Research On Theory And Method Of State Estimation Of Space Debris Group

Space surveillance is the main approach of the space situational awareness,also is the basis of the space control and space safety maintaining.The collision and breakup events produce a large number of space debris objects.Due to the small size and closely-spaced debris objects,intrinsically challenging issues of low object detection probability,complex data association and unresolved objects will be faced in space object surveillance.Thus,more effective radar information processing technology is needed for ground-based radar,which is always used for low earth orbit surveillance.This dissertation essentially concerns the important and difficult problems in space debris group surveillance,in which,the object of catalogue is no longer single object,but rather group object.Researches of the group catalogue are conducted.A series of more advanced and efficient information processing methods have been proposed.The primary work and achievements of this dissertation are as follows:First of all,the structures and orbits of the debris groups,which are produced in three typical collision and breakup events,are introduced and analyzed.The necessity of group catalogue is emphasized.The basic knowledge of the space debris group surveillance is presented from four aspects,i.e.,physical knowledge,object characteristics,data modeling and data processing.These are fundamental contents of the succeeding researches.Secondly,state estimation of space debris group is studied and two cases—resolved and unresolved objects—are considered.When debris objects are resolved,the group characteristics are supposed to be estimated,including the object distribution,number of objects in group and the state estimation of conspicuous object.When debris objects are unresolved,the group needs to be partitioned and the parts with possible unresolved objects are handled separately.When debris objects are resolved,a new filter based on the Random Finite Sets theory is proposed for state estimation of debris group.For debris group tracking,the issues of the Bayes moment approximation filters are analyzed,and addressed by the proposed labeled moment approximation filter.The proposed filter is a combination of two filtering processes and addresses the issues in miss detection information loss and mismatch in unknown detection model.By means of the information interaction of two filtering processes and a preprocessing procedure,the proposed filter has the capability of maintaining track continuity with low computational complexity.The new filter is implemented by Beta Gaussian mixture approximation.With a post-processing,the object distribution,number of objects in group,the state of conspicuous object and the group error ellipse have been estimated.Simulation scenario is similar with the data of Fengyun 1C debris group,in which the new filter demonstrates robust and accurate performance in different complex situations.When debris objects are unresolved,the data processing procedure on the level of echo signal is proposed,maximizing the information of signal.Debris group are filtered and partitioned,firstly.Then,the parts with possible unresolved objects are processed separately.The moment approximation filter is not able to estimate existence probability and probability density.Thus,a new labeled multi-Bernoulli filter is proposed for this situation.For dealing with unresolved objects,a jointly detection and tracking approach is proposed to estimate object number and state.Simulation has two parts.The good performance of the proposed filter is demonstrated in the first part.The second part shows performance of the proposed jointly detection and tracking approach in different measurement environments.The proposed approach distinguishes two objects well in high signal-to-noise ratio,and its performance degrades with the decrease of signal-to-noise ratio.Last,the study is taken from group characteristics further into characteristics of objects.After a period time of observation,the length of measurement data is enough,the knowledge of individual objects is supposed to be known.The characteristics of individual object can be deduced from the group characteristics.Because that debris objects in a group have similar space environment and orbital elements,there are relations between parameter variations of different objects.Based on this,a new method for ballistic coefficient estimation of low earth orbit space debris is proposed,using the ratios and greatly mitigating the effects of the model error.When two objects have similar orbital elements,the new method obtains more accurate and stable results than the directly estimation method.This algorithm is also validated for low Earth orbit space objects with perigee altitude higher than 600 km.This algorithm can be used to analyze the stable variable relationship between the values of objects' parameters in a group,which provides new information of debris group characteristics.The relationship between the parameters of different objects is considered as a criterion for judging which debris group a particular piece of debris belongs to.This will be an important basis for belonging judgment in group catalogue.