Minimizing the coarse-fine spectrum sensing time of cognitive radios with ideal secondary detectors subject to noise uncertainty: An analytical approach

In previous work a method was derived for generating estimates of design curves for Coarse-Fine spectrum sensing. One of the problems that the method had was it could not account for variations in noise power. If the noise was uncertain or not the same across the bands, then the model could not predict the performance correctly. Here a method is provided to solve this problem, together with a Markov Chain analysis of the system. It will be shown that, by accounting for the variation of noise power in the probability distribution functions, the analysis can be performed as previously, thus reducing simulation time. The Markov model of the process provides a more rigorous analysis of the expected speed gain of the system.