Research On Anti-jamming Communication In UAV Based On Reinforcement Learning

The Internet of Things node can transmit messages aided by the unmanned aerial vehicle who acts as the relay node,which are vulnerable to be attacked,such as the jammer who can change the interference intensity randomly.The existing anti hostile jamming schemes of the unmanned aerial vehicle usually adopt the frequency hopping,single flight path planning and power control strategies which can resist the attack of the fixed interference power,but they are helpless to the intelligent jammer with variable interference signal strength.Therefore,this thesis designed a system model about the unmanned aerial vehicle aided the Internet of Things,and assumed that there was the intelligent jammer in this system model.Aiming at solving the problem of anti hostile interference in the system model,this thesis optimized the transmission power of the Internet of Things node,the signal power and moving trajectory of the unmanned aerial vehicle to improve the anti-jamming performance of the system.It was assumed that all the three nodes in the system model have known the channel model and the interference model,this thesis established the attack and defense game model of the unmanned aerial vehicle aided the Internet of things,the Stackelberg equilibrium and Nash equilibrium of the game were solved respectively,deduced the existence conditions of the equilibrium and verified the uniqueness of the equilibrium point,and revealed the influence of the channel gain and the unmanned aerial vehicle flight distance on the anti-jamming performance of the system.The simulation results show that compared with Nash equilibrium,the jamming performance of the intelligent jammer under the Stackelberg game model is the larger,so the anti hostile jamming algorithm were designed based on the idea of the Stackelberg game.Assumed that the ground sensor node and the unmanned aerial vehicle in the system model do not know the interference model and the channel model,but the intelligent jammer knowed all the transmission information in the system.In this thesis,the ground sensor node and the unmanned aerial vehicle were regarded as two agents,and reinforcement learning technology was introduced to design two algorithms to dynamically optimize the transmission power of the Internet of Things node,the transmission power and moving trajectory of the unmanned aerial vehicle.The first algorithm combined with the Q＿Learning,the greedy strategy was used to learn the optimal behavior of the agents.The second algorithm was combined with the Wo LF-PHC technology,two different learning rates and the average strategy were introduced on the basis of the first algorithm,and the two agents can learn with better strategies.Simulation results show that the proposed two anti-jamming algorithms can improve the anti-jamming performance of the system.Compared with the anti hostile jamming algorithm based on Q＿Learning,the proposed anti hostile jamming algorithm based on Wo LF-PHC can improve the efficiency of the ground sensor nodes by 145%,the efficiency of the unmanned aerial vehicle by 233%,and the signal to interference noise ratio of receiver by 85.7%.