Research On USRP RIO Communication Jamming System Based On Machine Learning

With the continuous development of information technology,wireless communication technology has developed rapidly and has been widely used in both civil and military communication fields.In the field of communication countermeasures,various anti-jamming technologies have been researched to deal with various kinds of interference,while the traditional interference technology is relatively simple,and effective interference becomes more and more difficult.With the development of machine learning technology in the field of artificial intelligence,communication interference technology has a new development direction.This paper designs the communication jamming system framework based on the USRP RIO software radio platform,and further studies the communication jamming system based on reinforcement learning decision-making and the communication jamming system based on deep learning prediction.The intelligent jamming technology is realized through machine learning algorithm,which effectively enhances the jamming system.The main work of this paper is as follows:(1)In order to study the communication interference problem,this paper designs a communication interference system framework based on the USRP RIO software radio platform.According to the basic functional requirements of the communication jamming system,two main modules,the communication user and the jammer,which are composed of a transmitter and a receiver,are designed,and the functions and physical implementations of each module are described.(2)This paper proposes an interference channel selection scheme based on Q-learning,and realizes the construction of the communication interference system model and system flow design based on reinforcement learning.First,spectrum sensing is realized by using dual-threshold energy detection,and spectrum state information of communication users is obtained.Secondly,using the spectrum state as input,select the interference channel according to the Q-learning algorithm and calculate the reward,and update the Q-table through iterative training of the Q-learning algorithm.Finally,through the analysis of the experimental results,it is verified that the proposed algorithm has a higher effective interference probability than the traditional single-tone interference,multi-tone interference and random interference methods.(3)In this paper,a spectrum state prediction scheme based on CNN-LSTM network is proposed for communication spectrum prediction,and the model construction and system process design of the communication interference system based on deep learning are realized.First,the spectrum energy data of the communication user is acquired by using the energy detection method of spectrum sensing.Secondly,using the spectrum data as input,the CNN-LSTM network model is learned and trained,and the CNN-LSTM network prediction model is built by using the model parameters obtained from the training to predict the spectrum data by time slot,and the prediction results are used to select the interference channel and implement the interference.Finally,through the analysis of experimental results,it is verified that the proposed scheme can effectively predict the spectrum state for interference channel selection and achieve effective interference.