Research On Reinforcement Learning In Wireless Communication And Physical Layer Security

With the widespread popularity of mobile devices with high mobility and convenience,the scale of wireless communication services is increasing.At the same time,a large amount of data communication will also bring about the increase of energy demand and the problems of wireless transmission safety.How to solve these problems has become an increasingly important focus.This paper mainly researches the security issues in the process of wireless communication transmission.The main contents are summarized as follows1.Considering a D2 D energy harvesting system,an online learning energy transmission scheduling scheme combined with the idea of Q learning is proposed.This scheme can determine the charging time of the battery according to the channel information and the remaining power state of the battery to make the energy consumption of the energy source,the number of exhaustion times of the battery,and the number of times of battery overflow are as small as possible.The simulation results show that the scheme can obtain better performance.2.Considering an energy harvesting system in which an attacker exists,in order to counter the attacker's interference to the wireless communication process,an anti-jamming transmission scheme based on reinforcement learning is proposed and compared with traditional game theory-based methods.Through simulation,the superiority of this method is verified.3.Considering a D2 D underlying cellular network with an eavesdropper,a D2 D pair channel allocation method based on reinforcement learning is proposed to use D2 D pair communication to interfere with eavesdropping,so as to improve the security probability of the system.The simulation results verify that this method can achieve better performance.