Novel quadrature spatial modulation techniques for multiple input multiple-output wireless communication systems

Compared with single-input single-output (SISO) transmission, multiple-input multiple-output (MIMO) techniques have made a breakthrough in terms of the capacity and reliability of the wireless communication systems. The growing demand for high data rates has led researchers to develop new transmission techniques such as spatial multiplexing (SMX) and space-time block coding (STBC) in order to obtain multiplexing and diversity gains, respectively. In traditional SMX techniques, as all antennas are used for transmission at the same time, the increase in the number of transmit antennas, linearly increases the spectral efficiency of the system; however this also exponentially increases the complexity of the optimum decoder. In order to reduce the receiver complexity, suboptimal decoders have been widely considered at the receiver of the SMX systems, which exhibit significant reduction in the error performance of the system. Spatial modulation (SM), has become one of the most popular transmission techniques in the past decade. In SM technique, unlike traditional SMX techniques, only one of the transmit antennas is activated and the information is transmitted through this active antenna to avoid inter-channel interference (ICI) at the receiver. Its attractive advantages such as low cost and energy efficiency, make SM an alternative transmission technique for 5G and beyond wireless networks. Quadrature spatial modulation (QSM) is one of the most recent SM-based schemes, in which a complexM-ary quadrature amplitude modulation (QAM) symbol is expanded into its real and imaginary components and independently transmitted through their corresponding active antennas. In this study, first, high-rate generalised spatial modulation (HR-GSM) scheme is proposed. In HR-GSM, two M-QAM data symbols are independently transmitted through their corresponding active transmit antennas using classical SM transmission approach. In the proposed system, while one of the data symbols is directly transmitted through its corresponding active transmit antenna, the other one is rotated with an angle, which is optimized for the correspondingM-QAM signal constellation, in order to distinguish the transmitted signals at the receiver. In the HR-GSM scheme, the rotation angle for the second M-QAM data symbol is determined by performing an exhaustive search between 0o 􀀀180o at a certain signal to noise ratio (SNR) value. The angle, which provides the best error performance, is determined as the optimum rotation angle of its corresponding M-QAM signal constellations. Through computer simulations, the optimum rotation angle for BPSK, 4-QAM and 8-QAM signal constellations, is determined as 90o, 45o and 60o, respectively. In this study, the error performance of the HR-GSM system is analysed as well as the average bit error probability (ABEP) of the system is obtained theoretically and supported via Monte Carlo simulations for differentM-QAM constellations. Moreover, for different spectral efficiencies and MIMO configurations, the HR-GSM scheme is compared with classical SM, QSM, generalised SM (GSM) and enhanced SM (ESM) transmission schemes. Simulation results show the superior error performance of the HR-GSM scheme over the reference transmission structures in the literature. Second, since the receiver complexity of the optimum QSM detector grows quadratically as the number of transmit antennas is increased, to reduce the optimum detection complexity of the large scale QSM systems, a compressive sensing (CS) based, low complexity detector is proposed. In the proposed detector, CS technique has been used for the signal reconstruction of the QSM transmission vector due to its sparsity property, which has one or two non-zero elements. In this work, for large scale QSM systems, the reconstruction of received signals has been regarded as a quadratic optimization problem and resolved by using computer software. Then, the most probable active antenna set and their corresponding transmission patterns are determined. After all, a maximum likelihood (ML) type detector determines the complexM-ary constellation symbol and the corresponding active antennas of the real and imaginary parts of the data symbol by performing a search over these transmission patterns and M-ary constellation. This algorithm has been also applied to zero forcing (ZF) detector. For different MIMO configurations, spectral efficiency values and active antenna sets, the BER performance of the ML, QP and ZF detectors are compared. The computer simulation results show that the low-complexity QP algorithm has achieved a near-ML performance by significantly reducing receiver complexity. Third, a system called space-time QSM (ST-QSM), which achieves a second order transmit diversity gain by combining the QSM system with different STBC transmission principles, is proposed. In ST-QSM system, the total group of transmit antennas is subdivided into two groups each of which has a number of transmit antennas that is an integer power of two. In the first time slot, using QSM approach, two complex M-QAM data symbols are decomposed into their in-phase and quadrature components and independently transmitted through their corresponding active antennas belonging to these two antenna groups. In the second time slot, Alamouti, ABBA or 8-antenna STBC principle is applied to each group depending on the number of transmit antennas in each of them. In this work, ST-QSM system is designed for MIMO schemes with 4, 6, 8 and 16 transmit antennas. As an upper bound, the ABEP of the ST-QSM system, for different MIMO configurations and spectral efficiency values, is theoretically derived and supported via Monte Carlo simulations. Furthermore, at the same spectral efficiency values, the ST-QSM scheme is compared with the space-time block coded SM (STBC-SM), Alamouti STBC and QSM schemes. It is observed that ST-QSM scheme shows better BER performance than the reference MIMO transmission structures. Finally, a new cooperative relaying protocol for the QSM scheme, which is called two-way QSM, is proposed. In the proposed system, two users exchange their data with each other via a half-duplex relay that receives or transmits data at one time slot. In the first time slot, using QSM transmission principle, the users at the same rate transmit their-own data to the relay. In the second time slot, by applying bit-wise exclusive OR (XOR) operation to the detected transmission vectors of the users, a new QSM transmission vector is constructed and retransmitted to the users. After detection processes, by applying bit-wise XOR operation to its own data and the detected data transmitted by the relay, each user extracts the data of the other user. Also, the proposed system has been applied for the classical SM transmission principle, as two-way SM scheme. The pairwise error probability of the proposed two-way QSM and SM systems are determined, then the upper bound on the ABEP of the systems derived analytically and through computer simulations. Also, via Monte Carlo simulations, BER performance of two-way QSM and two-way SM schemes have been obtained for different MIMO schemes and spectral efficiency values.