Applying triangular fuzzy number for multi‐sensor object recognition
Abstract
Purpose
Multi‐sensor data fusion (MSDF) is defined as the process of integrating information from multiple sources to produce the most specific and comprehensive unified data about an entity, activity or event. Multi‐sensor object recognition is one of the important technologies of MSDF. It has been widely applied in the fields of navigation, aviation, artificial intelligence, pattern recognition, fuzzy control, robot, and so on. Hence, aimed at the type recognition problem in which the characteristic values of object types and observations of sensors are in the form of triangular fuzzy numbers, the purpose of this paper is to propose a new fusion method from the viewpoint of decision‐making theory.
Design/methodology/approach
This work, first divides the comprehensive transaction process of sensor signal into two phases. Then, aimed at the type recognition problem, the paper gives the definition of similarity degree between two triangular fuzzy numbers. By solving the maximization optimization model, the vector of characteristic weights is objectively derived. A new fusion method is proposed according to the overall similarity degree.
Findings
The results of the experiments show that solving the maximization optimization model improves significantly the objectivity and accuracy of object recognition.
Originality/value
The paper studies the type recognition problem in which the characteristic values of object types and observations of sensors are in the form of triangular fuzzy numbers. By solving the maximization optimization model, the vector of characteristic weights is derived. A new fusion method is proposed. This method improves the objectivity and accuracy of object recognition.
Keywords
Citation
Wan, S. (2011), "Applying triangular fuzzy number for multi‐sensor object recognition", Kybernetes, Vol. 40 No. 3/4, pp. 589-599. https://doi.org/10.1108/03684921111133755
Publisher
:Emerald Group Publishing Limited
Copyright © 2011, Emerald Group Publishing Limited