Intelligent Anti-jamming Decision-making Scheme Based On Reinforcement Learning

With the increasing number of wireless communication devices and the increasingly complex and changeable spectrum environment,the traditional spread spectrum communication technology is also difficult to resist dynamically changing interference signals in this case.The intelligent anti-jamming communication technology has the ability of spectrum sensing.Through intelligent decision-making algorithm,the communication parameters can be dynamically adjusted in real time according to the interference signal,which can realize the anti-jamming communication ability in the complex environment.The core of intelligent anti-jamming technology is to realize anti-jamming decision knowledge base,solve the mapping from environmental interference to anti-jamming strategy,and guide the selection of anti-jamming actions in the communication process.In order to solve this problem,this paper studies the method of learning anti-jamming strategies from actual jamming signals.Firstly,using the energy detection method,the data processing process of calculating the actual received interference signal as the channel interference distribution vector is designed,and the vector is used to describe the environmental interference and serve as the input of the anti-interference decision algorithm.The realization principle of anti-jamming communication is to dynamically switch the communication channel to avoid the interference signal.In this paper,the learning process of anti-jamming policy is modeled as a Markov decision process,the state set,action set and reward function are defined,and the mapping problem is described by the Bellman equation.The policy solving process requires a lot of iterative operations.This paper studies the use of reinforcement learning techniques as a tool for policy solving.First,the basic principles of reinforcement learning,exploration and utilization problems,and commonly used exploration algorithms are introduced,and an improved ε-greedy algorithm is proposed in this paper.Then a two-stage Q-learning hybrid anti-interference decision-making algorithm is designed based on Q-learning,and the strategy is stored in the form of Q-value table.Through five modes of interference signals,the data processing process of the whole text and the overall feasibility of the anti-jamming decision algorithm are simulated and analyzed,and compared with the random selection algorithm of idle channels and the selection algorithm of maximum idle probability of channels,it is proved that the anti-jamming decision in this paper is the effectiveness of the algorithm.Finally,this paper designs a strategy synchronization algorithm based on a fixed sequence to realize the strategy synchronization of the sender and receiver,and ensure that the sender and receiver can jump to the same channel to continue data transmission after the antiinterference channel switching is performed.