Research On Key Technologies Of The Faster Than Nyquist Transmission

Faster than Nyquist(FTN)signaling can achieve a higher transmission rate than the Nyquist rate in a certain range without consuming additional spectrum resources by artificially introducing inter-symbol interference(ISI).As an effective method to improve capacity,FTN transmission has a very broad application prospect under the general trend of future integrated satellite-terrestrial communications.At the same time,FTN technology faces many theoretical and practical problems and challenges in its application,which means that further research work is required.In this paper,we try to address the key technologies of FTN signaling,including capacity analysis,receiver-related technologies,transmission system design,security and technology expansion,etc.The main contributions are as follows.Firstly,in terms of FTN capacity analysis,most of the current work focuses on the derivation of upper bounds on capacity.Relatively little research has been done on numerical computation.In this paper,we try to solve the problem of numerical calculation of FTN capacity and derive the numerical expression of the spectral efficiency of FTN signaling under additive white Gaussian noise(AWGN),Rayleigh and Nakagami-m channels.The expressions are in the form of a summation of items,which is convenient for practical calculations.The computational accuracy achieves a fast convergence with the number of items.Finally,the proposed capacity expressions can also be used to compose and analyze the subsequently proposed VPR-H and VPR-S high-speed secure transmission systems.Secondly,in terms of receiver technology for FTN transmission,this paper focuses on FTN signal detection and packing ratio recognition problems.For signal detection,dedicated to lower complexity and higher performance,this paper proposes FTN-DNet,a deep learningbased FTN signal detection algorithm,and FTN-JNet,a joint FTN signal detection and decoding algorithm.FTN-DNet achieves a near-optimal performance with moderate complexity and good generalization ability for signal-to-noise ratio(SNR).Meanwhile,with the internal successive interference cancellation structure,FTN-DNet can effectively eliminate the distortion in soft information caused by ISI,and then generate the accurate log-likelihood ratio(LLR)for subsequent channel decoding operations.Further,the FTN-JNet algorithm replaces the signal detection and channel decoding parts of the conventional FTN receiver with a deep learning structure,which achieves a near-optimal performance.In addition,given the critical impact of packing ratio on the performance and security of FTN signaling,this paper focuses on the recognition of packing ratio within FTN receivers and proposes two deep-learning-based recognition algorithms,FTN-RC and FTN-RU,for cooperative and noncooperative communication scenarios,respectively.The two algorithms can achieve blind recognition for packing ratio in FTN transmissions with the help of received symbols.The algorithms have low complexity,fast convergence and the ability to generalize to other SNR values.Especially,the FTN-RC algorithm for cooperative communication has significantly lower complexity.Thirdly,in terms of FTN transmission system design,this paper focuses on improving the spectral efficiency and security of communications with FTN technology.With the help of VPR and FTN-RC algorithms,this paper proposes the efficient transmission system VPR-H and the secure transmission system VPR-S.Among them,the transmitter of the VPR-H system can dynamically switch the packing ratio at specific moments,while the receiver uses the recognition algorithm FTN-RC to infer the parameters for subsequent signal detection and other operations.The system can achieve a higher spectral efficiency without consuming any in-band or out-band resources(e.g.,pilot,dedicated channel,etc.)for packing ratio transmission.The VPR-S system ensures the security of the transmission system through dynamic switching moments and dynamic packing ratios.In particular,the dynamic switching moment generation method,which is proposed in the VPR-S system,can effectively prevent detection and attack by eavesdroppers by exploiting the reciprocity,uniqueness,randomness and time-varying nature of the channel information between the transmitter and receiver.Finally,in terms of the extension of FTN technology,a joint system of FTN and non-orthogonal multiple access(NOMA)is considered to combine the advantages of both technologies.This paper focuses on the user pairing and power allocation problems in the FTN-NOMA system and proposes solutions for single-carrier and multi-input-multi-output(MIMO)scenarios.First,a power allocation algorithm for single-carrier FTN-NOMA systems is proposed,which can guarantee the fairness of users where their spectral efficiency under FTN-NOMA systems is not lower than that under Nyquist-OMA systems.Then,the optimal user pairing strategy under this power allocation scheme is proposed.Simulations show that the proposed algorithm can effectively improve the achievable sum rate(ASR)of the system and improve the quality of service for users.Finally,a dynamic user pairing and power allocation algorithm is proposed for the MIMO-FTN-NOMA system.The proposed algorithm shows significant gains in ASR and outage probability compared to the conventional Nyquist-OMA system.