【PATENT NUMBER】7356773
【APPLICATION NUMBER】2023-111983
【TITLE OF THE INVENTION】Abnormality detection system and method
【REGISTRATION DATE】September 27th, 2023
Objective
The detection of abnormalities in the systems of increasingly electronic and complex vehicles (automated vehicles and EVs, hybrids, and plug-in hybrids) is based on an information flow algorithm, which enables early detection, localization, and visualization of parts anomalies by calculating the transfer function and convolution of the anomaly.
Overview
The Multi-Level Flow Model method is used to identify the location of electrical anomalies by mathematical calculation using information from sensors as input. For single circuit loops, the state of each part can be calculated using transfer functions, while for multiple circuits, several sub-loops are joined to the main loop and the state of each part is estimated using a convolution method.
The estimated state matrix for each part is the following equation.
If Qi=matrix P of sensor values, then the parts of the anomaly is identified.
Applications
This research is not limited to electric vehicles but is also applicable to anomaly detection in various critical systems such as electric power plant and production line. We anticipate that this patent, now part of our company’s mobility solutions, will be instrumental in future developments. The experts at FUTU-RE are committed to evolving continuously to realize a future of safe and comfortable mobility!
For more information, please visit the Japan Platform For Patent Information website.
English 
Japanese