Turning foes to allies in cognitive radio networks

We study a class of problems in Cognitive Radio Networks where multiple half-duplex unlicensed (secondary) users can eavesdrop and jam the communications of licensed (primary) users unless granted access to communicate over the same spectrum band. The problem is to characterize the optimal rule for the primary system that grants spectrum access to selected secondary users and the optimal resource allocation for the secondary users. We model the problem as a Stackelberg game with the primary system as the leader. The equilibrium analysis shows that it is not always optimal to grant access to the strongest eavesdroppers. In addition, it is shown that transmitting secondary users can limit the eavesdropping capabilities of other secondary users, possibly leading to improved primary secure transmission data rate. Thus, interestingly, the outcome reveals a recruiting process that turns selected eavesdroppers into helping jammers under certain conditions. Finally, we propose a low complexity algorithm to select a subset of secondary users for transmission and evaluate the performance of the primary system when different number of secondary users are granted access through simulations.