Research On Single-carrier Faster-than-Nyquist Signaling Technologies

Thanks to the rapid growth of semiconductor technology,non-orthogonal transmission technologies that trade computational complexity for spectral efficiency(SE)gain have become a research hotspot in the field of air interface waveform design in recent years,in order to alleviate the contradiction between the explosive growth of data traffic demand and limited spectrum resources.Faster-than-Nyquist(FTN)transmission,as a representative non-orthogonal transmission scheme,uses a symbol interval shorter than that defined by the Nyquist criterion,and then improves baud rate and the SE,at the expense of inter-symbol interference(ISI).This dissertation is concerned with two key issues that are currently confronting the development of single-carrier FTN transmission technology: 1)evaluation of FTN capacity gain in the context of spatial multiplexing; 2)waveform design and low-complexity symbol detection for precoded FTN signaling.The following are the main contents and contributions.1.A discrete-time observation model of the FTN signal transmission over continuoustime multiple-input multiple-output(MIMO)Rayleigh fading channels is established.Moreover,ergodic capacity formulas are proposed,which lay a theoretical foundation for MIMO-FTN transmission.The capacity analysis takes into account the correlations of channel tap coefficients in time,delay and spatial domains,as well as the noise correlation caused by the FTN signaling.The capacity calculation is simplified by adopting the widely used assumption that the spatial correlation matrices of the transmit and receive antennas are in Kronecker product form and by utilizing the mean-value theorem of integrals.The capacity results show that the MIMO-FTN transmission can obtain multiplexing gain and FTN capacity gain at the same time; the FTN gains are nearly consistent in different channel conditions,while the gains obtained in the frequency-selective fading channels are slightly higher than those obtained in the flat fading channels.2.A time-frequency compressed FTN(TFC-FTN)signaling technique is developed to overcome the limitation that classic FTN signaling can only compress the signal in the time dimension.The extra frequency compression is created by precoding.Moreover,an optimization subject to practical constraints,such as spectrum emission mask,is performed on the precoder to suppress the ISI,according to a mean square error criterion.The simulation results show that the ISI level in TFC-FTN is lower than that in the existing TC-FTN.In terms of detection performance,the TFC-FTN scheme also beats the TC-FTN scheme with high-complexity iterative BCJR detection.3.For the FTN transmission in multipath fading channels,linear pre-equalization(LPE)is introduced to cancel the FTN-ISI at the transmitter side.As a result,the linear frequency-domain channel estimation(CE)and subsequently channel equalization are not affected by the FTN-ISI,which benefits for improving the accuracy of CE and the equalization performance.In addition,compared with the traditional FTN CE scheme,the LPE assisted CE scheme not only reduces the complexity,but also decreases the pilot overhead by at least half.4.For the FTN transmission systems that employ high-order quadrature amplitude modulation(QAM)formats and are corrupted by phase noise(PHN),FTN-ISI and PHN are cancelled separately in order to implement low complexity symbol detection.The transmitter,in particular,adopts Tomlinson-Harashima precoding(THP)to eliminate the FTN-ISI,while the receiver uses pilot-assisted phase noise compensation.Following that,two pilot design schemes based on pilot energy reduction(ER)and phase rotation are proposed to address the issue that THP will damage the traditional QPSK symbol-based pilot design.Furthermore,a pulse shape optimization method is proposed to improve detection performance and boost the FTN CRs that can be attained.The simulation results demonstrate that the 4096-QAM FTN scheme with a compression ratio(CR)of 7/6 improves the detection signal-to-noise ratio by around 6 d B when compared to the same-SE 16384-QAM Nyquist scheme,and the increased complexity is almost negligible.5.For the MIMO-FTN transmission systems deteriorated by PHN,THP is applied to equalize the FTN-ISI at the transmitter as in the SISO case,and the ER-based pilot design is extended to multi-antenna systems.Then,two low-complexity PHN and inter-antenna interference elimination algorithms based on orthogonal pilot sequence and extended Kalman filter,respectively,are presented.Furthermore,by embedding dummy symbols in the pilot sequence,the transmission can operate at larger CRs or achieve a better compromise between pilot overhead and detection performance for a given CR.The single-carrier FTN capacity theory is extended for the case of spatial multiplexing.The precoding-based TFC-FTN waveform design,as well as the decoupling method of FTN-ISI and link interference,increase the achievable FTN CR and SE gain,and effectively reduce the complexity of FTN symbol detection.To summarize,the research presented in this dissertation lays a theoretical foundation for promoting the application of FTN transmission technology in high-capacity and high-SE communication scenarios.