On the performance of STBC-NOMA assisted overlay cognitive system under CEEs and imperfect SIC

Abstract

In future wireless networks, the synergy of non-orthogonal multiple access (NOMA) and cognitive radio (CR) can provide higher spectral efficiency and increased data rate. In this context, this article studies the performance of a space-time block code-aided overlay cognitive NOMA system based on Alamouti's (2 x 1, multiple-input single-output) code under the practical consideration of channel estimation errors (CEEs) and imperfect implementation of successive interference cancellation (SIC) at the receiver. We evaluate the closed-form expressions, after approximation, for the system throughput (ST) and outage probability (OP) over Rayleigh fading channels for both the secondary and primary networks. Asymptotic expressions of the OP and ST at high signal-to-noise ratio region are also evaluated to obtain further insights into the proposed scheme. Further, the article investigates the impact of various parameters on the OP and ST of the proposed scheme. Moreover, the OP of STBC orthogonal multiple access (OMA) based overlay cognitive network is also determined and compared with the presented network. Finally, simulations are used to validate the performance analysis of the proposed network and the accuracy of the derived analytical results. The simulation result shows that the presented system achieves significant gain in terms of the OP and ST compared to the STBC-OMA based overlay cognitive system and a recently proposed scheme under imperfect SIC and CEE conditions.

Introduction

In future wireless communication networks, among various multiple access technologies, non-orthogonal multiple access (NOMA) has been considered as a key technology. NOMA has the capability to improve connectivity and provide essential advantages like larger network capacity, increased data rate, and higher spectrum efficiency [1], [2], [3]. Particularly, in case of power domain NOMA technique, at the transmitter, symbols of various users are superimposed after allocating dissimilar powers and the multiplexed signal is forwarded on the same time and frequency resource. The assigned power level can be allocated generally depending upon the channel conditions. At the receiver, the symbols are separated and decoded after employing successive interference cancellation (SIC) process [4]. However, the improved spectral efficiency in NOMA typically comes by compromising in the bit error rate (BER), due to the interference effect [5].

Cognitive radio (CR) is another prominent concept in which the secondary network can be allowed to opportunistically or collaboratively access the licensed frequency band of the primary network to improve spectrum efficiency. CR has received significant attention as a concept to enhance the performance of upcoming technologies [6]. Basically, CR operates under three modes, namely, the interweave CR, the underlay CR, and the overlay CR. In interweave CR, the secondary network transmits only when the frequency band is not utilized by the primary network. In underlay CR, the secondary and primary networks transmit simultaneously given that the secondary network interference is tolerable by the primary network. In overlay CR, the primary network collaborates with the secondary network to relay its messages reliably. Therefore, it may be observed that spectrum sharing is the common principle between NOMA and CR, due to which researchers attempt to combine CR and NOMA technologies to further improve the spectrum sharing. The integration of CR and NOMA is summarized in the following subsection.

Integrating NOMA with CR, denoted as CR-NOMA, has the ability to achieve massive connectivity, low latency, high data rate, and high throughput. The possibility of combination of two CR modes, namely, the overlay and underlay integrated with NOMA are proposed in [7]. Further, in [8], authors proposed the challenges and the benefits of integrating NOMA with CR networks. Therefore, CR-NOMA expects to attain more intelligent sharing of spectrum.

In underlay CR-NOMA systems, the secondary network can communicate with the multiple users using NOMA technique to improve its performance satisfying the primary network's interference constraint. Authors in [9] presented an underlay CR-NOMA scheme based on cooperative relaying, where a secondary transmitter broadcast a NOMA signal to the secondary receivers subject to the restraint that the interference at the primary's receiver is less than a threshold. In [10], the application of NOMA with power harvesting was analyzed in cooperative underlay CR networks, in which constant power allocation is used at the secondary transmitter for the NOMA principle, while the relay uses a power splitting scheme to harvests power from the secondary transmitter. A comparison of decode-and-forward (DF) with amplify-and-forward (AF) relaying techniques in underlay CR-NOMA network was presented in [11] and [12], respectively. The analyses in these articles have revealed that for underlay CR networks, NOMA can outperform traditional orthogonal multiple access (OMA). However, compared with conventional NOMA systems, underlay NOMA systems need to handle challenges of more strict interference management [9], [10], [11], [12].

