Influence of Noise Uncertainty Source and Model on the SNR Wall of Energy Detection

Recently, we conducted a comprehensive analysis of energy detection (ED) signal-to-noise ratio wall (SNRw) due to noise uncertainty (NU) in cognitive-radio (CR)-based non-cooperative spectrum sensing (nCSS) and cooperative spectrum sensing (CSS) with soft-decision (SD) and hard-decision (HD) fusion under the k-out-of-M rule. It derived the SNRw for a novel NU source and model adopting a truncated Gaussian NU distribution at the CRs and proposed empirical algorithms for SNRw estimation. Based on it, this article conducts another extensive ED study by deriving new closed-form SNRw expressions combining novel and traditional NU sources and models in nCSS and CSS with SD and HD k-out-of-M rule. Besides the conventional test statistic in CSS with SD, it also considers a more general one that, to our best knowledge, was never studied under NU. This new ED computation improves detection performance when CRs are under unequal noise powers and leads to a more conservative (higher) SNRw when CRs are under unequal NU levels in the novel NU source and model combination. Yet, this article maps new and previous derivations for easier comparisons involving any NU source and model combination, more easily highlighting its advantages. Simulations validate the theoretical findings.