基于POMDP模型的分布式机会频谱接入算法

针对认知无线电（cognitive radio,CR）信道的动态特性,以部分可观测马尔科夫决策过程（POMDP）为模型对认知无线电网络用户的频谱感知和频谱接入过程进行研究,提出了基于POMDP模型的分布式机会频谱接入算法.该算法利用网络信道的历史频谱感知信息对主用户接入信道的状况作出估计,以认知用户吞吐量最大化为目标进行频谱接入.同时,通过贪心算法得到此优化策略的次优解,降低了最优策略的计算复杂度.论文分析了认知用户接入吞吐量与网络中信道数目以及信道状态转移概率之间的关系,将贪心算法与随机检测接入算法进行了仿真比较.仿真结果显示,该算法获得的吞吐量比随机检测接入算法提高了约25％,能够更有效地做出接入策略.

Decentralized Opportunity Spectrum Access Based on POMDP Model

According to the dynamic features of cognitive radio channel state,aimed at the spectrum sensing and access in cognitive radio networks,this paper proposes a decentralized opportunistic spectrum access algorithm,based on partially observable Markov decision process （POMDP） model.The algorithm uses the historical information of spectrum sensing to estimate the next channel access state of authorized users.Based on the estimated channel access state,cognitive users access channel to maximize the throughput.Meanwhile,the greedy algorithm is used to obtain the suboptimal solution,thus reducing the computational complexity of the optimal strategy.The relationship is analyzed between the throughput of cognitive users and the number of channels as well as the transition probability of channel state,and the greedy algorithm is compared with the random access algorithm.Simulation results show that the throughput of the greedy algorithm is higher than that of the random access algorithm about 25％,thus it can provide the access strategy more effectively.