Application of neural network for sensing primary radio signals in a cognitive radio environment

The use of radio spectrum is currently regulated by a fixed spectrum allocation policy. This has led to an under-utilization of licensed spectrum by licensed users, while other users are experiencing spectrum shortages. To correct this situation, dynamic spectrum access (DSA) is emerging as a promising technology to enable unlicensed users to operate in unused licensed spectrum. To enable DSA, the unlicensed user must posses the capability of detecting the presence/absence of licensed users so that all users, licensed and unlicensed, can coexist without interference. In this paper, different sensing methods for detecting primary users´ signals are reviewed before proposing automatic modulation classification detection´s method for sensing the presence/absence of licensed users. The research presented in this paper focuses on the sensing of digitally modulated primary radio signals. In achieving this objective, a digital automatic modulation classifier was developed using an artificial neural network. The classifier results show accurate performance with an average success rate of above 99.50%. With the success recorded so far, the outcomes of this on-going research will produce a single sensing method capable of sensing all forms of primary radio signals in a cognitive radio environment. Compared with other detection methods, this sensing method promises better performance as all radio devices in the cognitive radio environment make use of one form of modulation technique or another when transmitting their signals.