| 基于博弈学习的认知网络安全通信信道选择算法 |
| |
| 作者姓名： | 卢娜 |
| |
| 作者单位： | 商丘职业技术学院机电工程系 |
| |
| 基金项目： | 河南省科技攻关项目“基于强度滤波器的紧邻多目标高效率跟踪技术研究”(182102210116);河南省高等学校重点科研基金计划资助项目“基于概率假设密度滤波的多目标连续跟踪算法研究”(17A520052) |
| |
| 摘 要： | 认知用户针对多窃听节点场景下认知网络物理层安全通信,为最大化安全通信速率进行动态信道选择。非合作博弈建模认知用户可以在以下三种基于博弈学习的信道选择算法中实施安全通信信道选择行为,即虚拟行动学习算法、经验权重吸引(experience-weighted attraction,EWA)学习算法以及强化学习算法。算法仿真结果表明:虚拟行动算法与EWA算法性能相近,但虚拟行动算法收敛较慢且计算复杂度高,EWA算法收敛速度快且算法复杂度适中,而强化学习算法性能较差且收敛慢但计算复杂度低。
|
| 关 键 词： | 认知无线电 物理层安全 动态信道选择 博弈学习 |
Cognitive Network Secure Communication Channel Selection Algorithm based on Game Learning |
| |
| Authors: | LU Na |
| |
| Institution: | (Department of Mechanical and Electrical Engineering,Shangqiu Polytechnic,Shangqiu 476000,China) |
| |
| Abstract: | The cognitive user performs dynamic channel selection for maximazing the secure network communication rate for the cognitive network physical layer secure communication in the multi-eagle node scenario.Non-cooperative game is used to model the behavior of cognitive users’channel selection,and three game learning-based channel selection algorithms are proposed,i.e.,fictitious play learning algorithm(FPL),experience-weighted attraction learning(EWAL)algorithm and reinforcement learning(RL)algorithm.The simulation results indicated that,FPL and EWAL algorithms have the similar performance,however,FPL converges slowly and it is computation expensive,while EWAL converges fast and its complexity is moderation,RL’s performance is the worst and converges more slowly but with the lowest complexity. |
| |
| Keywords: | cognitive radio physical layer security dynamic channel selection game learning |
| 本文献已被 维普 等数据库收录！ |
|
