QoS aware admission and power control for cognitive radio cellular networks

In cognitive radio cellular networks (CogCell), the secondary users (SUs) are allowed to access the channels licensed to the primary users (PUs) including Primary Transmitters (PTs) and Primary Receivers (PRs), only if the interference to the PRs is less than the predefined threshold, and the quality of service (QoS) requirements of PTs are guaranteed. In addition, different SUs may require different levels of QoS, and pay differently depending on the provided QoS. The network operator achieves different secondary revenues by admitting SUs in different QoS levels. The problem we address in this paper is to maximize the total secondary revenue relative to the interference constraints on PRs, and QoS requirements for both PTs and SUs. We formulate this optimization problem, and propose a power control scheme for both PTs and SUs. Then, we introduce three solutions including an exact solution using dynamic programming, a greedy heuristic algorithm, and a minimal signal-interference-plus-noise-ratio (SINR) removal algorithm. Based on these algorithms, we propose three QoS aware admission and power control (QAPC) schemes, one optimal solution called QAPC-dynamic, and two approximate solutions called QAPC-greedy and QAPC-minimal SINR removal algorithm (MSRA), respectively. Numerical results show that QAPC-dynamic always achieves the highest secondary revenue while QAPC-MSRA gives the lowest secondary revenue. Since the time complexity of QAPC-dynamic is much higher than the other two schemes, QAPC-greedy is recommended considering the trade-off between the computation complexity and performance gain. Copyright © 2009 John Wiley & Sons, Ltd. We study the operator problem in cognitive radio cellular networks to maximize the secondary revenue, while limiting the interference with primary receivers and guaranteeing the QoS requirement of both primary transmitters and the admitted secondary users. We propose three QoS aware admission and power control schemes, including one optimal scheme QAPC-dynamic, one greedy heuristic scheme QAPC-greedy, and one minimal SINR removal scheme QAPC-MSRA. QAPC-greedy is recommended considering the trade-off between the computation complexity and performance gain.