In overlay CR-NOMA systems, unlike underlay CR-NOMA systems, the secondary network works as a relay in order to help forward the primary transmitter's symbols. As a reward, the secondary transmitter gets permission to simultaneously forward its information symbol with the help of the NOMA principle [7]. Therefore, a trade-off between reception reliability and efficiency of the spectrum can be obtained by employing overlay CR-NOMA networks. CR-NOMA systems in overlay mode have been studies in [13], [14], [15], [16], [17], [18], [19], [20]. Particularly, an overlay CR-NOMA system was presented in [13], in which the unlicensed or secondary user, in lieu of utilizing the licensed or primary user's frequency band, performed as a relay for the primary user. In [14], a NOMA-assisted overlay multi user CR network is studied, where a secondary network is selected in order to re-transmit the primary's symbols and transmit its own symbols as well using the NOMA technique. Furthermore, in [15], in place of half-duplex (HD) relay used in [13], relay works in full-duplex (FD) mode. The authors in [16] analyzed an overlay CR-NOMA system, in which only one primary receiver at the ground is considered without a direct channel between the primary satellite transmitter and ground station. Furthermore, authors in [17] presented an overlay CR-based satellite-terrestrial system, where a primary satellite transmitter uses the NOMA principle to serve all its primary users, while the secondary transmitter accesses the spectrum in lieu of assisting the primary's communication by cooperative relaying. In [18], CR-NOMA network in overlay framework for multiple receivers in CR networks was investigated, where the secondary network helps in communicating the primary's signal and transmit its own signal as well by employing the NOMA concept. Moreover, in [19], a NOMA-based spectrum-sharing system with cooperative relaying is analyzed, where one of the secondary transmitters, working in FD mode, is picked for transmitting the primary's symbol and its own signal to the secondary and primary users, respectively. However, in [20], a NOMA-based overlay cognitive scheme was proposed, in which various assumptions of the SIC conditions at the secondary and primary receivers are considered depending on the channel conditions. Compared with CR-NOMA working in underlay mode, there is no interference restraint in CR-NOMA working in overlay mode and the outage probability (OP) of the primary user is decreased because of higher diversity gain achieved by the secondary network relaying.

Although, the combination of NOMA with CR is advantageous, the problem of interference mutually between the secondary and primary networks due to transmission on the same spectrum needs to be tackled. However, NOMA uses SIC to mitigate the interference, but since the signals of the weak user may be recovered incorrectly at the near user, the SIC realization at the receiver may be performed imperfectly [1]. The analysis in [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20] considers the perfect implementation of SIC, however, practically the perfect SIC is difficult to realize. The outage performance of a cooperative CR-NOMA system in underlay mode with imperfect SIC has been studied in [21]. The OP of CR-NOMA system working in overlay mode with imperfect SIC was analyzed in [22] and the proposed system is proven to achieve better performance than the CR-NOMA network working in underlay mode.

In addition to this, the perfect information of channel condition, at the receiver, is still a main challenge because of limited overhead of pilot signals in time-division multiplex networks and the finite capacity of feedback channels in frequency-division multiplex networks. However, in recent years, limited works have investigated the impact of imperfect channel state information (CSI), at the receiver, on CR-NOMA systems. For instance, the OP of a downlink cooperative CR-NOMA system in underlay mode with channel estimation error (CEE) has been studied in [23]. Further, authors in [24], proposed an AF cooperative overlay CR-NOMA network, and the analysis was performed under practical conditions of CEEs and imperfect SIC.

