On Pilot Sequence Design And Channel Estimation In OTFS System

With the development of high-speed railway(HST),wireless communication with high reliability and high data rate is required to satisfy the communication needs of Vehicle-toInfrastructure(V2I)of mobile networks.The performance of the conventional orthogonal frequency-division multiplexing(OFDM)system is degraded because of the doubly-selective channels.Through a series of 2-demensional(2D)transformations,orthogonal time frequency space(OTFS)system transforms the doubly-selective channels into the time invariant delayDoppler 2D convolution channels,therefore,OTFS can achieve significant performance gains,especially in high mobility scenario.This thesis focuses on pilot sequence design and channel estimation scheme for OTFS system.The thesis studies the feature of the pilot with matched filter scheme for OTFS system and the analysis shows that the optimal pilot sequence should possess ideal local periodic ambiguity function(PAF).According to the feature of the optimal pilot sequence,the thesis analyzes the PAFs of different unimodular sequences.The m-sequence,Zadoff-Chu sequence,Frank sequence and Golomb sequence as pilot sequence are compared with each others,the simulation results show the estimation performance of Zadoff-Chu sequence outperforms others.A novel sequence design algorithm under spectrum hole constraint is proposed which generates sequences with good local AF and low PAPR.At the same time,the multi-stage accelerated iterative sequential optimization(MS-AISO)algorithm to generate a set of unimodular sequences with good local auto-PAFs and cross-PAFs over specific Doppler bins and delay bins of interest is proposed,which can be used as the pilot sequences in MIMOOTFS system.The simulation results show the performance of the proposed sequences is better than the m-sequences and the Zhou sequences in terms of channel estimation error and bit-error rate(BER).In addition,the thesis analyzes the matrix expression of the received pilot signals,in which the channel estimation problem can be converted into the sparse vector recovery problem.Therefore,the orthogonal matching pursuit(OMP)algorithm which is widely used in compressed sensing theory is introduced in channel estimation for OTFS system.Channel parameters such as path gains,delays,Doppler shifts can be estimated by utilizing the nonzero values in the sparse vector and their corresponding locations.The OMP channel estimation scheme can be applied into both SISO-OTFS and MIMO-OTFS systems.The simulation results show that the BER performance of OMP channel estimation scheme is close to the BER performance of ideal channel estimation.