Skip to Content
Report an accessibility problem
Center for Efficient Vehicles and Sustainable Transportation Systems

PI: Hwan-Sik Yoon, University of Alabama.

Project Description:

As conventional OBD systems can provide information only on impending or occurring engine failure, a new approach is desirable to change the Fault Diagnosis and Isolation (FDI) paradigm by predicting what will happen to the system before anything occurs. By taking proactive actions based on predictive diagnostics, customers can avoid or minimize system downtime as well as catastrophic system failure initiated by a minor component failure. The goal of this research is to develop a new predictive diagnostics methodology by applying machine learning algorithms to massive data called big data produced by a large number of engines in order to extract meaningful information to make a predictive decision on the system failure. Machine learning algorithms categorized by unsupervised and supervised learning can find natural patterns hidden in massive data and help make better decisions and predictions.