Control and estimation through cognitive radio with distributed and dynamic spectral activity

Cognitive radio systems are currently a popular topic and are a good prospect for research when combined with control technology. In this paper, we consider the control and estimation via a two switch model which can represent a cognitive radio system. We provide an optimal estimator for this model and demonstrate that the optimal LQG controller is not a linear gain of the estimate of the states, as well as showing how the separation principle does not hold. Several conditions of stability are also discussed. Numerical examples are provided to verify the results.