Frequency spectrum sensing of cognitive radio based on Bayesian network

Cognitive radio has become a key technology to improve the low utilization of spectrum. Although some traditional detection methods could solve the problem of spectrum utilization, the detection performance is not high enough. This paper presents a cognitive radio spectrum sensing method based on Bayesian network in order to improve its detection performance. The spectrum of the primary user has two states, which are binary detection signals. Bayesian criterion and maximum a posteriori probability criterion used in the Bayesian network are suitable for the binary signals´ detection. Firstly, the posterior probability of each state can be obtained from the Bayesian network structure, then the posterior probability is used as the prior probability for the maximum a posteriori probability criterion, and thereby the spectrum can be sensed. It can be seen from the MATLAB simulations that detection probability is improved with the maximum a posteriori probability detection criterion and its detection performance is higher than that of the traditional methods.