An Improved Recovery Algorithm Based on ISD for Multiband Signals

Abstract

Multiband signal is an important signal in radar and communication systems, which is generally sparse and with a wide spectrum. Due to the wide spectral range of multiband signals, traditional ADC sampling at Nyquist rate may be too difficult to achieve. Modulated wideband converter (MWC), which is based on compressed sensing (CS), is a subtle scheme to solve wide spectrum sampling problem. In addition, reconstruction algorithm is the key of MWC which can decide the performance of the architecture. In this paper, we propose an improved reconstruction method, which expands iterative support detection (ISD), an iterative recovery algorithm for one- dimensional signal, to continue to finite (CTF) system. Our algorithm uses ISD to compute the solutions which satisfy that every column's reconstructions is l1 minimization. Then, utilizing the resulting matrix and support detection rule, we can find the support set and obtain the recovered signal. Our algorithm does not require the number of support set as prior knowledge. And the property of our method is close to orthogonal matching pursuit multiple measurement vectors (OMPMMV), and better than Sparsity adaptive matching pursuit multiple measurement vectors (SAMPMMV). Simulation results confirm the conclusion.