基于Q学习的航空集群网络认知抗干扰策略  

Q-Learning Based Cognitive Anti-jamming Strategy for Aeronautic Swarm Network

作　　者：黎海涛[1] 罗佳伟 吕鑫 方正[2] 仇启明[2] LI Hai-tao;LUO Jia-wei;LV Xin;FANG Zhen;QIU Qi-ming(Beijing University of Technology,Beijing 100124,China;China National Aeronautical Radio Electronics Institute,Shanghai 200233,China)

机构地区：[1]北京工业大学,北京100124 [2]中国航空无线电电子研究所,上海200233

出　　处：《中国电子科学研究院学报》2021年第10期985-990,998,共7页Journal of China Academy of Electronics and Information Technology

基　　金：航空科学基金资助项目(2018ZC15003);。

摘　　要：为了增强复杂电磁环境中航空集群战术网络的抗干扰能力,文中针对部署同时收发电台的航空集群战术网络(ASN),提出改进能量检测的多时隙干扰感知方法,推导出电台干扰感知的虚警概率和检测概率的闭式表达。然后,以机载电台的吞吐量最大为目标并结合干扰感知结果,分别提出基于Stateless Q学习和Fairness Q学习的认知抗干扰策略,为电台配置最佳的发射功率和频率信道。仿真结果表明,所提认知抗干扰策略均能提高航空集群网络容量,且ASN采用Fairness Q学习抗干扰策略的公平性最高而适用于网络容量最大化的场景,Stateless Q学习抗干扰算法的平均吞吐量最高,故而适用于最大化各电台容量的应用场景。In order to enhance the anti-jamming capability of aeronautic swarm tactical network(ASN)in complicated electromagnetic environment,we present a multi-slot jamming sensing scheme based on the improved energy detection method for ASN employing airborne radios with the simultaneous transmit and receive feature,and derive the closed expression of false alarm probability and detection probability.Then,in order to maximizing the throughput of airborne radio,different Q-learning anti-jamming strategy based on jamming sensing is developed to allocate an optimal configuration of transmitting power and spectrum channel to each radio.The simulation results validate that cognitive anti-jamming performance of ASN can be improved by Q learning anti-jamming strategy,and the provisionally Fairness Q-learning based anti-jamming strategy has the highest fairness,but its average throughput is lower than the anti-jamming strategy based on the Stateless Q-learning algorithm.

关 键 词：航空集群网络 多时隙干扰感知 改进能量检测 Q学习 认知抗干扰 

分 类 号：TN92[电子电信—通信与信息系统]