Research on data fusion of multi-sensors based on fuzzy preference relations

• Published in: Neural computing & applications Vol. 31; no. Suppl 1; pp. 337 - 346
• Main Authors: Hao, Huijuan, Wang, Maoli, Tang, Yongwei, Li, Qingdang
• Format: Journal Article
• Language: English
• Published: London Springer London 01.01.2019; Springer Nature B.V
• Subjects: Accuracy; Adaptive algorithms; Algorithms; Artificial Intelligence; Computational Biology/Bioinformatics; Computational Science and Engineering; Computer Science; Computer simulation; Correlation analysis; Data integration; Data Mining and Knowledge Discovery; Fault diagnosis; Image Processing and Computer Vision; Machine Learning Applications for Self-Organized Wireless Networks; Multisensor fusion; Parameter estimation; Probability and Statistics in Computer Science; Sensors; Vibration analysis; Weight; Data fusion; Multi-sensors signals; Fuzzy preference relations; Adaptive weighted estimate algorithm; Weight
• ISSN: 0941-0643; 1433-3058
• DOI: 10.1007/s00521-018-3778-5

For the data fusion of multi-sensors, the determination of weight directly affects the accuracy and performance of the fusion algorithm. In order to improve the accuracy of fusion algorithm, an adaptive weighted algorithm based on fuzzy preference relations is proposed. The degree of preference between signals is represented by introducing the improved logsig function, and then, the weight is calculated by fuzzy preference relations. Simulation results show that the proposed algorithm is significantly better than the mean value method, and the accuracy is basically equivalent to the method based on correlation function. The analysis of the actual vibration signals in axis system verifies the validity of the algorithm in the practical application. The algorithm in this paper has good dynamic performance and is easy to be implemented. It can be applied to the actual multi-vibration signal estimation to provide more accurate parameters for the next step of fault diagnosis.