The Study Of The Rough Set Methods On Information Fusion

Information fusion is a kind of formal framework. Its process is to combine different sources of information by mathematical approaches and technical tools, and its purpose is to obtain useful information with high quality. The objective threat assessment is one of its key topics. In fact, it can be considered as a comprehensive evaluation process, which not only needs to consider some threats of enemy elements, but also gives full consideration to both the integrated battlefield situation and the results possibly occurring after confrontation.Rough set theory, as a useful mathematical tool for data presentation and analysis, is an effective method for analysis and research of information fusion. Its biggest feature lies in having no prior information beyond the data collection process in addition to the problem. Rough set theory can be applied to the data for analysis which is collected by information fusion, without special experimental process.This paper analyzes the conflict between Pawlak’s classical rough set and the incompleteness occurring in practical information systems, and then introduces the model of the set-valued information system proposed by Zhang et al. Moreover, we explain the dominance based rough set model presented by Greco, extend the multi-granulation rough set idea given by Qian et al, and put forward the idea of considering multi-granularity in relationship levels so as to define the upper and lower approximations. In addition, examples of data processing are employed to illustrate that the lower approximation of the new dominance-based multi-granulation rough set model is more accurate by means of a comparative study.For the application of the theater missile defense system, it is necessary to fusion multi-source radar information, make threat evaluation, and obtain the operational command information system (i.e. data chain). As is well known, the existing work has studied some typical cases of target threat assessment based on complete information systems. In order to deal with the more complex environment of multi-source data, this study takes the generalized dominance-based multi-granulation rough set model as an information fusion technology for target threat estimation, which can depict the case of set-valued data better in the practical target threat estimation. Moreover, compared with the existing literature [43] and [47], the proposed model is better in terms of adaptation.