Research On The Key Technology Of Communication Security Based On Reinforcement Learning In Mobile Fog Computing

Fog computing is a technology that provides distributed computing,storage,and other services between cloud data centers and Internet of Things(IoT)devices,while leveraging network edge authentication to provide services that interact with the cloud.With the advent of 5G era,the security of the communication between fog nodes and mobile terminal nodes in the framework of mobile fog computing network on the basis of ensuring high efficiency has become a focus of current research.In the face of various attack methods and threats such as eavesdropping attack,impersonation attack and user-oriented intelligent attack,new challenges are presented for key technologies such as attack detection and security defense in mobile fog computing network.Therefore,this paper proposes the detection scheme of impersonation attack and the scheme of intelligent attack defense based on DQL(Double Q-learning)algorithm,aiming to enhance the security performance of mobile fog computing.Main research contributions are as follows:(1)Physical layer security technology(PLS)for wireless channel and wireless network communication security technology based on reinforcement learning are analysed.The reinforcement learning algorithm for mobile fog computing security and its potential advantages combined with game theory are studied.The types of attacks in mobile fog computing and the harm of attacks are summarized.(2)A scheme of detecting impersonation attacks based on DQL algorithm in mobile fog computing is proposed.In the face of the impersonation attack in the fog mobile computing environment,firstly,the excessive Q-value estimation problem in the Q-learning algorithm is solved with the aid of the PLS technology in the channel parameters and the zero-sum game based on Expected Utility Theory(EUT)calculation.Then,the scheme obtains the best impersonation attack test threshold and reduces the fault alarm rate and miss detection rate of the system,Thus,the accuracy of the detection of impersonation attack between the users and the fog nodes is improved.(3)An intelligent attack defense scheme based on DQL algorithm in mobile fog computing is proposed.At present,researchers seldom study the subjectivity of the attacker.To solve this problem,first of all,the scheme describes a system model involving malicious users in mobile fog computing.A variety of intelligent attack types are comprehensively considered.Secondly,different from EUT,a static method of subjective zero-sum game between malicious users and legitimate users is constructed based on Prospect Theory(PT).And then,a dynamic subjective game scheme based on DQL algorithm is proposed to obtain the optimal defense strategy and resist the threat of active intelligent attack.Finally,by comparing with the schemes about other related algorithms,this defense scheme can enhance the security of mobile fog computing network and improve the protection performance.