시변 다중경로가 스파스 신호 파라미터 추정에 미치는 영향

Abstract

Shallow underwater acoustic channel is typically characterized as sparse channel and the sparsity has been actively exploited for accurate channel estimation. However, time-varying multipath distortion degrades the performance of sparse approximation, and the amount of the distortion is dependent on the channel condition each multipath component encounters during the transmission. In this research, we study the signal distortion of time-varying multipath channel. Especially, we are interested in the spatial difference of signal distortion which affects the sparse approximation of multichannel receiver data. To this end, we propose a method to decompose multipath signals from received composite signal. The proposed method decomposes each multipath signal using the measured input delay-spread function of selected multipaths, and the decomposed signal is shown to have the same Doppler shift as the physical channel condition. Analysis of experiment data under regular surface gravity wave shows how the spatial coherence varies as the reflection at the boundary multiples and also that the spatial coherence has strong correlation with the Doppler shift variation which is affected by the surface wave. approximation, and the amount of the distortion is dependent on the channel condition each multipath component encounters during the transmission. In this research, we study the signal distortion of time-varying multipath channel. Especially, we are interested in the spatial difference of signal distortion which affects the sparse approximation of multichannel receiver data. To this end, we propose a method to decompose multipath signals from received composite signal. The proposed method decomposes each multipath signal using the measured input delay-spread function of selected multipaths, and the decomposed signal is shown to have the same Doppler shift as the physical channel condition. Analysis of experiment data under regular surface gravity wave shows how the spatial coherence varies as the reflection at the boundary multiples and also that the spatial coherence has strong correlation with the Doppler shift variation which is affected by the surface wave.