Channel State Prediction in Cognitive Radio

Abstract

Spectrum sensing is the cornerstone of cognitive radio, which detects the availability of a spectrum band for the current time. In theory, the result of spectrum sensing reflects the current channel state, which is the ideal case. However, according to the author’s measurements, hardware platforms can introduce a non-negligible time delay on the signal path, which undermines the accuracy of spectrum sensing. To reduce the negative impact of the hardware platform time delay, channel state prediction in cognitive radio is proposed and presented in this chapter. As examples, channel state prediction algorithms based on a modified hidden Markov model (HMM) are given and tested using recorded real-world data. Moreover, as a second stage, cooperative channel state prediction is also proposed and experimentally evaluated. The experimental results approve that channel state prediction in cognitive radio indeed helps improve the accuracy of spectrum sensing in practical cases.