An Improved CBBA Generation Method Based on Triangular Fuzzy Numbers

Abstract

Abstract Acknowledged as a robust tool for managing uncertain information, Dempster–Shafer evidence theory has seen significant progress in recent years, especially in the refinement of mass functions, also known as basic belief assignments (BBAs). This progress is particularly noticeable in complex domains where the effective handling of uncertainty is considered of paramount importance. Despite these advancements, the generation of complex mass functions, referred to as complex basic belief assignments (CBBAs), continues to be viewed as an open and challenging aspect within the framework of complex evidence theory. A method for CBBA generation based on triangular fuzzy numbers was introduced by Xiao, specifically applied to target recognition. However, despite its application, there is notable room for improvement in the recognition rate achieved by this method. In response to this gap, an improved CBBA generation method based on triangular fuzzy numbers is proposed in this paper. Notably, the consideration of attribute weights is incorporated into the CBBA generation process by this approach. This refinement is rooted in the recognition that, in practical scenarios, different attributes carry distinct levels of importance. Hence, adopting a more rational approach by assigning higher weights to crucial attributes becomes imperative. The proposed method is subjected to rigorous testing in the paper of target recognition, with its performance systematically compared against Xiao’s method and the conventional Dempster–Shafer evidence theory. The ensuing simulation results unequivocally demonstrate the superior efficacy of the proposed method in achieving enhanced target recognition rates.

Keywords

• Complex Dempster–Shafer evidence theory
• Complex basic belief assignment (CBBA)
• Attribute weights
• Triangular fuzzy number model
• Target recognition