Experimental Assessment of Eigenvalue-Based Spectrum Sensing Using USRP-Based MIMO Testbed for Cognitive Radio Applications

Abstract

One of the key tasks of cognitive radio systems is spectrum sensing (SS), which allows secondary user devices to detect the presence or absence of a primary user (PU) in the frequency band. Among the various proposed sensing techniques, the eigenvalue-based SS (ESS) is known for having a high detection rate without requiring any knowledge about the PU or the noise level. Although some experimental studies on ESS techniques have been conducted in the literature, more practical assessments are still needed before industrial integration. Therefore, we propose an experimental testbed for ESS algorithms performance evaluation with up to four receiving (Rx) channels using the USRP N310 software-defined radio. Furthermore, we experimentally demonstrate that the maximum-to-minimum eigenvalue and the mean-to-square extreme eigenvalue techniques have identical performances, even with up to four Rx channels. In addition, our simulations show that proper selection of the ESS technique can enhance the detection probability by 70.8 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\%$</tex-math></inline-formula> with a relatively low number of samples (500) at a low signal-to-noise ratio (−12 dB) when four Rx channels are used.