Imperial experts work with Cubic to explore AI-powered public transport

Imperial experts are working with Cubic Transportation Systems, a global leader in public transport innovation, to explore new applications of AI.

These include tackling fare evasion, which costs transport authorities in cities such as New York and London hundreds of millions of dollars each year, and predictive maintenance, which allows providers to anticipate mechanical problems and optimise maintenance schedules.

Professor Danilo Mandic and members of his team in Imperial’s Department of Electrical and Electronic Engineering are working independently with Cubic through Imperial Consultants to help the company launch an innovation centre at its Surrey headquarters to progress applications such as these.

Irregular domains

Professor Mandic’s lab specialises in using AI techniques to analyse so-called irregular domains such as the London Underground network.

Irregular domains are systems that consist of multiple entities, called nodes, that stand in inconsistent relationships to one another. In the Underground network, for example, some stations are geographically closer to their neighbours than others, while some are better connected – and during the morning rush hour, some stations are more popular as boarding points and others as destinations.

Because no consistent rule can describe the relationship between any two individual stations, the team’s models make extensive use of graph theory, a technique for representing individual relationships between pairs of nodes. These can be used to discover anomalies, such as fare evasion or mechanical failures in station gates, that may not be detectable through individual data points or aggregated trends alone.

Given the data-hungry nature of machine learning models, the team’s work with Cubic will also explore the use of generative AI for creating synthetic data to help train the models.

Tube station of the future

The new centre, set to open in September 2025, aims to support the development of AI technologies based on innovative approaches such as these by convening experts from the public and private sectors, and will feature an immersive ‘tube station of the future’ to demonstrate digital technologies. Building on the collaboration with Professor Mandic and his lab, Cubic also plans to offer internship opportunities to Imperial students.

“This collaboration is an opportunity for us to develop the transportation experts of the future,” said Professor Mandic. “Our students and researchers will be working closely with Cubic to harness the latest innovations in technology to solve transportation’s most significant challenges. We look forward to working together to develop new ideas, technologies and industry leaders.”

“By developing a centre that bridges industries, academic partners and government agencies, our aim is to discover better solutions for our customers and the wider public transit ecosystem, which will contribute to increased ridership and job opportunities,” said Peter Montgomery-Torrellas, president of Cubic Transportation Systems. “Our long-term vision is to replicate this model in key markets around the world, expanding our industry-leading partnerships and ensuring innovation is at the core of everything we do.”