Ansik Inc.

Published purpose

Pitstop’s mission is to obtain predictive maintenance insights from vehicle data and service data. Some drawbacks with machine learning approaches have been that data sets are not always complete. The simulation approach taken in this project will improve this situation and increase the benefit of a machine learning approach. In this project the plan is to train algorithms where there are common cases of service records that correlate to time-series sensor streams across different kinds of components (ex. engine and battery) and let the machine learning take care of small variations. This is an important step in our plan to apply neural network approaches to build proprietary predictions models on vehicle wear and failures.