An efficient spectrum sensing over η-μ fading on sub 6 GHz bands: A real-time implementation on USRP RIO

Abstract

Massive machine-type communications (mMTC) face spectrum scarcity problems at 5G mid-bands (i.e., sub 6 GHz) due to tens of billions of machine-type devices connected in a smaller region. The mMTC facing significant challenges are scalability and efficient connectivity. To enable cognitive radio (CR) in 5G mMTC technology to resolve issues and gives promising solutions. Dynamic spectrum assignment (DSA) is mainly used to solve spectrum scarcity issues that provide efficient connectivity to massive devices. Spectrum sensing (SS) is an integral part of the CR used to detect spectrum opportunities in the spectrum. Energy detection (ED) is a prominent sensing mechanism, and a threshold computation is essential to find accurate spectrum opportunities in the spectrum. The massive devices in mMTC communication send small messages and face short-term fluctuations during sensing at the sub 6 GHz bands. Hence, threshold calculation is a critical challenge in SS at sub 6 GHz bands that depend on two key factors: signal and noise variance. At 5G mid bands, the high random nature of the signal input gives poor accuracy to find spectrum holes under the assumption that the channel is either static or quasi-static in a non-line-of-sight (NLOS) scenario. Thus, a novel spectrum sensing with an efficient threshold over $\eta -\mu$ fading channel is proposed. The key implementation is the moving average energy (MAE) mechanism to minimize short-term fluctuations in signal energy computation and delay variance that reduces sensing delay under fading conditions. The noise variance is measured from the upper bound limit of the noise interval. The SS mechanism implemented on the real-time testbed and GNU radio processing blocks on 3.3–3.8 GHz frequency bands using universal software radio peripheral (USRP)-2953 RIO. The experimental outcomes show a high probability detection rate and low sensing delay at 5G mid-bands.