Opportunistic spectrum access over fading channels with collision constraints: A DEC-POMDP approach

In cognitive radio network, one of the most significant issues is to design an effective framework in which multiple cognitive radio users coexist with each other without causing pernicious collision to primary users. Despite manifold works on this issue, complete solution with considering entire conditions has not been presented yet. By regarding this problem in an opportunistic spectrum access scenario, this paper investigates how cooperative cognitive radio users maximize their common utility functions through finite horizon in distributed manner. The model for transmission of primary user is assumed as a continuous-time on-off Markov process. Meanwhile, cognitive users transmit at constant time slot. Due to hardware limitation, a secondary user cannot sense the entire spectrum. So, we adapt a cooperative stochastic game, namely Decentralize Partially Observable Markov Decision Process, to this framework which guarantees maximization of throughput over fading channel. Furthermore, we propose an iterated algorithm to satisfy the collision constraints for protection of primary users. We alleviate this problem to near optimal solution which can cope for large time horizon. Simulation results indicate that our method outperforms the existing strategy.