Research On Data Fusion Algorithm Based On Perceptual Cue

With the development of technology, the increasing methods of information acquisition provide huge data which can help people observe the world, and these data also reflect the world that we need to observe form different characteristics and different dimensions. Facing with the huge redundant, mutual and contrary information, people hope the technology of data fusion can achieve fast, accurate and reliable data analysis.However, due to various reasons and the uncertainty introduced by the information detection process, data fusion need to deal with the massive, redundant and ambiguous and even contradictory information, which reduces the efficiency and accuracy of data fusion system. So developing the autonomous, reactive and pre-mobile intelligent data fusion system is the pivotal approach to research the uncertain information. After putting more research for judgmatic guide and control management, this paper presents the perception guide data fusion model to solve the uncertain information and it achieves several innovations and achievements. The main contributions of this paper are summarized as follows:1. Aiming at the problem that prediction model can not deal with incomplete data, and the problem that predictions can not performan data credibility, this paper presents a predictive model of the fuzzy perception. The model uses the causal factors to assessment the degree of influence on the predicted results caused by incomplete information, when it uses the credible measure mapping to reflect the trust worthiness of the data. The Fuzzy perception prediction results not only reflect the relevant causality of data, but also measure the uncertainty of the data processing, the fuzzy membership is the information given by the system in two dimensions. The nature, computation rules and method for determining of the Fuzzy perception prediction model reasoning have laid the basis of the model in the actual environment.(Reflected in Chapter Ⅲ3.2and3.3)2. Aiming at the problem that the efficiency of the information system’s retrieval is low and the not high discrimination of identification problem, this paper proposed a Perceptual Decision-Making Rule model. This model realizes the increase of information retrieving by achieving the maximum of anticipant discrimination which is used to plan the Decision-Making Rule. Because decision direction is the biggest direction of information discrimination, the model can better distinguish the Decision-Making information. This paper analyses the method which can produce perception Decision-Making Rule combining the Local Attribute Importance proposed by Rough Set to produce perceptive Decision-Making Rule after getting some information. The Local Attribute Fuzzy Importance and Local Attribute Fuzzy Importance based on tolerance relation are used to deal with the fuzzy and incomplete problem of information in uncertain environment, so it can achieve best decision programming, and improve the efficiency and accuracy of identification, and the Perceptual Decision-Making Rule can make data realize the self-cognizing and quick-response.3. Aiming at the existing deviation problem of covering efficiency caused by the inaccurate detecting model in traditional fixed network detection model, this paper build the target perceptual detection model to obtain more accurate netted radar coverage efficiency. For conceal airplane used as radars, the best network’s deployment form of it has been simulated and analyzed.(Reflected in Chapter V5.2)For sensor mobile programming problem in mobile networking collaborative detection, this paper uses the correlation measure mapping and credible measure mapping of fuzzy perceptual prediction model to predict distribution probability, then the enlightening information is formed to realize the best searching path planning of mobile networking collaborative detection. The higher detecting efficiency can be achieved by Fuzzy Perceptual Prediction Model which was proved by the Monte Carlo simulation.(Reflected in Chapter V5.3)4. According to the initiative and pre-mobility requirements established by the intelligent data fusion system, this paper puts forward the Sensors Management Framework based on the Perceptual Decision-Making Rule guide. Through the multiple source sensors data fusion perception rule mining, this framework plan sensor management to obtain the biggest information incremental of perception, so that the multiple source sensors’management efficiency is increased.(6.1) Aiming at the instant and best perception prediction, this paper used dynamic distinguish entropy incremental model to realize the extraction of perceptual information, and statistical perception prediction realize the information extraction by using Local Attribute Fuzzy Importance The simulation based on space target detection verifies the efficiency and accuracy of the sensors management which is guided by Perceptual Decision-Making Rule.(6.2,6.3)Summaries and prospects have also been put forward in the final of this dissertation.