Performance improvement with predictive channel selection for cognitive radios

Prediction of future availability times of different channels based on history information helps a cognitive radio (CR) to select the best channels for control and data transmission. Different prediction rules apply to periodic and stochastic ON-OFF patterns. A CR can learn the patterns in different channels over time. We propose a simple classification and learning method to detect the pattern type and to gather the needed information for intelligent channel selection. Matlab simulations show that the proposed method outperforms opportunistic random channel selection both with stochastic and periodic channel patterns. The amount of channel switches needed over time reduces up to 55%, which reduces also the delay and increases the throughput.