Optimal Channel Sensing in Wireless Communication Networks with Cognitive Radio

In this paper, designing channel sensing policies for cognitive radio networks is discussed. A discrete-time semi- Markov channel model is first introduced, which facilitates the analysis on more general channel occupancy behaviors and possible asynchronism among primary and secondary users. Based on the characteristics of the channels at the stationary state, multiple channel sensing policies have been proposed for different network scenarios with homogeneous channels, heterogeneous channels, and sequential channel sensing. Both analytical and simulation results are given to demonstrate the effectiveness of the proposed channel sensing policies.

I. Introduction

Recently, spectrum utilization begins to attract attention from researchers. On one hand, unallocated frequency bands become scarce with the increment of wireless applications, while on the other hand, most allocated spectrum is severely under-utilized [1]. In order to improve the spectrum utilization, a new spectrum allocation method, called cognitive radio, is proposed [2]. With cognitive radio, users are classified into two groups, i.e., primary users (PUs) and secondary users (SUs). The PUs are licensed users for a given frequency band and have highest priority to access the allocated band, while the SUs opportunistically share the bandwidth with the PUs only when the bandwidth is not currently used by PUs. Therefore, in order to avoid severe interference to the transmission from PUs, the SUs need to sense the channel availability first and then carry out data transmission over idle channels.