Game Theoretic Channel Selection for Opportunistic Spectrum Access with Unknown Prior Information

The issue of distributed channel selection in opportunistic spectrum access is investigated in this paper. We consider a practical scenario where the channel availability statistics and the number of competing secondary users are unknown to the secondary users. Furthermore, there is no information exchange between secondary users. We formulate the problem of distributed channel selection as a static non-cooperative game. Since there is no prior information about the licensed channels and there is no information exchange between secondary users, existing approaches are unfeasible in our proposed game model. We then propose a learning automata based distributed channel selection algorithm, which does not explicitly learn the channel availability statistics and the number of competing secondary users but learns proper actions for secondary users, to solve the proposed channel selection game. The convergence towards Nash equilibrium with respect to the proposed algorithm also has been investigated.