An novel spectrum sensing scheme combined with machine learning

Cognitive radio (CR) network technology is widely used as a approach to the solve the scarce radio spectrum by allowing the unlicensed users to access the licensed spectrum. Since in the CR network, the licensed users are easily be affected by the introduce of the unlicensed user, we have to avoid to bring the interference to the licensed users when the unlicensed users try to transmit the data on the licensed radio spectrum. It is very difficult to solve this problem especially when the licensed users are mobile. The situation becomes even worse when the distribution of the licensed users is unknown. In this paper, in order to improve the throughput under the mobile network, we propose a novel algorithm which combines the random forest to decrease the interference of the unlicensed user to the licensed users, thus, the network throughput can be dramatically improved. The simulation results show that our proposed novel algorithm has good performance in improving the mobile network throughput.