Uplink NOMA, IEEE Wireless Communications Letters, 2018, 7(3), pp. 400403, https://doi:10.1109/LWC.2017.2779778.
 [26]
 [27] [[14]] Z. Ding, M. Peng and H. V. Poor, Cooperative NonOrthogonal Multiple Access in 5G Systems, IEEE Communications Letters, 2015, 19(8), pp. 14621465, https://doi:10.1109/LCOMM.2015.2441064.
 [28]
 [29] [[15]] Y. Alghorani and M. Seyfi, On the Performance of ReducedComplexity Transmit/ReceiveDiversity Systems Over MIMOV2V Channel Model, IEEE Wireless Communications Letters, 2017, 6(2), pp. 214217, https://doi:10.1109/LWC.2017.2661759.
 [30]
 [31] [[16]] M. B. Shahab and S. Y. Shin, A Time Sharing Based Approach to Accommodate Similar Gain Users in NOMA for 5G Networks, IEEE 42nd Conference on Local Computer Networks Workshops (LCN Workshops), Singapore, 2017, pp. 142147, https://doi:10.1109/LCN.Workshops.2017.76.
 [32]
 [33] [[17]] Wei Duan, Jinjuan Ju, Qiang Sun, Yancheng Ji, Zhiliang Wang, Jeaho Choi, Guoan Zhang, Capacity Enhanced Cooperative D2D Systems over Rayleigh Fading Channels with NOMA, CoRR abs/1810.06837 (2018), https://arxiv.org/abs/1810.06837.
 [34]
 [35] [[18]] Ahmed Al Amin, Soo Young Shin, Investigate the Dominating Factor of Hybrid SWIPT Protocol by Performance Analysis of the Far User of Hybrid SWIPT based CNOMA Downlink Transmission, International Conference on Electrical, Computer and Communication Engineering (ECCE 2019), Coxs Bazar, Bangladesh, 2019.
 [36]
 [37] [[19]] M. F. Kader, S. Y. Shin and V. C. M. Leung, FullDuplex NonOrthogonal Multiple Access in Cooperative Relay Sharing for 5G Systems, IEEE Transactions on Vehicular Technology, 2018, 67(7), pp. 58315840, https://doi:10.1109/TVT.2018.2799939.
 [38]
 [39] [[20]] Kader, M. F., Uddin, M.B., Islam, A., Shin, S.Y.,: Cooperative Non Orthogonal Multiple Access with SWIPT over Nakagami m Fading Channels, Trans Emerging Tel Tech, 2018 [Accepted].
 [40]
 [41] [[21]] Y. Huang, M. Liu and Y. Liu, EnergyEfficient SWIPT in IoT Distributed Antenna Systems, IEEE Internet of Things Journal, 5(4), 2018, pp. 26462656, https://doi:10.1109/JIOT.2018.2796124.
 [42]
Comments
There are no comments yet.