Physical Layer Security in Wireless Networks With Ginibre Point Processes

Abstract

In this paper, we investigate wireless networks consisting of a legitimate transmitter (Alice), a legitimate receiver (Bob), eavesdroppers (Eves), and friendly jammers. Two network scenarios are considered depending on whether Alice and the jammers have the ability to detect the existence of Eves in their vicinity. If they do not have the ability, as a means to enhance the secrecy, Alice transmits artificial noise and each jammer selectively radiates a jamming signal based on the channel gain between the jammer and Bob. On the other hand, when they have the ability, Alice sends a confidential message to Bob if no Eve is detected within its guard zone, and the jammers transmit jamming signals when there exists at least one Eve in their vicinity. We model the spatial distributions of Eves and jammers as $\beta $ -Ginibre point processes, which can characterize repulsion among the nodes and include the Poisson point process (PPP) as a special case. Then, we analyze both the probability that Bob successfully decodes the confidential message and the probability that the message is secure against eavesdropping. Also, we show that our analysis is a generalization of previous works on the networks with PPPs by recovering them from our analytical results.