Land Battlefield Armored Targets Aggregation Algorithm

As an important function which the situation assessment, the second level ofInformation Fusion, needed to implement, the issue of the target aggregation is aprolongation of the first level of Information Fusion and a key point of the advancedfusion levels. In the battle environment, the targets in different situation have differentstructures, and the different parts in which act differently. In the real environment of theunited battle, there are various battle objects, which have complex relationships.Meanwhile, the battle situation changes so quickly that the commanders couldn't makeany effective decision in front of the abundant information. By using the technology oftarget aggregation, the information of the threat units are abstracted and partitioned, andthe hypothesis of the martial systematic units is formed. tn that case, not only therelationships between the situation elements could be determined, but also variousactions of the armored targets of land battlefield could be explained. Consequently, thereported information could be the basis of the battle attempt estimation and the threatassessment. In this paper, these above problems are discussed and the main work can besummarized as follows:1,The fundamental theory of multi-sensor Information Fusion, including themodel of the system and the basic technologies, is systematically introduced.The functions that the situation assessment needed to complete are addressed. Athree-level functional processing model, which consists of current situationperception, current situation recognition and future situation projection, is setup and each level is analyzed in detail. The characteristics and the attributes ofthe target, clustering based on situation assessment are analyzed.2,The clustering algorithms in the field of pattern recognition are studied,including the definition and the criterion of the algorithmic efficiency. Theapplicability of the clustering technology in battle target problem is discussed.The implementations of common algorithmic routines are also presented indetail.3,The armored target aggregation model based on the environment of the landbattlefield is built, and common algorithms are emulated by using theMATLAB language. Then, two improved algorithms are presented.4,The relationship and hierarchy of the target are studied, a hierarchicalaggregated clustering algorithm is proposed, and then, the different hierarchiesare abstracted and partitioned, and the relationship between the situationelements is opened out.Although the algorithms in this paper are proposed under the study of the armoredtargets of land battlefield, they could also be used in the issue of targets in the air andocean battlefield, if the algorithm improved in some certain way.