Random projection method for ultra wideband signal acquisition

Abstract

High-speed high-resolution analog-to-digital (A/D) conversion demanded by ultra wideband (UWB) signal processing is a very challenging problem. This paper proposes a parallel random projection method for UWB signal acquisition. The proposed method can achieve high sampling rate, high resolution and technical feasibility of hardware implementation. In the proposed method, an analog UWB signal is projected over a set of random sign functions. Then the low-rate high-resolution analog-to-digital convertors (ADCs) are used to sample the projection coefficients. The signal can be reconstructed by simple linear calculation with the sampling matrix, without complying with optimization algorithm and prior knowledge. In other aspects, unlike other approaches that need to utilize an accurate time-shift at extremely high frequency, or design a hybrid filter bank, or generate specific basis functions or work for signals with prior knowledge, the proposed method is a universal sampling approach and easy to apply. The simulation results of signal to noise ratio (SNR) and spurious-free dynamic range (SFDR) validate the efficiency of the proposed method for UWB signal acquisition.