Big Telemetry Data Analytics for Vehicle Fault Detection

Abstract

Monitoring the correctness of vehicles in modern traffic is very difficult, considering that the condition of a vehicle is constantly changing due to driving and requires constant checking of various parameters. Systems that would deal with this would have to be accurate in their estimates and very fast given that vehicles constantly generate a large amount of data. A Vehicle Fault Detection system has been proposed as a solution to this problem. The system is distributed and autonomous and uses linear regression as a classifier. The proposed solution was implemented and tested and detailed performance and accuracy analysis were performed.