New Modulation And Coding Technology For Faster Than Nyquist Signaling

A mushrooming number of new applications and demands require higher data rates than that 5G networks can offer.Hence,more and more countries and research organizations have begun to study 6G.Future communications systems will pose higher challenges to spectrum efficiency(SE)and transmission reliability.However,the Nyquist criterion limits the further improvement of the capacity and SE.Faster than Nyquist(FTN)transmission packs the symbol interval in the time domain and packs the subcarrier interval in the frequency domain to improve the SE.The self-interference introduced by time-frequency packing reduces the transmission reliability.FTN employs the coded modulation(CM)to eliminate the self-interference and improve performance.This paper studies CM and its joint optimization for FTN including the multi-carrier FTN(MFTN).The main reason for studying MFTN is that future communication systems with ultra-wide bandwidth need to support parallel and multi-stream data transmission.This paper also studies high-performance equalization algorithms and highly efficient digital implementation for MFTN to verify the CM optimization.The main innovations of this paper include the following aspects:1.For precoder optimization of the CM,this paper proposes an improved bare-bones particle swarm optimization(BB-PSO)algorithm based on maximizing achievable information rate(AIR).The traditional precoder optimization methods take minimum Euclidean distance(MED)as a performance metric.This paper finds that an AIR is a more robust performance metric than the MED,and employs AIR as the objective function.This paper applies the BB-PSO algorithm to the precoder optimization by maximizing the AIR and changes the topology of the BB-PSO algorithm according to the characteristics of the precoder optimization.The optimized precoder achieved 0.5 dB and 0.6 dB precoding gains in low-order and high-order modulation FTN systems,respectively.However,the traditional precoder based on MED metric does not obtain precoding gain and has a 0.1dB precoding loss.2.For high-performance equalization algorithms,this paper proposes the inter-symbol interference(ISI)length optimization based on AIR criteria and an improved Ungerboeck-M-BCJR algorithm based on Channel Shortening(CS).The high-performance equalizer algorithm is important for the turbo-based CM-FTN receiver.The CS method can provide a low complexity and high-reliability auxiliary channel for high-performance equalizer.(1)This paper proposes an ISI length optimization method based on the AIR criterion for the CS method.Compared with the traditional BCJR algorithm and frequency domain equalization algorithm,the optimized CS-BCJR algorithm has gains of 1dB and 0.5dB,respectively.(2)This paper proposes the CS-Ungerboeck-M-BCJR to reduce the complexity of Ungerboeck-M-BCJR algorithm.The CS-Ungerboeck-M-BCJR can reduce the complexity of the receiver by up to 75%while obtaining roughly the same or better BER performance compared to the Ungerboeck-M-BCJR.3.For channel code optimization of CM,this paper proposes an improved BB-PSO algorithm based on extrinsic information transfer(EXIT)chart.The traditional channel code methods take the frame error rate(FER)as the objective function and the genetic algorithm(GA)as the optimization algorithm.This paper proposes an improved BB-PSO algorithm based on EXIT chart to optimize the generator polynomial of convolutional code(CC)and the degree distribution of low density parity check(LDPC)code.Compared with the traditional method based on the FER,the improved method avoids the parity-check-matrix optimization and the FER simulation for the sub-optimal degree.The BB-PSO algorithm has better ergodicity and is easier to implement than the GA algorithm.For high-rate FTN,the optimized CC has a code gain of 3.6dB compared to the reference CC;the optimized LDPC code has a code gain of 3.3dB compared to the reference LDPC code.The optimized LDPC has a code gain of 0.2dB compared to the reference LDPC for high-order FTN system.4.For CM-MFTN,this paper proposes a low-complexity digital implementation and a CS-based Turbo equalization method;this paper proposes a low-complexity linear pre-equalization(LPE)-MFTN(MFTN-LPE)with the effective digital implementation to reduce the complexity in traditional MFTN equalization.This paper proposes an efficient implementation based on inverse fast Fourier transforms(IFFT)/fast Fourier transform(FFT)and filtering,which avoids the following additional operations in traditional MFTN:reorder,rotation,stack,overlap,overlap-add,and circular expanding.This paper proposes a turbo equalization scheme based on CS and parallel interference cancellation(PIC)for the MFTN equalizer.The optimized MFTN has a 0.9dB gain compared to the traditional MFTN.The ultra-wide bandwidth communication systems require high processing speed,and the complexity of MFTN cannot be too high.So,this paper proposes an MFTN-LPE to reduce the complexity.MFTN-LPE has a 0.1 dB performance loss compared to the optimal MFTN,but MFTN-LPE does not require complex equalizer and turbo iteration.The MFTN-LPE with limited performance loss reduces the complexity and is deployed easily in high-speed and high-throughput scenarios.