当前位置: X-MOL 学术Complex Intell. Syst. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Dueling deep Q-networks for social awareness-aided spectrum sharing
Complex & Intelligent Systems ( IF 5.8 ) Pub Date : 2021-04-27 , DOI: 10.1007/s40747-021-00382-1
Yonghua Wang , Xueyang Li , Pin Wan , Le Chang , Xia Deng

In overlapping spectrum sharing, due to the complexity of cognitive environment, it is a real challenge for a secondary user (SU) to correctly sense the usage of the spectrum in real time. To tackle this challenge, a social awareness-aided transmit power control policy for SUs is developed. First, a social network composed of a group of third-party sensing nodes that do not share the spectrum with the PU is established, which helps an SU collect the power information of the PU. Then, we design a Dueling Deep Q-Network (DQN) model to achieve efficient dynamic spectrum sharing between the PU and the SU with the power information collected in the social network. Experimental results show that the spectrum sharing success rate is higher and the comprehensive performance is improved with the sensing nodes selected by the social relationship. Moreover, compared with other deep reinforcement learning (DRL) algorithms, the performance of Dueling DQN is more stable on our targeted spectrum sharing problem.



中文翻译:

借助深度Q网络进行社会意识辅助的频谱共享

在重叠频谱共享中,由于认知环境的复杂性,对于二级用户(SU)而言,实时正确地感测频谱的使用是一个真正的挑战。为了应对这一挑战,开发了一种针对SU的具有社会意识的发射功率控制策略。首先,建立由不与PU共享频谱的一组第三方传感节点组成的社交网络,这有助于SU收集PU的功率信息。然后,我们设计了决斗深度Q网络(DQN)模型,以利用社交网络中收集的功率信息在PU和SU之间实现高效的动态频谱共享。实验结果表明,通过社会关系选择感知节点,频谱共享成功率更高,综合性能得到改善。而且,

更新日期:2021-04-28
down
wechat
bug