Design and performance analysis of relay-based cooperative overlay cognitive radio networks

As the rapid growth of wireless communications to meet the fast growing demand for supporting an increasing number of advanced wireless applications and services, the availability of the radio spectrum becomes extremely scarce. In addition, several studies reported that most of the allocated spectrum is not efficiently utilized. To deal with such a problem, cognitive radio (CR) was proposed as a promising technology for boosting the utilization of the radio spectrum, since it could efficiently resolve the spectrum scarcity versus under-utilization dilemma. In CR networks, unlicensed secondary users (SUs) are allowed to use/share the radio spectrum licensed for primary users (PUs) when they are inactive/active, without degrading the performance of the PUs. Research on CR has been divided into three main spectrum-sharing paradigms: interweave, underlay, and overlay. In the interweave paradigm, SUs are allowed to use the spectrum only when PUs are inactive. In the underlay paradigm, SUs are allowed to share the spectrum with PUs with an acceptable levels of interference at PUs' receivers. Throughout this thesis we consider the third paradigm, the overlay model, in which SUs are assumed to be smart and capable of cooperatively maintain or improve the communication between PUs while also gaining access to the spectrum for their own communication. For further increasing the efficient utilization of the scarce spectrum, various modulation and multiple access techniques have been recently proposed such as spatial modulation (SM) and non-orthogonal multiple access (NOMA). NOMA exploits the power domain and SM exploits the spatial domain to efficiently improve the spectrum utilization and the network capacity. SM was proposed as a spatial multiplexing technique that effectively remove inter-channel interference (ICI) and inter-antenna synchronization in multiple input multiple output (MIMO) systems; by allowing only one transmit antenna to be activated at any time interval (i.e., only one radio frequency chain at any transmit interval). SM increases the spectral efficiency by utilizing the active antenna index for conveying information besides the usage of conventional amplitude and phase modulation (APM) techniques. For further enhancement in the performance of SM, transmit antenna selection (TAS) schemes were proposed for SM systems; specifically capacity optimized antenna selection (COAS) and Euclidean distance optimized antenna selection (EDAS). EDAS was shown to outperform COAS but with increased complexity. On the other hand, NOMA is a promising spectral efficient multi-user access candidate for 5G systems. NOMA supports multiple users within the same time/frequency/code resource block but with different power levels that significantly improves the spectral efficiency over conventional orthogonal multiple access (OMA) schemes. In a two-user scenario, NOMA allocates higher power to the user with poor channel gain and less power to the user with better channel gain. The user with better channel gain uses successive interference cancellation (SIC) for signal detection while the user with the poor channel gain uses direct decoding considering the signal for the other user as interference. The performance of NOMA can be improved by using user-pairing, i.e., by pairing the users with more distinctive channel gains with each other. Various researchers have investigated SM in overlay CR networks, where they considered using SM at a secondary transmitter that uses the APM technique to relay the symbol of a primary transmitter while using the antenna index to transmit its own message to a secondary receiver. The application of NOMA in CR networks has also attracted many researchers to investigate problems related with outage probability, capacity analysis, energy efficiency and power allocation schemes to further improve the system performance. In this thesis, we study the combination of NOMA, SM, and TAS schemes in overlay CR networks. Joining NOMA and SM altogether increases the spectral efficiency of the network by exploiting the power and spatial domains. The use of TAS schemes improves the quality of service provided for the primary receiver and enhances the energy efficiency of the system. We first started by applying NOMA in overlay CR networks in both the uplink and downlink with single antenna users. In chapter 3, we propose a two-way two-phase spectrum sharing protocol for the overlay CR paradigm using NOMA and physical layer network coding (PLNC). In our proposed protocol, not only the SU (working as a relay) maintains the two-way transmission between the PUs, it also establishes a two-way transmission with the other SU that results in higher spectral efficiency and hence improved system capacity over other protocols proposed in literature. Closed form expressions for the outage probabilities for both PUs and SUs have been derived and validated through Monte Carlo simulations. In chapter 4, we studied the combination of NOMA and SM in overlay CR networks. We considered an overlay CR network that consists of multiple antennas primary transmitter (PT), primary receiver (PR), secondary transmitter (ST) and secondary receiver (SR). The proposed protocol allows ST to transmit simultaneously using NOMA and SM to both PR and SR. The usage of NOMA and SM increases the spectral efficiency for both PR and SR. With NOMA, SR uses SIC to detect the signal intended for PR and then cancels it to detect its own signal, while PR uses direct decoding to detect its own signal considering the intended signal for SR as an interference. This decreases the detection complexity considerably compared to the case without NOMA in which the detectors are required to jointly detect both SM symbols at each receiver. Furthermore, to provide higher quality of service (QoS) for PR, we considered the application of TAS schemes at ST in accordance with PR. We first considered the usage of COAS where ST assigns the antennas with highest channel gains with respect to PR for SM transmission toward PR. From PR's side, this will ensure that the primary symbol will be sent over one of the channels with high gain (with respect to PR) and the secondary symbol will be sent over one of the channels with low gain (with respect to PR), which will result in decreasing the secondary interference at PR – compared with the case without antenna selection (AS) – and therefore will enhance the performance of PR. On the other hand, from SR's side, the new ordering of the antennas of ST is still random with respect to SR and hence will not affect the performance of SR on the average. The performance of the proposed protocol is investigated through derivations of upper bounds for the average symbol error probabilities (SEP) at PR and SR and validated through Monte Carlo simulations. In chapter 5, we studied the combination of NOMA, SM and EDAS in overlay CR networks. Employing EDAS at ST in accordance with PR results in decreasing the SEP of PR significantly; since it selects the antenna subset that maximizes the minimum squared Euclidean distance for the SM constellation at PR. Similar to chapter 4, this will not deteriorate the performance of SR since re-ordering of the antennas at ST remains random with respect to SR. We evaluate the performance of the proposed protocol by derivation of an upper bound for the SEP of PR for the case of using BPSK for both PR and SR and selection of two transmit antennas from the transmit antennas at ST. Also we compare the performance of the proposed protocol with that of COAS and with that of without-AS through Monte Carlo simulations. Finally, the use of NOMA and SM in CR increases the spectral efficiency and hence the system capacity by allowing more users to be serviced with higher data rates through the use of power and spatial domains. Introducing TAS schemes to NOMA-SM in overlay CR networks has been shown to improve the performance of the protocol for PR significantly without affecting the performance of SR, resulting in significant savings in the transmit power of the secondary transmitter that improves the energy efficiency of the network. It has also been shown that the performance of the proposed protocol CR-NOMA-SM with EDAS provides much more transmit diversity compared to CR-NOMA-SM with COAS; but with a trade-off of increased complexity which can be reduced by the use of low complexity EDAS algorithms.