Abstract
As wireless technology continues to expand, there is a
growing concern about the efficient use of spectrum resources.
Even though a significant portion of the spectrum is allocated
to licensed primary users (PUs), studies indicate that their
actual utilization is often limited to between 5% to 10% [1].
The underutilization of spectrum has given rise to cognitive
radio (CR) technology, which allows secondary users (SUs) to
opportunistically access these underused resources [2].
However, wideband spectrum sensing, the key of CR, is
limited by the need for high-speed analog-to-digital converters
(ADCs), which are costly and power-hungry.
Compressed spectrum sensing (CSS) addresses this
challenge by employing sub-Nyquist rate sampling. The
efficiency of active transmission detection heavily depends on
the quality of spectrum reconstruction. There are various
reconstruction methods in CSS, each with its merits and
drawbacks. Still, existing algorithms have not tapped into the
full potential of sub-sampling sequences, and their
performance notably drops in noisy environments [3,4].
The GHz Bandwidth Sensing (GBSense) project1)
introduces an innovative approach for GHz bandwidth
sensing. GBSense incorporates advanced sub-Nyquist
sampling methods and is compatible with low-power devices.
This project also prompted the GBSense Challenge 2021,
which centered on sub-Nyquist reconstruction algorithms,
with four leading algorithms to be presented and evaluated in
this paper.