Research On Secondary Data Fusion Based On Trust

In recent years, with the development of information network technology, wireless sensor network (WSN) as a new type of intelligent network monitoring system plays an increasingly important role in the economic development and life. The networks usually deploy a large number of sensor nodes on the same target task of monitoring and controlling, multiple sensors would have a large amount of data, and to get on the objectives and task of authentic data description, which requires data fusion technology. However, in practical applications, due to various causes of uncertainty, the information collection process often causes a lot of redundant, ambiguous and even contradictory information, resulting in that data processing efficiency of data fusion system is low, the accuracy is not high, credibility is low and other issues. Therefore, how to improve data fusion technology to get a more accurate and reliable results which has a practical significance.This paper analyzes the current situation and related technology of data fusion theory, focus on the batch estimation theory and the general trust degree algorithm and other traditional methods of multi-source data fusion algorithm in the presence of low processing precision and poor stability. The paper presents the grade fusion algorithm, which is based on trust degree secondary fusion method. First, the algorithm with the median estimate in batches theory, can avoid the influence of the extreme data, each sensor integration through improved algorithms for the first time, can obtain more reliable local decisions; second, in view of the actual integration, data is hard be judged to be "credible" or "not credible", resulting in the absolute level of trust degree of the data, based on the membership function theory, the paper defines a level of trust for the fuzziness exponential function of trust, then obtains trust matrix through local decisions, the trust degree value in the range of [0,1], avoiding the confidence level of absolute and subjective, as a basis for local decision-making data for secondary fusion, to get the final fusion result.Finally, the data is simulated by MATLAB software platform, the experimental results show that, comparing to the arithmetic mean method, adaptive weight algorithm, the general trust degree algorithm. The results of improved the integration of the optimized method, the grade fusion algorithm, closer to the actual situation, and the algorithm has a good stability and stronger anti-interference, the absolute error is smaller. Therefore, in the presence of extreme data, the algorithm shows some advantages in data fusion, and can effectively improve the accuracy of data fusion and integration of reliability, and has some theoretical significance for data fusion.