Dual-Channel Unlimited Sampling for Bandpass Signals

Abstract

Bandpass signals are fundamental to a number of applications. Such signal classes are characterized by high-frequency content. Thus, their digital acquisition is demanding due to the higher sampling rate requirement. Relaxing this requirement has been widely investigated in the literature in the context of bandpass sampling theory. However, such investigations do not address the problem of handling high dynamic range (HDR) inputs, which is inherent to real-life applications and the restrictions imposed by analog-to-digital converters (ADCs). Recent work has shown that one can acquire arbitrarily large bandpass signals at sub-Nyquist rates by leveraging the Unlimited Sensing Framework (USF). Taking a step further, in this paper, a dual-channel architecture that integrates USF and second-order bandpass sampling (SO-BPS) is presented. The result extends the valid sampling frequencies and enhances design flexibility. Furthermore, we show that the sampling rate of one channel can be further reduced, and provide a complementary recovery algorithm. Lastly, we validate the theory, asses algorithmic performance and demonstrate the effectiveness of the proposed solution for HDR bandpass sampling.