Research On Situation And Threat Assessment Based On Intuitionistic Fuzzy Reasoning

Information in the upcoming operational environments is massive, complex, and highly distributed and redundant. Units of operations need to access fusion information of a battlefield, in order to form appropriate perception, timely and exactly to understand battlefield situation, to assess foe courses of action and capabilities, and to make decisions rapidly. The circumstances for decision making are usually quite complicated. Therefore, researches on situation and threat assessment (STA) of information fusion in high levels should be oriented to military decision and continuously seek new techniques and solutions. The aim of this dissertation is to perform the deep research on problems of STA by utilizing intuitionistic fuzzy sets (IFS) with the background of net-centric warfare (NCW).Firstly, intuitionistic fuzzy (IF) semantic match and operations of relations get into investigation. To the extensions of Zadeh's fuzzy sets (ZFS), transforms of IFS to ZFS and equivalent relations of IFS with Vague sets are expatiated. The equivalent properties of transforms of L-IFS to Z-fuzzy sets (FS), interval value IFS to L-FS, interval value (IV) FS to IFS, and IFS to IV IFS are proved by means of D-K operators. The semantic operators of IF are defined. The properties of temporal logic operators and extended operations over IFS are exposed. The properties of reflexive, symmetric, and transitive on IF relations (IFR) are explored with the proof of the theorem of combination law on IFR operations. The effects of Atanassov's operator on the IFR properties are described.Secondly, investigations of techniques for IF reasoning and confidence factors spreading proceed. To IF logic (IFL) reasoning and predicate calculus, a synthetic method for finding the truth of IFL propositions is proposed with a symmetric synthetic method derived on the basis of the hypothesis of an equilibrium status as a special instance, thus the forms of logic calculus on IFL are enabled to be the same as Zadeh fuzzy logic (ZFL). To conditional reasoning on IFL, mathematical formulas of inference compositional operations on implications, conditions, multiplications, and the multi-dimensions are derived. To issues of IFL reasoning with confidence factors, the effects of confidence factors spreading in rules on those of conclusions are analyzed. The related formulas for finding truth of conclusions are exposed. To approximate reasoning on IFL, mathematical formulas of inference compositional operations on generalized modus ponens, modus tollens, and hypothetical syllogism are derived. An approach to formation of IF equivalent matrixes by finding the transitive closure is presented with the proof of theorems of transitive closure and resembling matrix. Amethod for IFL reasoning with truth qualifications is proposed, and the related sets of the compositional formulas of operations are derived.Thirdly, to the problems of battlefield situation assessment (BSA), a technique based on IF decision is proposed. First, BSA is concluded to an issue of synthetic judgement oriented to situation understanding, and an IF model for synthetic judgement is derived with a proof of the model normalized. Then, a system of evaluating goals for BSA is constructed, and approaches to effectiveness measures for evaluating goals and normalization of their values are described. Subsequently, methods of Delphi integrated with analytic hierarchy process (AHP) are utilized for finding weight vectors of goals. Finally, the efficient synthetic assessment and comprehension to current battlefield situation can be made out by using evaluating instances of electronic war, air-defense counter-missile and counterwork in the course of joint operations of air defense.Fourthly, to the questions of threat assessment (TA), a method based on IFL reasoning is proposed. First, the major factors of effects of attacking targets from space or air on TA in joint air defense operations and nondeterminacy to TA are analyzed. The functional model for TA is constructed and the measurement rank of threat degree is divided. Then, the membership and nonmembership functions for input state variables and inference rules are built with the algorithms for reasoning devised. Subsequently, completeness and interactivity and consistency of rules contained in the rulebase are checked with a verification method to the rulebase presented. The validity is checked by providing TA instances with 20 typical targets.Finally, to improve TA accuracy of synthetic value, a method of adaptive neuro-intuitionistic fuzzy inference system (ANIFIS) is proposed. A Takagi-Sugeno model for TA on the basis of ANIFIS is constructed. The global approaching property of the model is proven with the network learning algorithms devised. The validity is checked by providing TA instances with 20 typical targets. An analysis by comparison with simulated instances from the two methods is made. The results show that the latter is superior in performance.