Incentive for cooperation in Cognitive Radio Networks with group communities

In Cognitive Radio Networks medium access becomes more complex due to existence of competitive nodes with different detection and access strategies. Each node in a Cognitive Radio Network can try to utilize all or some of the channels that it detects as unused. Since the available spectrum can be shared selfishly or fairly among the candidate nodes, sharing the spectrum can be seen as a Prisoner´s Dilemma Game. In a simple N times repeated Prisoner´s Dilemma Game, it is proven that any rational player tends not to cooperate at all. Non-cooperative nodes try to consume all available channels for themselves, which in turn causes more collision and less effective utilization of the available channels. Punishment methods like Tit-For-Tat and Generous Tit-For-Tat are proposed to enforce cooperation according to the nodes´ previous actions. However, such harsh punishment algorithms that are based solely on previous actions result in lower utilization when the probability of misbehavior or false positive detection of a cooperative action is high as in the case of cognitive radio networks. This paper proposes a novel protocol (CR-COOP) that provides new incentives for cooperation, forgiveness and punishment based on group communities where certain reputation is needed for joining a community. Through the simulations, the channel utilization of each method is compared and CR-COOP is shown to be more resilient to increased false positive detections. Moreover, smarter selfish nodes that misbehave by exploiting the reputation logic are introduced to the system, and it is shown that even under these circumstances the channel utilization of the CR-COOP is remarkably better than the existing solutions in the literature.