Low-Complexity Iterative Equalization And Decoding In Faster-Than-Nyquist(FTN) System

With the advent of the era of big data,massive data transmission and increasingly tight spectrum resources put forward higher requirements for communication networks.The traditional Nyquist transmission theory points out the maximum spectral efficiency that can be achieved without inter-symbol interference.The FTN transmission theory proposed by Mazo breaks this limitation and further improves the spectral efficiency,but inter-symbol interference is introduced and the detection performance of the system is affected.Considering that the equalization algorithm in FTN system which has become a research hotspot is often high-complexity,designing a FTN receiver with the balance of performance and complexity is the subject to be studied in this thesis.In this thesis,a FTN receiver with high detection performance and low complexity is designed by using the idea of parallel interference cancellation,information interleaving technology and the combination of iterative equalization structure and channel decoder.The main research contents of this thesis include:Firstly,we study and compare the Nyquist and FTN transmission signal,and analyse the inherent ISI characteristics in FTN transmission and the Mazo limit theory of FTN system.In the single-carrier FTN system,the performance of ZF equalization and MMSE detection,as well as the optimal MLSE estimation and MAP estimation is simulated and analysed.After that,the convolutional codec algorithm based on neural network is studied and the latest LDPC codec scheme in 5G is simulated.We apply the two codec schemes in FTN transmission system.Finally,in this thesis,we design a high-performance and low-complexity FTN iterative detection scheme.The parallel interference cancellation technology is used to detect the single carrier FTN signal,and the detection performance of the algorithm is simulated and analyzed.Then the PIC algorithm is improved by introducing confidence estimation method.The simulation results show that the improved algorithm has some performance improvement compared with the original algorithm,but the detection performance is still not up to the optimal algorithm performance.To solve this problem,this thesis improves the iterative feedback information in PIC algorithm and introduces interframe information interleaving to further improve the detection performance of the algorithm.Simulation results show that the improved PIC-II algorithm can effectively demodulate while the original PIC algorithm cannot demodulate when the modulation order is 4QAM and the compression radio is 0.4 or even 0.35.In addition,it can achieve the detection performance of the M-BCJR algorithm under specific compression ratio,and its complexity is only one tenth of the M-BCJR algorithm.