Can artificial intelligence (AI) help with gridlocked traffic? While smart traffic management systems exist already, they typically work on a fairly localised basis. If the power of AI was applied across an entire city or region, could it bring another level of improved decision-making?
A tie-up between the UK’s Alan Turing Institute and the Toyota Mobility Foundation is intended to explore exactly this. Dubbed ‘Optimising flow within mobility systems with AI’, their collaboration is part of the Turing Institute’s new AI programme. Its aim is to shift complex traffic management from static systems to “dynamic, optimised systems, managed in real-time across many different types of mobility”.
The intention is to look at the entire breadth of a system, says William Chernicoff, senior manager, global research and innovation, at Toyota Mobility Foundation. As a change occurs in one location, how does that propagate across everything? Current systems, he feels, give some degree of intelligence at a single intersection but not how this has an impact three roads over, on cross-flow traffic or at five junctions ahead.
AI might be layered on top of existing, more advanced traffic management systems (such as that of Utrecht – https://smartercommunities.media/utrechts-radical-response-congestion-downgrading-inner-ring-road/) or as a replacement, says Chernicoff. It is intended to harness the machine-learning capabilities of AI so to improve and adapt over time. For one thing, new developments will change the dynamics and flows, including on-demand and shared mobility, autonomous vehicles, new regulations such as congestion charges, new forms of e-commerce delivery and shifting demographics.
“It is not about completely handing over control to the AI system, people will still have to be in charge and set the priorities,” says Chernicoff. “But AI can do simultaneous calculations that no humans can do.”
This is an 18-month project, at the end of which the intention is to move from testing and prototyping to actual use. Chernicoff cites the 2020 Tokyo Olympics as a possible focus for applying the technology. This isn’t about doing research for the sake of research, he says, nor about Toyota driving the development and sale of new products, but “about empowering and bringing capabilities to governments that serve the community”.
The Toyota Mobility Foundation was set up as a not-for-profit entity in August 2014 with the stated aim to “support the development of a better mobile society”. The Foundation aims to support strong mobility systems while eliminating disparities in mobility.
In its initial dozen or so projects it is clear the Foundation is taking a broad definition of mobility. Projects include developing devices for lower limb paralysis; a competition with mayors in Brazil to find solutions for this country’s congestion; work to improve “first and last mile” access to the metro in Bangalore; and an app-based solution to reduce food waste in supply chains in developing countries.
The Alan Turing Institute was set up in 2015 and is the UK’s national institute for data science and AI. The Institute and partner universities are working with the Greater London Authority on AI for transport and the tie-up with Toyota will build on this. It is the first time the two entities have worked together.