Information fusion diagnosis research based on fiber CAN communication in vehicle
Weirong Jiang, Haibo Huang, Tao Zuo · 2010
In order to transmit high-speed and large-capacity data with powerful ability of anti- electromagnetic interference in modern automobile, data communication and fault diagnosis platform based on fiber CAN communication is built up, from which every sensor in accessory nodes achieves at gathering fault signal and processing fault real time. Combine neural network with evidence theory organically, process the elementary diagnosis output of neural network as the base belief assignment value of evidence theory, and fuse with evidence theory to obtain the final diagnosis result. Fault diagnosis experiment on automatic transmission shows the proposed fusing diagnosis method can fully make use of redundancy and complement information of fault data, the veracity and reliability of fusing diagnosis result with neural network and evidence theory is much better than that of only neural network.