Research On Midcourse Target Tracking Technology Via LEO Infrared Constellation

In missile defense,the ballistic missile early warning system is an extremely critical component.The space-based infrared early warning system has become an important development direction of missile early warning system because of its advantages of wide coverage,long detection distance,fast response time and not limited by geographical conditions.Among them,the space-based low Earth orbit(LEO)early warning system can play a very important role in the space-based infrared early warning system because it can continuously monitor and track the midcourse ballistic target.This dissertation takes the space-based infrared LEO constellation tracking of midcourse ballistic targets as the research background,and studies the midcourse target tracking technology based on the LEO warning system in complex environments.The key contributions are as follows:· The multi-sensor management method for LEO infrared warning system is proposed for tracking a midcourse ballistic target.First,an LEO constellation sensor management method based on covariance control is proposed from the perspective of the control process.The covariance control method uses the filter estimated covariance as a reference,generates feedback information by setting the desired covariance,and each time selects the sensor combinations with the filter estimated covariance closest to the desired covariance as the current sensor selection scheme.Then,from the perspective of the optimization process,the sensor management of the LEO warning system is considered to minimize the tracking error of the midcourse target.The posterior Cramér-Rao lower bound(PCRLB)is used as the performance evaluation criterion,and a PCRLB-based sensor selection model is developed.Under the convex optimization framework,in order to adapt to the task requirements of large-scale sensor networks,an efficient algorithm based on the proximal gradient method is developed to solve the sensor management model.· A new filter named measurement-driven sequential random sample consensus Gaussian mixture probability hypothesis density(MD-S-RANSAC-GM-PHD)filter is proposed for estimating the trajectory of midcourse ballistic targets.In the proposed filter,a measurement-driven birth intensity estimation algorithm is developedto generate the new birth target intensity adaptively,and since the measurement set used for birth intensity estimation may contain a large amount of clutter,a measurement set pre-processing method based on density-based spatial clustering and sequential random sample consensus(S-RANSAC)algorithm is proposed to eliminate the interference of clutter on generating new target birth intensity.Specifically,the proposed filter extends the standard GM-PHD filter by distinguishing between the persistent and the newborn target.The proposed MD-S-RANSAC-GM-PHD filter can effectively track the midcourse multitarget in a complex environment with a lot of clutter and unknown new birth target intensity.· A new filter named graph assisted improved GM-PHD(IGM-PHD)filter is proposed for estimating the trajectory of midcourse resolvable group target.In order to improve the performance of the original GM-PHD filter for closely spaced target in a group,the proposed IGM-PHD filter contains a novel weight renormalization algorithm,a new merging algorithm and track continuity.Modeling the structure of resolvable group targets based on graph theory,and the established graph model can effectively handle the dynamically changing groups.Combining the graph model and IGM-PHD,the proposed graph assisted IGM-PHD filter can effectively track the midcourse resolvable group target in a complex environment.· A new filter named graph partitioning extended target GM-PHD(ET-GM-PHD)is proposed for estimating the trajectory of midcourse unresolvable group target.In the proposed graph-based partitioning algorithm,in order to reduce the interference of clutter on the measurement set partition,a measurement set pre-processing method based on density-based clustering algorithm is presented.An intuitive directed weighted k-nearest neighbor(k NN)graph model based on graph theory is established to represent the relationship between different measurements in the measurement set that needs to be segmented.In the framework of directed k NN graph,a novel similarity metric based on shared nearest neighbor(SNN)is used,and a pairwise similarity that integrates the number of elements in the set of SNN and the closeness of data points is constructed.Based on the established directed k NN graph,the measurement set partitioning is transformed into a graph cut problem.The proposed graph-based partitioning algorithm has the advantages of strong adaptability to clutter environment,simple parameter selection,and high computa-tional efficiency.Combining the graph-based partitioning algorithm and ET-GMPHD filter,the proposed algorithm can effectively track the midcourse unresolvable group target in a complex environment.In summary,this dissertation focuses on the difficulties of midcourse target tracking technology based on LEO early warning systems in complex environments,and conducts research on LEO constellation sensor management,midcourse multi-target tracking,midcourse resolvable group target tracking,and midcourse unresolvable group target tracking problems.The proposed method has certain theoretical significance and engineering application value for the construction of China’s missile defense system.