基于UCB改进的短波认知多信道选择算法  

Improved HF cognitive multi-channel selection algorithm based on UCB

作　　者：王董礼 王叶群[1] 孙启禄[1] 张雯鹤 王也 Wang Dongli;Wang Yequn;Sun Qilu;Zhang Wenhe;Wang Ye(College of Information and Navigation, Air Force Engineering University, Xi an 710077, China)

机构地区：[1]空军工程大学信息与导航学院

出　　处：《东南大学学报（自然科学版）》2019年第5期897-903,共7页Journal of Southeast University：Natural Science Edition

基　　金：国家自然科学基金资助项目(61701521);陕西省自然科学基金资助项目(2018JQ6074)

摘　　要：为了提高UCB算法在短波多信道选择时的成功传输率,结合短波宽带应用背景,提出一种基于UCB改进的短波认知多信道选择算法.该算法根据认知无线电原理,通过把应用认知无线电技术的短波电台作为认知用户,利用强化学习中UCB索引值对信道进行排序,并引入次序感知策略进行多信道选择,直到选出所需数量的空闲信道,从而为认知用户提供等效宽带传输.仿真结果表明,与原始UCB多信道选择算法相比,所提算法通过次序感知策略适当增加感知次数,在所需信道数量为4时,成功传输率最大提高了71.61%,累积收益平均提高了79.24%,累积数据传输时间平均提高了3.67倍.To improve the successful transmission rate of UCB (upper confidence bound) algorithm in HF (high frequency) multi-channel selection, combined with the background of HF broadband application, an improved HF cognitive multi-channel selection algorithm based on UCB is proposed. By regarding the HF radio station with cognitive radio technology as the cognitive user according to the principle of the cognitive radio, the UCB index in reinforcement learning is utilized to rank the HF channels, and the sequential sensing strategy is introduced to select multiple channels, until the number of idle channels required is selected to provide an equivalent broadband transmission for cognitive users. Simulation results show that compared with the original UCB multi-channel selection algorithm, the proposed algorithm appropriately increases the number of channel sensing times by the sequence sensing strategy: when the required number of channels is 4, the successful transmission rate can be improved by a maximum of 71.61%, the cumulative rewards can be improved by an average of 79.24%, and the cumulative data transmission time can be improved by an average of 3.67 times.

关 键 词：短波通信 认知无线电 UCB 多信道选择 

分 类 号：TN92[电子电信—通信与信息系统]