Research On Multi-sensor Data Fusion Methods In Wireless Sensor Network

Multi-sensor data fusion technology is the national key research project, has been listed as the first of the ten emerging technologies affecting human future.The technology has been widely used in different area, such as medical treatment, agriculture and industry.Multi-sensor information fusion technology by integrated from different sensors comprehensive measurement information to processing data preconditioning, data association,data decision making and data fusion,etc.Make a correct judgment and decision process is the purpose of the fusion technology.This paper focuses on the key issues of the fusion technology, the main research content of this paper includes data preconditioning,redundancy data elimination, data association, data fusion, etc.Firstly, this paper introduces the theoretical bases and application method, elaborate it from the confluent basic model, confluent basic level and confluent mathematical theory basis aspect, and analyze the application method of existing data fusion algorithm and its merits and demerits of processing data.Secondly, in particular network structure, aiming at the problem that as a result of failure of sensor measured outcome, such as excessive data, environment noise, zero drift in nodes in data processing, this paper proposed multi-sensor data fusion method based homogeneity and multi-sensor data fusion method based heterogeneity.When study multi-sensor data fusion method based homogeneity, the existing fusion algorithm set the fusion threshold merely according to the empirical value, leading to the low robustness of fusion performance even failure. Aim at the problem, this paper defines a fusion method based fuzzy adaptive weigh, through introducing the conception of the similarity degree among nodes in social computing, defines spatial correlation parameter and trust degree value, utilizes Consistent integration degree to reassign nodes Weight coefficient, gets the final fusion expression. And the efficiency of the algorithm is verified by the simulation example.When study multi-sensor data fusion method based heterogeneity, adding processing method based on relative gradient, proceeding pretreatment in the previous stage before data fusion, introducing the method of derivative function extreme and median filter to filter redundant data and wrong data, get the best value in pretreatment period; defining compatibility coefficient and correction factor, from a global perspective, considering the relative information among evidence, and adopting the improved D-S evidence theory to proceed data fusion, obtaining the final fusion result. The experiment has shown that the algorithm of this paper has faster convergence rate, higher stability, and its judgement to fusion results are more suitable to actual conditions.