Online Spectrum Sensing in Cognitive Radio

Spectrum sensing is an essential part for cognitive radio which is used for directly or indirectly estimating the states of primary users (PUs) to find spectrum holes for second users (SUs). Nowadays, some online spectrum sensing algorithms have been proposed but most of them ignore the time complexity which influences the efficiency of spectrum sensing. In this paper we propose a new filter algorithm named Mixed filter to address this problem. Leveraged by the Hidden Markov model, this algorithm provides an alternative way to estimate parameters of the model online. Extensive simulations have been conducted and the simulation results show that the proposed method can reduce the computational complexity and shorten the executing time withoug out reducing the accuracy.