基于Message Passing的认知无线电快速资源分配

Fast Resource Allocation for Cognitive Radio Based on Message Passing

针对多用户的OFDM认知无线电系统上行链路,提出一种基于Message Passing的分布式快速资源分配算法。该算法以认知系统总发射功率最小化为优化目标,综合考虑了认知用户对授权用户的干扰、总发射功率预算以及认知用户之间的比例公平性等约束条件,将资源分配分为子信道分配与功率分配相继2个步骤,构建资源分配的因子图,通过在节点间迭代地传递信息直至最终完成分布式的资源分配。分析和仿真结果表明,该算法在保证系统通信性能及资源分配公平性的前提下降低了系统总发射功率,并且运算效率得到了明显提升。

Aiming at the uplink of multiuser OFDM-based cognitive radio （CR） system,a fast distributed resource allocation algo- rithm based on message passing （MP） technique is proposed.The algorithm aims at minimizing total transmission power of the CR sys- tem,based on comprehensive consideration of constraint conditions such as the interference introduced by second users （SUs） into pri- mary user （PU） ,the total transmission power budget, the proportional fairness among SUs, etc.The algorithm adopted two steps：subcar- rier allocation and power allocation.By establishing a factor graph and passing messages iteratively from nodes to nodes,it can perform the distributed resource allocation to the end.The simulation results and the analysis show that, on the premise of ensuring communica- tion performance and fairness, the proposed algorithm can reduce the total transmission power and improve the computation efficiency obviously.