Abstract:

In cognitive radio (CR) networks with slotted channels, a brandnew definition，channel availability vector (CAV), which characterizes the state information of a group of channels, is proposed based on spectralcorrelation prediction in order that the secondary users can find the available channel as rapidly as possible. The problem of sensing time minimization is formulated as a Markov decision process (MDP), and the best sensing strategy for current CAV is figured out. Through real measurement data for spectrum availability that are acquired from Guangdong province, it is verified that, as a satisfactory solution for lowrate services in highcongested spectrum, the predictionbased scheme outperforms the random one and the sequential one to minimize sensing time for secondary users.

Key words: cognitive radio, spectrum sensing, channel availability vector, Markov decision process