Research On Signal Detection And Channel Estimation In OTFS

Orthogonal Time Frequency Space(OTFS)is a novel modulation technology derived from the Orthogonal Frequency Division Multiplexing(OFDM).The main difference is to use the twodimensional Fourier transform in the time-frequency domain to construct a new delay-Doppler domain,the signal in the time-frequency domain of OFDM system is modulated in the delayDoppler domain to counter the orthogonality damage of OFDM sub-carriers caused by high speed movement and high frequency band.OTFS modulation technology has significant performance advantages in high-speed Mobile scenarios,which can meet the requirements of the Sixth Generation Mobile Communication System(6G)in the tera-hertz frequency band and 1000km/h terminal mobile speed application scenarios.It is a key technology in the field of communication in the future,and has been widely concerned by the academic community.In this thesis,the channel estimation and signal detection of OTFS receiver are studied.This paper first introduces the OTFS related background,frequency selective fading and time selective fading channels.Then the architecture and basic principle of OTFS transmitter and receiver are introduced,and the concrete expression of OTFS delay-Doppler domain input/output relationship is obtained.Secondly,aiming at reducing the high complexity of Unitary Approximate Message Passing(UAMP)algorithm in the OTFS receiver under the practical rectangular waveform,a low complexity UAMP-based signal algorithm is proposed,through the equivalent channel matrix local loop reconstruction,the input variable of UAMP algorithm is reconfigured,and the calculation is simplified based on the cyclic feature,so the complexity is reduced.Then the low complexity UAMP-based signal algorithm is combined with Turbo iteration to convert the iteration of algorithm into the soft information interaction between equalizer and decoder,so as to improve the performance.This thesis also extends the low complexity UAMP-based signal algorithm under Signal Input Signal Output(SISO)to Multiple Input Multiple Output(MIMO),which increases the transmission capacity of the system.The simulation results show that the low complexity signal algorithm based on UAMP algorithm and Turbo iteration has excellent performance.Finally,the channel estimation technique in OTFS system is studied in this thesis.First of all,channel estimation algorithms based on compressed sensing are described.The Orthogonal Matching Pursuit(OMP)algorithm,suffers performance degradation due to continuous atom indexing errors under the low Signal-to-Noise Ratio(SNR).The Residual Network(Res Net)in deep learning is used to replace the atomic selection process in OMP algorithm,and the serial iteration of OMP algorithm is transformed into parallel Res Net network selection.Thus,the OMP algorithm can optimize the atomic index error deviation in subsequent iterations due to the atomic index error in the last iteration under the low SNR.In addition,this thesis also uses the Field Programmable Gate Array(FPGA)platform to design and implement the transceiver module of OTFS system,and verifies the excellent performance of OTFS modulation technology in high-speed mobile application scenarios through practical engineering tests.