Q-learning Based p-pesistent CSMA MAC Protcol for Secondary User of Cognitive Radio Networks

We consider a reinforcement learning based p-persistent CSMA protocol for cognitive radio secondary user when the primary user operates with the conventional CSMA/CA scheme. The learning scheme is applied to adjust the value of transmission opportunity p to reduce the idle time of channel and also avoid excessive collisions compared with fixed p scheme. Then the proposed method can keep the throughput of the primary user and the channel utilization rate can be improved by sharing the primary user and secondary user traffic. The simulation results show that the secondary user can utilize the available channel efficiently at the cost of additional delay of primary user.