Quantifying Covertness in the Presence of Primary Networks

Wireless covert systems use low probability of detection (LPD) techniques to hide transmission and protect the transmitter. Traditionally, spread spectrum has been used to hide the covert signal in noise. Recently, research has focused on covert systems that sense the RF channel and hide their transmission in the RF band of a primary network (PN). A widely used metric to quantify covertness is the ratio of the signal-to-interference-noise ratio (SINR) of the covert signal at the intended receiver to that at an intercept receiver (IR). Analyzing this metric is challenging in the presence of PNs and, hence, has been studied in the literature for simple specific cases using simulation only. In this paper, we use stochastic geometry to provide a comprehensive analysis of this metric for three main cases: noise-limited, interference-limited with perfect sensing of the environment, and interference-limited with imperfect sensing. We compute a lower bound on the system's achievable covertness, and demonstrate the advantages of using interference, instead of noise, to increase covertness. The effect of the system parameters on the covert system are discussed. Finally, simulation results confirm the theoretical analysis and the tightness of the proposed bound.