Research On Decision-making Method Of Multi-target Interceptor Cooperative Interception

The multi-object kill vehicle carries multiple sub-interceptors through a groundbased interceptor,which can effectively improve the interception effect on multiple targets.Considering the small size and large number of multi-target interceptors,it is necessary to make reasonable target allocation decisions and coordinated interception strategy decisions for multiple targets by combining their own interception capability range during the combat process.Based on the problem of intercepting multiple targets by multi-target interceptors,this paper studies the prediction of interceptor’s interception ability,target allocation method and cooperative interception strategy of multi-target interceptors.The main contents include:Aiming at the problem of interception capability prediction required before the interceptor makes a decision,the relationship between the interception capability range of the interceptor and its flight status and related basic parameters is analyzed.According to this,an absolute state of the interceptor is predicted by predicting the relative state of the interceptor Based on the deep learning method,and designed a deep learning neural network to predict the range of interceptor capabilities through the current flight status of the interceptor and the basic parameters of the interceptor.It has a wide application range,good scalability and fast calculation speed.advantage.Based on the predicted flight status of the interceptor and target in the case of undifferentiated,a relative coordinate system is proposed.Using this relative coordinate system,the relative motion relationship between the interceptor and the target in their respective clusters and the relationship between the interceptor and the target are proposed.The relative motion relationship of the system is characterized and the target allocation decision vector is calculated,and the randomness of the target’s steering maneuver direction is considered.Based on the deep reinforcement learning method,a deep reinforcement learning neural network for target allocation decision is designed,which has It has the characteristics of fast solution speed and good robustness.For the interception problem of multiple interceptors intercepting the same target,based on the idea of adjusting the interception capability range of the interceptor and the classic proportional guidance guidance law,a proportional guidance guidance law with target position offset is designed,and consideration can be given to Based on the estimated information of the target maneuvering direction,different collaborative interception strategies were designed.Based on the deep reinforcement learning method,a deep reinforcement learning neural network for decision interceptor interception strategies was designed.Finally,the network was verified by simulation for the collaborative interception strategies.The effectiveness of decision problems.This paper studies the key technologies of multi-target interceptor collaborative interception decision-making,studies the key issues of multi-target interceptor interception capability prediction,target allocation decision-making,and cooperative strategy decision-making,and explores emerging methods such as deep learning and deep reinforcement learning.The application of the traditional interception problem,the results obtained have certain reference significance for the combination of artificial intelligence technology and traditional ballistic missile interception technology.