中文摘要：

认知频谱接入技术通过感知侦测并伺机接入空闲频带，大大缓解了频谱资源稀缺和利用率低下的矛盾．在未知信道占用信息情况下，认知用户如何快速找到最优空闲信道，并避免相互之间的冲突，是认知接入算法的难点和关键技术．本文中提出一种低复杂度的分布式认知接入算法，通过学习物理层的感知结果和链路层的发送确认信息，同时获取主用户的信道占用概率和从用户之间的冲突概率，能快速选择感知信道，并有效避免了从用户之间的竞争．仿真结果表明，该算法在未知信道占用概率信息的情况下，网络平均吞吐量明显优于贪婪等算法，很好地解决了认知网络中多个从用户频谱共享问题，同时收敛速度较快．

英文摘要：

Cognitive medium access, built on sensing and accessing idle frequency bands, greatly releases the confliction between the frequency scarcity and the low utilization of a precious radio resource. How to quickly probe and optimally exploit multiple vacant channels with unknown parameters as well as to avoid the competition among cognitive users, is key to the cognitive access strategy. This paper proposed a low complexity and efficient distributed access strategy based on joint design of sensing results in physical layer and acknowledgment information in medium access layer, which not only learned from primary time-varying traffic information but also considered spectrum-sharing among competitive cognitive users. Finally, numerical results show that the algorithm can gain higher average throughout compared with greedy strategy etc. without a prior transition probability of each channel and has faster conver- gence.