Atrás

RMIT University in Melbourne and LeadMind join forces on the road to excellence in Data Science

RMIT University students have participated in a Data Science collaborative project with CAF´s Digital Platform: LeadMind; in order to develop a classifier to identify unknown Multifunction Vehicle Bus (MVB) communication signals from trains.

Collaboration and being open to partnership has always been linked to CAF values. This commitment could not have been better in this occasion: RMIT University students have participated in a Data Science collaborative project with CAF´s Digital Platform: LeadMind.

RMIT is an innovative University based in Melbourne, Australia. As a technology research and teaching institution, RMIT encourages students to collaborate with industry experts through a specialized business collaboration program. This time RMIT students have had the opportunity to work tightly with CAF, a global railway industry supplier of rolling stock and digital services. 

CAF’s experts have challenged RMIT students to develop a classifier to identify unknown Multifunction Vehicle Bus (MVB) communication signals from trains in order to enhance LeadMind’s capabilities. The selected dataset in question comprises time-based signals from Heating, Ventilation and Air Conditioning system (HVAC) of a single train. Students developed an Exploratory Data Analysis (EDA) reporting interesting relationships between different signals and how they evolve over time.

Based on this information, they were able to train a classification Machine Learning (ML) model that could identify 20 HVAC signals from a set of 100 from other systems in the train with a >90% accuracy performance.

According to Eugenio Marinetto (Artificial Intelligence Lead at CAF Digital Services), “This is a very promising and valuable result. RMIT student results give us the opportunity to validate our hypothesis that intrinsic properties of train signals could contain enough information to identify each of the signals that we collect from the train systems. This knowledge can be transferred to the developing of ML-based tools for LeadMind’s improvements which means a better service for our clients

The experience on both sides has been completely satisfactory. Students have been able to experience a real world ML challenge while CAF has collected precious information for the improvement of LeadMind features.

THE EXPERIENCE by Maree Carroll, Student at RMIT

What new learning has given you this opportunity?

"The CAF Digital Services MVB classification project was a highlight of my studies, giving me a chance to experience a real world machine learning problem that I can see has potential in the LeadMind platform for predictive maintenance and efficient operation of rolling stock. This experience gives me confidence to pursue machine learning and data science interests to progress my career." 

Would you recommend other students to have this experience?

“I would highly recommend other students seek out this experience. Exposure to industry projects, overcoming the challenges of real data and finding solutions that meet the needs of an organisation cannot be replicated easily in academic projects. The application of what has been learned in coursework, in a genuine setting, with great support, is paramount to assessing how industry ready a student is, and this project was an excellent challenge to test our skills. I thank CAF Digital Services for the opportunity.”

 Amir Homayoon, Lecturer and Coordinator of the program at RMIT says “Its real-world experience, projects are at the edge of technology and industry partners are professional and supportive. All students like such an experience”.