Tunnelized Cyclostationary Signal Processing: A Novel Approach to Low-Energy Spectrum Sensing

We present novel tunnelized second- and higher-order cyclostationary signal processing algorithms to simultaneously detect and characterize RF signals. Techniques that exploit second- and higher-order cyclostationary features to detect and classify signals possess many desirable properties. However, their pervasive use and hardware implementation have been hampered because such features are highly complex, and consume substantial processor energy. In this paper we present a novel concept, where we observe that severe but purposeful under-sampling of the signals through tunneling preserves sufficient exploitable cyclostationarity, even when the tunnel bandwidth is much smaller than the signal bandwidth. This phenomenon is then exploited to create a low complexity and flexible suite of algorithms to simultaneously detect and characterize signals using their tunneling-distorted cyclostationary features. We also demonstrate that such algorithms can detect and characterize signals for a highly adverse signal-to-interference-plus-noise ratio, even when multiple signals completely overlap in time and frequency.