Jointly Optimal Sensing Selection and Power Allocation for Cognitive Communications

In this paper we develop the jointly optimal sensing-channel selection and power control scheme for cognitive communications over time-varying fading wireless channels. With judicious formulation, it is shown that this joint design can become a convex optimization problem. Optimal scheme is then derived in closed-form under the constraints of average power budget and maximum allowable probability of collisions with the primary communications. A stochastic optimization algorithm is further developed to dynamically learn the statistics of the wireless channels on-the-fly to approach the optimal policy. Simulation results demonstrate the merit of the proposed scheme.