Multi-relay Cooperative Mechanism with Q-Learning in Cognitive Radio Multimedia Sensor Networks

Multimedia data transmission requires strict QoS in Wireless Multimedia Sensor Networks. However, the limited frequency spectrum and serious interference between users is a challenge in communication system. In this paper, with the introduction of cognitive radio technique in wireless multimedia sensor networks, a Q-learning based multi-relay cooperative mechanism is proposed to guarantee QoS requiements during data transmission. Firstly, according to the results of spectrum sensing of cognitive nodes, a service aware criterion is designed to judge whether a node needs cooperative relays. Secondly, the reward value of Q function is considered as the ratio of the residual energy and communication energy consumption. Then, a satisfaction function based on transmission distance and SNR is proposed. Considering the reward value and the satisfaction, cooperative relays are obtained. Simulation results show that our proposed mechanism can significantly decrease the transmission delay and balance the network load.