Data fusion algorithm based on ultrasonic sensor network

Abstract

For observation information from multiple sensors to analyze consolidated. To arrive at decisions and the information needed to estimate task processing, a novel fusion method is proposed based on the approach degree and weights. The method calculate mean and variance based on the measured sensor's data, Using the maximum and minimum approach degree of this fuzzy set, the approach degree of the measured data from various sensors is processed quantitatively, eliminating outlier data by Grubbs method, assigned the weight's of data measured in the fusion process reasonably, so that the final expression of the data fusion is obtained, thus the data fusion of multi-sensor is realized. The results demonstrate that this method can bring higher fusion precision and more suitable for microcontroller and embedded systems applications. © Springer-Verlag Berlin Heidelberg 2013.