Research On Intelligent Anti-jamming Decision Algorithm In Wireless Network

In the confrontation environment,there is malicious electromagnetic interference in the wireless network,which makes the environment more and more complex.The traditional anti-jamming methods have certain limitations and cannot meet the requirements of communication transmission in battlefield environment.When the enemy adopts intelligent jamming,it is difficult for a single node and a single link to effectively resist enemy jamming due to limited resources.Therefore,how to effectively combat interference has become an urgent problem to be solved.Therefore,focusing on the anti-jamming problem,this thesis studies the channel selection and power selection of nodes,and puts forward two intelligent anti-jamming decision methods: off-line learning and on-line confrontation for different jamming strategies.The main work is:(1)In this thesis,an anti-jamming decision algorithm based on Elite Particle Swarm Optimization(EPSO)is proposed for dynamic interference.In this algorithm,the communicator retains the global and local optimal location information through offline learning,and obtains the best anti-jamming strategy under different jamming power.Although the optimization effect of the algorithm is good,it cannot meet the needs of real-time decision-making due to the slow iteration speed.Therefore,an anti-jamming decision method based on back propagation(BP)neural network is proposed.This method uses the EPSO algorithm to obtain a large number of data sets,train the neural network model,and then use the trained model for online decision-making.The simulation results show that under dynamic interference,the relative error of throughput is very small and real-time decision-making can be realized by using BP neural network model compared with the EPSO algorithm.(2)Aiming at intelligent jamming,an intelligent anti-jamming decision algorithm based on Q-Learning online confrontation is proposed.In the process of online confrontation,the communicator and the jammer decide to adopt the "exploration" or "utilization" mode according to the exploration rate parameters respectively.If you choose to continue "exploration",the communicator randomly selects the action strategy and updates their Q table according to the transition probability;If the "use" mode is selected,the communicator will select the action strategy with the highest Q value in the corresponding state of the Q table and update the Q table.The simulation results show that the algorithm can realize on-line confrontation and achieve good antijamming effect in the face of intelligent interference.In conclusion,the cognitive anti-jamming method based on off-line learning has a good anti-jamming effect.Based on BP neural network model,real-time decisionmaking can be realized,and the relative error between the throughput of the EPSO algorithm is small.The Q-learning anti-jamming method based on online learning can realize online confrontation,and the anti-jamming effect is very good.