Power allocation for cognitive radio networks with statistical QoS provisioning of primary users

In this paper, we investigate the optimal power allocation strategy for underlay-based cognitive radio networks (CRN) with statistical quality-of-service (QoS) protection of primary users (PU). Instead of utilizing commonly used average/peak interference power constraints to protect PU´s transmission, our proposed power allocation strategy will satisfy PU´s statistical QoS requirement characterized by the queue-length bound violation probability. By applying the effective capacity theory, we convert PU´s queue-length bound violation probability constraint to the equivalent maximum sustainable traffic load requirement. Then, we formulate the optimization problem aiming at maximizing the average transmission rate of secondary user (SU) while meeting PU´s statistical QoS requirement as well as SU´s average and peak transmit power constraints. Unfortunately, such a problem is non-convex. By employing the theories of convex hull and probabilistic transmission, we successfully convert the original non-convex problem to an equivalent strictly convex problem and obtain the optimal power allocation strategy of SU through Lagrangian approach. Simulation results are also provided to demonstrate the impact of PU´s statistical QoS requirement on the SU´s maximum achievable transmission rate as well as the superiority of our proposed optimal strategy as compared to the fixed power allocation scheme.