Generalized Similarity Measure for Multisensor Information Fusion via Dempster-Shafer Evidence Theory

Dempster-Shafer evidence theory (DSET) stands out as a mathematical model for handling imperfect data, garnering significant interest across various domains. However, a notable limitation of DSET is Dempster's rule, which can lead to counterintuitive outcomes in cases of highly conflicting evid...

Full description

Saved in:
Bibliographic Details
Published in:IEEE access 2024, Vol.12, p.104629-104642
Main Authors: Liu, Zhe, Hezam, Ibrahim M., Letchmunan, Sukumar, Qiu, Haoye, Alshamrani, Ahmad M.
Format: Article
Language:English
Subjects:
Accuracy
Area measurement
belief logarithmic similarity
Data integration
Dempster-Shafer evidence theory
Evidence theory
information fusion
Mathematical models
Measurement uncertainty
Multisensor fusion
Pattern classification
Pattern recognition
Power line communications
Similarity
Target recognition
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Dempster-Shafer evidence theory (DSET) stands out as a mathematical model for handling imperfect data, garnering significant interest across various domains. However, a notable limitation of DSET is Dempster's rule, which can lead to counterintuitive outcomes in cases of highly conflicting evidence. To mitigate this issue, this paper introduces a novel reinforced belief logarithmic similarity measure ( \mathcal {RBLSM} ), which assesses discrepancies between the evidences by incorporating both belief and plausibility functions. \mathcal {RBLSM} exhibits several intriguing properties including boundedness, symmetry, and non-degeneracy, making it a robust tool for analysis. Furthermore, we develop a new multisensor information fusion method based on \mathcal {RBLSM} . The proposed method uniquely integrates credibility weight and information volume weight, offering a more comprehensive reflection the reliability of each evidence. The effectiveness and practicality of the proposed \mathcal {RBLSM} -based fusion method are demonstrated through its applications in target recognition and pattern classification scenarios.
ISSN:2169-3536
2169-3536
DOI:10.1109/ACCESS.2024.3435459