Unified Framework for Secrecy Characteristics with Mixture of Gaussian (MoG) Distribution

The mixture of Gaussian (MoG) distribution was proposed to model the wireless channels by implementing the completely unsupervised expectation-maximization (EM) learning algorithm. With the high convenience for density estimation applications, the focus of this letter is supposed to investigate the secrecy metrics, including secrecy outage probability (SOP), the lower bound of SOP, the probability of non-zero secrecy capacity (PNZ), and the average secrecy capacity (ASC) from the information-theoretic perspective. The above-mentioned metrics are derived with simple and unified closed-form expressions. The effectiveness of our obtained analytical expressions are successfully examined and compared with Monte-Carlo simulations. One can conclude that this letter provides a simple but effective closed-form secrecy analysis solution exploiting the MoG distribution.