Research On Classification Of Tactical-action And Prediction Of Penetration-capability Of Hostile Target

As the development of missile penetration technology,the traditional algorithm for ballistic forecasting,which uses current status information of hostile targets,has been unable to meet the operational requirements of missile defense combats.In order to improve the capability of the interception combats,the Early-Warning System needs to extract more information based on the original target state to provide effective auxiliary decision-making suggests for the interception strategy generation and target matching.As the researching hotspot in recent years,behavior pattern judgment and trajectory prediction are the key technologies to improve the ability of generating confrontatial assisted-decision.In this paper,we suggest a model of generating confrontatial assisted-decision information based on a compound neural network model for the judgment of tactical maneuvering behavior and the prediction of penetration capabilities of hostile targets.The results of tactical behavior-type judgement and penetration ability prediction points can be given by processing the historical state and extracting information features,which provide an instructive evaluation basis for the decision-making stage of the interception system.The main research contents of the paper include:Research on the judgment method of hostile target tactical actions.Establish a mathematical model of the tactical behavior of hostile targets,and use mathematical simulation methods to generate samples of the tactical behavior of hostile targets either with weapons interception or interaction.Build a neural network model based on a deep fully connected layers.Through continuous learning of training sample data,the network can quickly and accurately judge the tactical behavior of hostile targets.Research on the prediction algorithm of penetration ability based on Long Short-Term Memory Network(LSTM).Build a neural network based on LSTM.Through continuous learning of the training sample data,the LSTM network can predict the complete tactical behavior and obtain the target’s penetration ability boundary.Research on the generation of hostile target confrontation auxiliary decision information based on convolutional interaction pool.Consider under the offensive and defensive interactive environment,simulate and model the engagement process of tactical behavior of hostile targets in the face of coordinated interception by multiple interceptors.Build a complex neural network based on convolutional interactive pools,using the LSTM network as the front-end data encoder and back-end multi-decoder combination method,through the learning of the data set,the network can not only perform tactical actions on hostile targets Judgment can also accurately predict the penetration ability.