Distributed rate allocation in MISO cognitive networks

We consider multiple-input-single-output (MISO) cognitive radio networks that have been conferred concurrent spectrum access with the primary (licensed) users. We treat the problem of distributed rate allocation in the cognitive network such that some notions of optimality and fairness are sustained and the interference imposed on the primary users is tightly controlled. Allowing multi-user decoding at each cognitive (secondary) receiver, we provide an explicit formulation for the rate allocation problem which seeks to maximize the minimum assigned secondary user rate. Unfortunately, the resulting optimization problem is non-convex and hence cannot be solved efficiently even with centralized processing. As a remedy, we propose a distributed two-stage suboptimal approach. In the first stage, assuming that each secondary user employs singleuser decoding, we design optimal secondary beamformers in a distributed manner, such the primary interference margin constraints and a secondary power budget constraint are satisfied and the minimum secondary user rate is maximized. In the second step, we let the secondary users employ multi-user decoding, which allows them to support higher rates beyond those achieved via single-user decoding. We then offer optimal distributed and low-complexity algorithms for fairly allocating these potential excess rates among the secondary users.