Abstract: In cognitive radio network, where Cognitive Users（CUs） share spectrum bandwidth with the Primary Users（PUs）, it is necessary for cognitive users to perform power control to guarantee their QoS and avoid affecting PUs. In this paper, an improved power control algorithm based on non-cooperative game（IPC-NG） is proposed, with cost functions based on link gains and interference temperature. It is proved that the Nash equilibrium exists uniquely for the new net utility function through mathematical derivation and theoretical analyses. In the meantime, the convergence rate of the algorithm is analysed. Simulation results show that the algorithm not only avoids the influence for PU and guarantees the QoS of CUs, but also increases the system throughput and net utility.

Key words: cognitive radio, non-cooperative game, power control, utility function, price function

摘要： 在认知无线电网络中，当认知用户（CU）与主用户（PU）共享频谱带宽时，需要对认知用户的发射功率进行控制，以确保认知用户在满足自己的QoS且不影响主用户的前提下与主用户共享频谱带宽。利用博弈论的方法，设计了一种基于链路增益和干扰温度的代价函数，并据此提出了一种改进型非合作博弈功率控制算法（IPC-NG）。通过数学理论推导分析，证明了新的净效用函数存在纳什均衡且均衡点唯一，同时还分析了该算法的收敛速度。仿真结果表明，该算法不仅避免了对主用户的影响，也保证了每个CU的QoS需求，同时还提高了认知系统的吞吐量和净效用。

关键词: 认知无线电, 非合作博弈论, 功率控制, 效用函数, 代价函数