Research On Waveform Technology Of Next Generation Wireless Communication System

ITU-R defines eMBB,u RLLC and mMTC as three main application scenarios of 5G and these scenarios have different demands obviously on the key characteristics of the system.Therefore,5G needs a new air interface to meet these requirements.Due to its inherent problems caused by the rectangular window function,CP-OFDM can't meet requirements of 5G new air interface.So 5G needs new waveform technologies.Nowadays,FBMC,GFDM,UFMC and F-OFDM have been widely recognized as the candidate waveform for 5G.In this thesis,the four waveform technologies are summarized on the basis of analyzing the waveform theory.Besides,this thesis mainly focuses on the research and analysis of FBMC and F-OFDM.In FBMC,each subcarrier is filtered by the prototype filter with good time-frequency localization characteristics,thus providing excellent performances: anti-ISI,anti-ICI,low out-of-band leaks,and efficient spectrum without CP.FBMC with OQAM modulation(FBMC/OQAM)is widely concerned because it can keep orthogonal in the real number domain.This thesis deduces the principle of the FBMC/OQAM system and two fast implementation structures are studied: frequency spreading-extended FFT(FS-FFT)and polyphase network-FFT(PPN-FFT).The implementation complexity of FBMC/OQAM system is simulated and analyzed compared with CP-OFDM system.Because the prototype filter plays a decisive role in FBMC/OQAM system,the design methods and optimization criterions of the prototype filter are studied.In addition,the time-frequency localization characteristics of the commonly used prototype filter are also analyzed and simulated.Based on the analysis of the characteristics of wireless channel,the expression of received signal after multipath channel in FBMC/OQAM system are derived and the channel estimation model of FBMC/OQAM system is given.The POP,IAM and ICM algorithms have been proposed as the preamble-based channel estimation methods of FBMC/OQAM system.Based on the IAM and ICM algorithms,this thesis proposes two channel estimation algorithms with efficient spectrum,which are called ES-IAM-R and ES-IAM-C.At the receiver,channel estimation performance of these two proposed algorithms is close to IAM-R and IAM-C after eliminating the real interference generated by the imaginary pilot,while the pilot occupies only two symbol periods.The main difference between F-OFDM and CP-OFDM is that F-OFDM filters the subband occupied by the user.In this thesis,an F-OFDM system with two asynchronous subbands is designed and compared with CP-OFDM system.The simulation results show that F-OFDM has better sub-band asynchronous capability.Besides,this thesis analyzes the design and implementation method of subband filter for F-OFDM.On the baisi of fast convolution algorithm(Overlap-Save method),a new fast convolution implementation method based on absolute value of the frequency response of the subband filter is proposed in the thesis.Then it has been implemented on FPGA,where a complex multiplier is converted into two real multipliers.FPGA implementation results show that: the new implementation method can save 4099 clock cycles through the rational design of intermediate data input and output order.