Joint Time-Frequency Transform Domain Signal Processing And Underdetermined Receiver In Wireless Communication Channels

With Artifical intelligence(AI),the new technological revolution is taking place.Wireless communications,which provides convenient,reliable and high-speed communications for people and smart devices,supports the new technologies and new applications.In the revolution of mobile communications technologies,the mobile communications systems before 5G focus on the commu-nications between people,and they endeavour to provide wider bandwidth.5G,however,aims to provide ubiquitous connectivity for both human and machines,and to support various applications with different requirements.Massive machine-type communications(mMTC)and ultra-reliable low-lantency communications(URLLC)bring more requirements and challenges.As the basis of wireless communications,physical layer adjusts the signal structure,the modulation and demodu-lation techniques according to the requirements of communications,and compensates the channel fading to provide efficient and reliable wireless connections.We focus on the time-variant channel equalization and signal processing in wireless communications,and study the time-varaint signal processing of two typical circumstances,the high-mobility communications and the non-orthogonal communications.The high-mobility communications mainly occur in vehicle communications,for example,in high-speed train.One of the key performance indicators of 5G is mobility,that is,it will support the broadband wireless communications under speed as high as 500Km/h.The relative motion between transmitter and receiver results in rapid change of the channel parame-ters.The combination of multipath propogation and Doppler effect brings about delay spread and Doppler spread to the transmitted signal,and resulting in fast-varying intersymbol interferences in the received signal.The interference in non-orthogonal transmissions,however,is deliberate Compared with the interference caused by the channel,there is more a priori information about the interference in non-orthogonal transmissions.We estiblish time-varaint signal model accord-ing to the characteristics of the signal transmission process,and design various time-variant signal processing techniques to mitigate channel fading and interferences.The main contributions of this dissertation are as follows1.Transform Domain Multiplexing and Transform Domain Equalization for doubly selective channels.When there exists high-speed relative motion between the transmitter and the receiver,the sig-nal often experiences both time and frequency selective fading during transmission,and the signal experiences delay spread in time domain and Doppler spread in frequency domain.It results in se-vere intersymbol interference(ISI)and intercarrier interference(ICI).We design the modulation and demodulation of the time-frequency signal,which converts the transmission of the time-frequency signal over the doubly selective to a two-dimensionl(2D)convolution of the transmitted signal ma-trix and the channel scattering coefficience matrix.By adding cyclic prefix(CP)to both time and frequency domain,the 2D linear convolution is converted to 2D cyclic convolution.Because the 2D cyclic convolution is equivalent to dot multiplication in its 2D Fourier transform domain,we propose Transform Domain Multiplexing(TrDM)and Transform Domain Equalization(TrDE)to mitigate the doubly selective fading.Simulation results show that both techniques can mitigate the doubly selective fading with very low comlexity.They have good performance in doubly selective channels and are able to realize reliable communication with high-mobility2.Cooperative multipath-Doppler diversity in vehicle mobile relay systemsThe wireless communications on high-speed train suffers from severe signal attenuation,fast-varying channel and frequent handover between cells.They degrade the quality of service and the reliability of the communications.Relay is an effective solution for high-mobility communications,and it also provides potential space diversity.In addition,delay spread and Doppler spread not only brings about ISI and ICI,but they also provide degrees of freedom of the channel which imply potential diversity.By designing the transmitted signal structure,the forwarding mode of the relay and the time-varaint signal processing at the destination,the vehicle mobile relay system can not only mitigate the channel fading,but can also obtain time,frequency and space diversity.Both theoretic analysis and simulation results verify that our proposed scheme is able to improve the quality and reliability of the wireless communications on high-speed train3.Data recovery in sub-Nyquist samplingSampling with at least Nyquist rate is sufficient for sampling without information loss,but not necessary.Sub-Nyquist sampling can also sample the sparse signals without information loss.But in this dissertation,we study the sub-Nyquist sampling of the baseband signal which has no sparsity in both time and frequency domain.The finite alphabet of the transmitted signal contraints the data recovery process.When the baseband signal is sampled with sampling rate less than the transmitted symbol rate,the number of samples is less than the number of transmitted symbols,and the sam-pling is modeled with an under-determinded time-varaint linear system,with transmitted symbols as inputs and samples as outputs.The equivalent channel that includes the sampling process is a special time-varaint channel,which has less outputs than inputs,and loses part of the spectrum in-formation.We propose time-variant Viterbi Algorithm(TVVA)to recover the transmitted symbols from samples.Through theoretic analysis,we get the upper bound of the performance of the data recovery.Simulation results verify the effectiveness of the data recovery algorithm.