On the probability of appearance of primary user in IEEE 802.22 WRAN using an artificial neural network learning technique

IEEE 802.22 wireless regional area networks promotes the use of TV white spaces to provide wireless broadband access to rural area using cognitive radio techniques. While selecting the best unused channel for transmission, a secondary user (SU) can use various parameters, among which the probability of appearance of the primary user (PU) in the channel. If that probability is high, the SU has to vacate the channel quickly when the PU will appear. This increases the hand-off and degrades the throughput. In this paper, the probability of appearance of a PU in a particular channel is predicted using an artificial neural network learning technique that learns the previous behavior of the PU. Simulation results are presented, showing that our proposed approach can be well fitted with the Mean Square Error (MSE) method as low as 0.08.