Evolution of Cooperation: A Case with Interference-Aware Cooperative Spectrum Sensing in Cognitive Radio Networks

We consider in this work a group of secondary users with backlogged traffic to transmit in the primary network. To avoid interfering with the primary user, each secondary user must perform interference-aware spectrum sensing before transmission. Unlike conventional sensing techniques, interference-aware spectrum sensing allows a secondary user to adjust its sensing parameters for optimal performance depending on the probability of interfering with the primary user. While interference-aware sensing can achieve better performance for individual users, challenges arise when secondary users collaborate with each other for cooperative spectrum sensing due to their unequal interference probabilities that result in a conflict for setting the optimal sensing parameters. To model this problem, we consider an interference-aware cooperative sensing game and analyze player behaviors under such a game. We find that there is a unique pure Nash equilibrium of the game, but it tends to deviate from the desirable solution of social optimum. We then design a repeated game based on evolutionary game theory to address this problem. Players in the repeated game have the chance to revenge “uncooperative” players in ensuing repetitions for driving the equilibrium to the social optimum. We show through numerical results that the proposed game of evolution does achieve the desirable performance for interference-aware cooperative sensing in dynamic spectrum access.