The analyses in overlay CR-NOMA-based literature [13], [14], [15], [16], [17], [18], [19], [20] have shown that due to single antenna systems, the maximum diversity order of two can be gained for the primary network when secondary network works as a relay. Keeping in mind the demands of beyond 5G networks, the reliability and performance of the primary networks should be of main concern. So, to increase the diversity gain of the system, researchers have integrated NOMA with Alamouti's space-time-block-code (STBC), denoted as STBC-NOMA, owing to the low decoding complexity and to achieve full diversity without sacrificing the data rate. OP and ergodic sum rate of the cooperative relay network assisted by STBC-NOMA have been analyzed in [25]. In [26], to achieve the advantages of both diversity and spectral efficiency, the authors study the OP of the integration of the NOMA with conventional Alamouti's coding scheme over Nakagami-m fading environment. Further, in [27], the authors evaluated the performance of a STBC-assisted cooperative NOMA network under practical conditions such as imperfect SIC, CEEs, and imperfect timing synchronization between the users. Moreover, authors in [28], investigates STBC-NOMA technique in overlay CR network under the condition of imperfect SIC. In [29], the STBC-NOMA network was analyzed in which diversity gain with lower SIC overhead compared to the conventional cooperative NOMA technique was achieved. In this article, the impact of imperfect CSI on the system performance, which is more realistic in real-time processing was investigated. Furthermore, the authors in [30] analyzed the ergodic rate, OP as well as the energy efficiency of the cooperative NOMA concept in underwater acoustic sensor networks by considering the underwater specific characteristics, such as acoustic spreading, propagation loss, and fading effects for both shallow and deep water scenarios, under imperfect CSI as well as imperfect SIC. From the above works of literature, the STBC-NOMA system models are basically categorized into two types: (1) one transmitter equipped with two transmit antennas forward signals simultaneously using STBC and (2) two transmitters with single antenna transmit signals simultaneously according to STBC. In the first type, there is no time offset between the symbols received at the receiver. However, in the second type, time synchronization between the transmitters is required, which is practically very difficult to achieve.

Motivated by the advantages achieved by implementing STBC in the cooperative NOMA networks and improvement in the performance of the primary network performance with the help of the secondary network relaying in overlay CR-NOMA systems, this article presents an STBC-NOMA assisted overlay cognitive network under CEEs and imperfect SIC. So, in this work, $2\times 1$ Alamouti's STBC and the overlay cognitive radio with NOMA are integrated together. Specifically, in the 1st phase, the primary transmitter broadcast two symbols with the help of Alamouti's STBC using two transmit antennas in 1st and 2nd time-slots. If the primary's symbols are recovered successfully at the secondary transmitter in the 1st phase, then the secondary transmitter superimposes its own symbols with primary user's symbols using the NOMA technique. In the 2nd phase, the secondary transmitter forwards the multiplexed symbols using the Alamouti's STBC by two transmit antennas in 3rd and 4th time-slots. At the primary receiver, maximum-ratio-combining (MRC) is employed to decode primary's symbols, received through a direct link in the 1st phase and through secondary transmitter link in the 2nd phase, by treating secondary's symbols as noise without implementing the SIC. At the secondary receiver, secondary's symbols are decoded after employing the SIC process. However, if the secondary transmitter unable to recover the primary user's symbol in the 1st phase, then it will transmit its own symbol with full power in the 2nd phase using Alamouti's STBC. In this case, the primary's symbol is decoded at the primary receiver, received through direct link only in the 1st phase, and the secondary's symbol is recovered at the secondary receiver, received in the 2nd phase. The proposed system model is summarized in Algorithm 1. Unlike the work in [22], where single antenna nodes are used, the proposed scheme consists of a $2\times 1$ multiple-input single-output (MISO) network, and the outage analysis is carried out under CEEs and imperfect SIC. The key contributions of this work are summed up as follows:

• We investigate the OP and system throughput (ST) of STBC-NOMA assisted overlay cognitive system assuming CEEs and imperfect SIC.

• The closed-form expressions, after approximation, of the OP are derived for both the secondary and primary networks. The analysis shows that, due to the existence of CEEs, an error floor exists in the OP of the system. In addition, asymptotic OP is also evaluated in high SNR region.

• The ST of the STBC-NOMA assisted overlay cognitive network is studied and to maximize the ST of the network, optimal power allocation coefficients are evaluated.

• We evaluate the OP of STBC-OMA based overlay cognitive network for comparison. The performance of the proposed network is compared with that of an STBC-OMA based overlay cognitive system and with the work presented in [22].

• The analytical results are verified through simulation process and the impact of various parameters on the OP and ST is studied.

The remaining article is organized as follows: The signal and system model of the STBC-NOMA assisted overlay cognitive system is explained in detail in section 2. The performance of the presented scheme is investigated in section 3. In Section 4, we evaluate the OP of STBC-OMA based overlay cognitive network for comparison purpose. Section 5, discusses the simulation and numerical results. Lastly, section 6 concludes the article. To make this article convenient for the readers, Table 1 summarizes the notations and symbols used in this article.