Fujitsu Launches ‘Digital Twin’ Trial on Isle of Wight for e-Scooter Sharing Service


Fujitsu today announced the development of a new ‘digital rehearsal’ technology that can help better inform public policy and business planning. Fujitsu’s first demonstration of the technology provides realistic simulations of the effects of traffic measures by reproducing people’s movements on a digital twin. Fujitsu began trials of the new technology on April 1, in cooperation with Beryl, a shared mobility company based in the UK. Together they will demonstrate the business and societal value of the technology, by improving the operation of shared e-scooter services on the Isle of Wight.

The new technology, which combines the behavioral economics model Prospect Theory and AI, allows for simulations that can infer the behavior of people in the real world, reproducing not only human biases such as our tendency to overestimate losses and underestimate potential gains and situational factors that influence behavior such as weather. By combining these models with digital twins – digital reproductions of physical objects and entities, sometimes entire cities – the new technology makes it possible for city planners and businesses to more accurately predict how changes in human behavior interact with evolving conditions in the environment to better inform decision-making.

The trials on the Isle of Wight, use digital rehearsal technology to test in advance the effects of people switching from cars to e-scooters. Fujitsu will further estimate how the usage of e-scooters instead of cars will affect CO2 emissions in local areas, and in how far measures including discounted fees for users who return e-scooters to specific places will affect people’s behavior regarding their choice of transport. The ultimate aim is to bring business benefit to Beryl, reduce the damaging environmental and social effects of car use, inform transport policy on the Isle and positively contribute to the Isle of Wight’s wider economy.

The trial is part of wider initiatives which Fujitsu is taking as a Lead Technical Partner for the UK’s National Digital Twin Programme(2) with the Department for Business and Trade, which aims to develop techniques to use connected Digital Twin models to benefit society, the economy, business and the environment, supported by HM Treasury. The Programme is running a demonstrator in Isle of Wight which is a fundamental element of the socio-technical change aspects of the Programme.

Moving forward, Fujitsu plans to leverage the results of this project to support the sustainability transformation (SX) of mobility service providers and contribute to the realization of a sustainable, fair, and diverse society through converging technologies that combine computer sciences with knowledge from the humanities and social sciences.

Fujitsu will showcase this technology at the Fujitsu ActivateNow Tech Summit held in Madrid, Spain on April 20, 2023 and the G7 Digital and Technology Ministers’ Meeting held in Takasaki, Gunma Prefecture, Japan from April 28, to April 30, 2023.

Simulations based on models that are closer to the real world

In 2021, Fujitsu initiated intensive R&D into Social Digital Twin(3), a promising field that merges technologies like AI and big data analytics with models from the social sciences, to address some of the increasingly complex challenges facing humanity.

To achieve more accurate reproductions of human behavior and ultimately realize more appropriate solutions to various societal issues, Fujitsu developed a digital rehearsal technology combining big data analysis using AI with knowledge from behavioral economics.

Where many behavioral models are based on expected utility models, assuming people always take the decision that brings them greatest immediate benefit, Fujitsu developed a model for human behavior based on Kahneman and Tversky’s prospect theory in behavioral economics. Kahneman and Tversky showed that individuals tend to overestimate potential losses while underestimating potential gains, Fujitsu aims to show that by weighting agents decisions according to prospect theory, and training an AI model with characteristics of indirect factors that influence human behavior such as weather conditions, it can produce more accurate simulations than conventional models.

Combining this model with a digital twin, the new technology enables pre-verification of various scenarios based on human psychology such as the effect of crowds of people moving to surrounding transportation systems after an event, or the change of people’s behavior according to traffic situations after an accident. The field trial will aim to validate and verify this approach under real-world conditions.