Research On Adaptive Jamming Technology In Cognitive Electronic Countermeasures

In the face of the challenges brought by the high dynamic,high realtime and high complex electromagnetic environment in the information battlefield,the automation and intelligence degree of modern electronic equipment keeps improving,and the cognitive electronic countermeasures technology becomes the inevitable trend of the electronic countermeasures.Adaptive interference is one of the core functions of cognitive electronic countermeasures,which requires the jammer to break through the traditional preloading mechanism and have self-adaptive jamming ability.In other words,based on the real-time perception of the electromagnetic environment,the jammer can efficiently and autonomously adjust the parameters of the transmitter and receiver to adapt to the changes of the environment,so as to improve its quick-response ability and decision-making reliability.In this thesis,three key technologies in the adaptive interference system are studied,including target state identification,interference decision and interference effect evaluation.A complete adaptive interference closed-loop is established,where the algorithms are proposed to achieve optimal interference in complex electromagnetic environment under the constraint of limited interference resources.The main work of this thesis can be divided into the following three parts:Aiming at the problem of radar working state recognition under the condition of little prior data,a classification model of working state based on "supervised+unsupervised" structure is constructed in this thesis.According to the distinction of radar working state in timing and range characteristics of pulse parameter,firstly,the signal timing characteristics are extracted for feature fusion,and the combination modulation of pulse parameters is preliminarily classified.Secondly,the density clustering of parameter range is carried out to obtain the secondary division of radar working state.Simulation experiments verify the advantages of the model structure in unknown radar working state recognition.In order to solve the problem of adaptive interference decision against complex objective states,an interference strategy optimization method based on dual reinforcement learning is proposed in this thesis.On the basis of modeling the interference process as a finite Markov decision,the dimension of high-dimensional interference action space is reduced and two interactive Q-learning models are established,where the jamming mode and waveform are hierarchically selected and jointly optimized during the confrontation process.Simulation experiments verify that this jamming model has the ability to learn the complex state transition strategy of radar mode switching and waveform agility.Besides,because of the reduced dimensionality of jamming action space,the globally optimal solution can easily be found with a shorter convergence time.In order to evaluate the jamming effectiveness based on the jamming side,an evaluation indicator vector space is constructed in this thesis,and a method of measuring the offset distance of the indicator vectors before and after jamming is proposed to evaluate the jamming effectiveness.The weights of indicators are updated in real time according to the radar data,forming a dynamic jamming effectiveness evaluation model.The simulation results show that this dynamic measurement can accurately reflect the radar state transition and obtain reliable evaluation results.In addition,the evaluation result of jamming effectiveness is served as feedback information of the dual reinforcement model to assist the optimization of jamming strategy,which overcomes the dependence on the accurate estimation of information from radar side when model updates.