In wireless communications, interference is generally regarded as an undesired phenomenon. In multiuser systems, interference management and avoidance are essential for acceptable system performance [1, 2]. In systems including cognitive radios with secondary spectrum privileges, a system objective is detecting the channel occupancy in an intelligent way to limit interference to primary users . In contrast with the conventional wisdom, in secure communications, interference can be a potentially beneficial phenomenon and hence a welcome addition, if injected to the system properly. This recently developed interesting form of cooperation enlists the help of nodes that are legitimate entities in the system and essentially asks them to jam the eavesdropper from whom the information flowing in the system is to be kept secret. The idea in essence is to put the eavesdropper at a disadvantage as compared to the legitimate parties. This chapter is devoted to exploring the technique based on this idea which we affectionately term "cooperative jamming" and its applications in different system models with Gaussian channels representing (wireless) communication scenarios of interest. Before going into the details, a note on the naming conventions are in order. Though we will refer to this technique as "cooperative jamming", as it was proposed in , the reader may encounter variants of this technique under names such as "artificial noise" , "noise forwarding" , or "interference assisted secret communication" . All involve introducing interference into the system in one form or another to increase secrecy rate. Indeed, cooperative jamming has become such an essential part of achievability proofs of multi-terminal channel models with secrecy constraints that the readers may encounter it in other chapters of this book. In this chapter, we focus on several interesting aspects of cooperative jamming and illustrate them with examples. In doing so, we must note that we will consider the three forms in which cooperative jamming can be accomplished by the friendly terminal(s): • Cooperative Jamming with Noise, • Cooperative Jamming with a Random Code, • Cooperative Jamming with a Structured Code. The next three sections are devoted to exploiting the above methods to improve achievable secrecy rates in the basic Gaussian Wiretap channel that consists of a transmitter, the legitimate receiver and an external eavesdropper, with the addition of a friendly cooperative jammer. The model can also be thought of as a two-user multiple access channel where one user transmits data while the other does cooperative jamming. Section 4.5 changes the scenario and considers an example where no external eavesdropper is present, but one of the parties involved in communication, in this case the relay node between the two nodes, is not trustworthy. The impact of cooperative jamming by one of the end-nodes on the secrecy rate of the other is considered. Sections 4.6 and 4.7 consider the multiple access channel with an external eavesdropper, where cooperative jamming with noise is established as a method to improve the secrecy sum-rate of the channel. There we show, to maximize the secrecy sum rate, the users need to be divided into two groups: the transmitting nodes and the cooperative jammers, and no user does both, coming to a full circle with the original model we considered in Sect. 4.2.
All Science Journal Classification (ASJC) codes