Research And Application Of Key Techniques Of Battlefield Situational Reasoning

With the development of information battlefield,more and more information and events need to be perceived on the battlefield.These information are complex,distributed and deceptive,which will greatly impede the real-time analysis of battlefield Situation,making the battlefield Commanders unable to obtain accurate battlefield description for the first time,and lose the opportunity to grasp the battlefield Situation.It is more and more important to evaluate the current Situation of battlefield by integrating the information between many sensors and other channels.There are some problems in traditional Situation assessment,for example,the logic that knowledge base can express is limited,the data between multiple platforms can not be shared,and the construction of reasoning templates is complex.In view of the above problems,this paper focuses on the idea of "Situation knowledge representation and analysis-the realization of battlefield data organization and data sharing-the selection and improvement of Situation reasoning algorithm".Knowledge representation mainly solves the expression of battlefield logic.Through the classification and modeling of battlefield data,combined with the Ontology Knowledge Base to complete data sharing.Situational reasoning algorithm is mainly reasoning the current battlefield Situation,and exporting the target intention in the current battlefield environment through the input of battlefield data.The main work of this article includes:1.In view of the knowledge representation of battlefield Situation.On the one hand,by comparing the commonly used knowledge representation methods and analyzing the needs of battlefield Situation environment,the ontology is identified as the knowledge representation method in this paper.On the other hand,the Situation elements in the battlefield environment are analyzed.First,combining the battlefield operational logic and the field characteristics,the construction method of the Situation ontology model is improved,and then the concrete implementation steps are given.Then,the construction of the Ontology Knowledge Base is completed.The construction of Ontology Knowledge Base mainly describes the battlefield operational logic,and provides a logical basis for the unification of data semantics of different platforms.2.In view of the diversity of data types in the battlefield environment,this paper first analyzes the data types in the battlefield,and proposes that the battlefield data should be divided into direct data and indirect data.On this basis,firstly,further refinement of data so that these data can be used directly.Secondly,the data are organized according to the classification results and battlefield logic,and Storing battlefield data in the Relational Database.Finally,based on the idea of "single platform database conversion to single platform ontology model-single platform ontology model integrated into the battlefield ontology knowledge base",to achieve multi platform logical interoperability and semantic specification,it provides a reasonable method for the interacting and sharing of battlefield data.3.In view of battlefield Situation reasoning,three kinds of algorithms are analyzed which based on Template Matching,BP Neural Network and Bayesian Network.The inference mechanism based on battlefield data as input and Situation mode as output is realized.Finally,the reasoning result simulation is given.Based on the simulation results,the advantages and disadvantage of the three methods in the Situation area are discussed.Because the Template Matching algorithm has the characteristics of high time delay and high accuracy,the Bayesian Network algorithm has the characteristics of low time delay and low accuracy.This paper proposes a " Template-Bayesian Network Fusion Reasoning Algorithm " which uses the advantages of Template Matching and Bayesian Network to compensate and amend their respective weaknesses,provides a accurate and real-time Situation observation model for the Commanders.