Stochastic Multichannel Sensing for Cognitive Radio Systems: Optimal Channel Selection for Sensing with Interference Constraints

This paper considers the problem of sensing and transmission strategy of multiple parallel channels owned by the primary user, referred as stochastic multichannel sensing. The traffic parameters follow the Markovian traffic assumption and are not identically distributed among the channels. In order to obtain the optimal probabilities of channel selection for sensing, we formulate a maximization problem for the secondary user throughput with interference constraints to the primary user. The solution to the problem is obtained via linear programming. Numerical results show that the proposed stochastic sensing achieves higher normalized effective throughput and lower average collision probability than the conventional deterministic sensing in a non-identical traffic environment. Additionally, the proposed method greatly reduces computational overheads and memory space.