基于UCB的短波认知信道选择算法

HF Channel Selection Algorithm Based on UCB for Cognitive Radio Application

针对短波频段用户间冲突碰撞严重的问题,应用认知无线电原理,提出一种改进的UCB(Upper Confidence Bound)算法进行动态频谱接入以缓和相互干扰。该算法设置信道质量差异因子,利用探索结果动态调整信道数量和探索系数,缩减探索范围,平衡探索和利用比重。通过建立适当的评价准则,分析该算法在机器学习和认知背景下的性能。仿真结果表明,所提算法能够快速收敛于最优信道,较原始UCB算法和随机信道选择算法具有较高的成功传输率和较低的累积接入损失。

Aiming at the heavy collision among communication users in the High Frequency（HF）band,with the cognitive radio principles,a novel channel selection algorithm based on Upper Confidence Bound（UCB）is proposed to mitigate mutual interference through dynamic spectrum access.By introducing the difference factor of channel quality as well as dynamically changing the channel number and exploration-exploitation coefficient with the explored results,the proposed algorithm can decrease the exploration range and balance the proportion of exploration and exploitation.Evaluation criteria are established appropriately to analyze the performance of the proposed algorithm in machine learning and cognitive radio contexts.Finally,simulation results demonstrate that the proposed algorithm can converge to the optimal channel fast and has a higher successful transmission probability and a lower cumulative regret in comparison with the original UCB algorithm and the random channel selection algorithm.