This paper addresses sparse channels estimation problem for the generalized linear models(GLM)in the orthogonal time frequency space(OTFS)underwater acoustic(UWA)system.OTFS works in the delay-Doppler domain,where tim...This paper addresses sparse channels estimation problem for the generalized linear models(GLM)in the orthogonal time frequency space(OTFS)underwater acoustic(UWA)system.OTFS works in the delay-Doppler domain,where timevarying channels are characterized as delay-Doppler impulse responses.In fact,a typical doubly spread UWA channel is associated with several resolvable paths,which exhibits a structured sparsity in the delayDoppler domain.To leverage the structured sparsity of the doubly spread UWA channel,we develop a structured sparsity-based generalized approximated message passing(GAMP)algorithm for reliable channel estimation in quantized OTFS systems.The proposed algorithm has a lower computational complexity compared to the conventional Bayesian algorithm.In addition,the expectation maximum algorithm is employed to learn the sparsity ratio and the noise variance.Simulation and experimental results show that the proposed algorithm has superior performance and low computational complexity for quantized OTFS systems.展开更多
基金supported by National Natural Science Foundation of China(No.62071383)。
文摘This paper addresses sparse channels estimation problem for the generalized linear models(GLM)in the orthogonal time frequency space(OTFS)underwater acoustic(UWA)system.OTFS works in the delay-Doppler domain,where timevarying channels are characterized as delay-Doppler impulse responses.In fact,a typical doubly spread UWA channel is associated with several resolvable paths,which exhibits a structured sparsity in the delayDoppler domain.To leverage the structured sparsity of the doubly spread UWA channel,we develop a structured sparsity-based generalized approximated message passing(GAMP)algorithm for reliable channel estimation in quantized OTFS systems.The proposed algorithm has a lower computational complexity compared to the conventional Bayesian algorithm.In addition,the expectation maximum algorithm is employed to learn the sparsity ratio and the noise variance.Simulation and experimental results show that the proposed algorithm has superior performance and low computational complexity for quantized OTFS systems.
