Robust Power Control in Cognitive Radio Networks: A Distributed Way

Conventional distributed power control algorithms in cognitive radio networks are based on the assumption of perfect channel state information (CSI) which may lead to performance degradation in practical systems. In this paper, we investigate the robust distributed power control problem in cognitive radio networks by considering the uncertainty of channel gains. Our objective is to minimize the total power consumption of cognitive transmitters under both QoS constraint at each cognitive receiver and interference constraint at primary receiver. The uncertainty of channel gain is described using ellipsoid sets and the robust power control problem can be formulated as a semi-infinite programming (SIP) problem. It can be transformed to a second order cone programming (SOCP) problem by considering the worst cases of constraints. We apply the dual decomposition theory to solve the robust power control problem in a distributed way. To reduce the overhead of message passing among all cognitive users, an asynchronous iterative algorithm is then proposed and its convergence is also proved. Numerical results show that when there is uncertainty of channel gains, by using the proposed robust algorithms, both the primary user interference constraint and the target SINR requirement of each cognitive receiver can be guaranteed.