Analysis of optimal threshold selection for spectrum sensing in a cognitive radio network: an energy detection approach

The spectrum sensing is a key process of the cognitive radio technology in which the cognitive users identify the unutilized/underutilized primary users (PUs)/licensed users spectrum for its efficient utilization. The sensing performance of cognitive radio (CR) is generally measured in terms of false-alarm probability (Pf\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ P_{f} $$\end{document}) and detection probability (Pd\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ P_{d} $$\end{document}). IEEE 802.22 wireless regional area network is one of the typical cognitive radio standards to access unused licensed frequencies of TV band and according to this standard, the false-alarm probability of CR should be ≤ 0.1 and the detection probability must be ≥ 0.9. Further, the detection and false-alarm probabilities are greatly affected by the selected threshold value in the spectrum sensing approach and selection of threshold is a crucial step to yield the status (presence/absence) of PU. In most of the available literatures, the threshold is decided by fixing one parameter (Pf\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ P_{f} $$\end{document} or Pd\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ P_{d} $$\end{document}) and optimizing the other parameter (Pd\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ P_{d} $$\end{document} or Pf\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ P_{f} $$\end{document}). Moreover, at low SNR, while achieving one of the targeted sensing parameter, there is significant degradation in the other sensing parameter. Therefore, in this paper, we are motivated to decide the optimal threshold at low SNR (signal-to-noise ratio) in such a way where we can jointly achieve both sensing matrices (Pf\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ P_{f} $$\end{document} ≤ 0.1 and Pd≥0.9\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ P_{d} \ge 0.9 $$\end{document}) and provided better sensing performance in comparison to that of the traditional constant false-alarm rate and constant detection rate (CDR) threshold selection approaches. Further, we have illustrated that at low SNR, the proposed optimal threshold selection approach has provided better throughput as compare to that of the threshold selected by traditional CDR approach. The proposed approach has improved throughput approximately 24.63% when compared with CDR at chosen SNR.