A Soft-Calibration System of Wideband Compressed Sampling Receivers based on MUSIC Subspaces Method

The emergence of compressed sensing (CS) theory provides potential hardware architecture to sub-Nyquist sample the wideband signals. However, applying this discrete CS model to continuous analogue signals is not an easy task. The modulated wideband converter (MWC) is an efficient wideband compressed sampling architecture for the sparse multiband signals. In this paper, a soft-calibration system is proposed to calibrate the sensing matrix with the sub-Nyquist sampling data. It utilizes a calibration branch with known signal source to estimate the actual measurement time-delay and calibrate the sensing matrix in measurement branch. The proposed soft-calibration algorithm based on MUSIC subspaces method greatly reduces the hardware requirement and computational complexity of sensing matrix calibration. Keywords—Compressed sampling (CS), modulated wideband converter (MWC), multiple measurement vector (MMV), sensing matrix calibration, MUultiple SIgnal Classification (MUSIC).