Smart modelling leaves European gridlock in the rear-view mirror
Europe’s urban populations continue to grow, so the road infrastructure becomes increasingly congested, negatively impacting the economy, public health and the environment. One study found that inefficiencies in urban mobility – road congestion in particular – cost the EU an estimated EUR 110 billion per year, more than 1 % of its GDP. To mitigate this, the CONDUCTOR project has been developing an advanced transport management system to help authorities and operators such as public bus networks move people and goods more efficiently, making use of the growing availability of automated and connected vehicle capabilities. “Driven by next-generation simulation tools, we are developing a prototype that dynamically assesses and prioritises the transport needs of various users, to optimise traffic flows,” explains project coordinator Flavien Massi, senior project manager at Netcompany-Intrasoft in Luxembourg.
Next-generation transport
At the heart of CONDUCTOR’s solution lies a customised platform created by merging various pre-existing traffic and fleet management technologies, with computer models developed by the project. These optimisation, simulation and prediction models are applicable to a range of real-life scenarios and are driven by algorithms trained on a variety of real-world traffic data using machine learning. These can also be used to evaluate cooperative routing strategies for large-scale connected, cooperative and automated mobility (CCAM) vehicle fleets. To ensure the solution’s compatibility with others already deployed across European networks, APIs have been developed for exchanging critical data including General Transit Feed Specifications (such as transport routes, timings, stops, and locations) alongside information related to historical performance such as travel times and speeds, and incident management. A portfolio of solutions is being deployed and tested across a range of use cases and pilot sites. These simulations model conventional vehicle traffic alongside that of CCAM on existing road infrastructure. The first use case focuses on integrated traffic management and comprises three pilots. In Athens, bus, metro and tram schedules are synchronised to maximise passenger options. In Madrid, a model is being developed to better manage unexpected transport events and ensure efficient network recovery afterwards. While in Almelo in the Netherlands, a logistics corridor is being optimised by installing traffic prioritisation at designated intersections. The second use case is improving the transport platform of project partner GoOpti, which uses cooperative routing (identifying transport patterns to maximise fleet resources) to shuttle passengers to and from airports. The third use case, also in Madrid, demonstrates how traffic swelled by surging demand for last-minute deliveries can be reduced by transporting passengers and goods in the same vehicles. Supply and demand is calculated using the FleetPy simulation environment, enriched with the project’s transport control strategies and linked to Aimsun.next software which provides a realistic representation of traffic conditions. “While we only recently started the validation phase, the initial results are extremely promising. Several models have already demonstrated improved traffic flow, shortened journey times and/or reduced emissions,” says Massi.
Future-proofing
CONDUCTOR contributes to the Europe-wide CCAM initiative, set up to create a more user-centred and inclusive mobility system increasing road safety while reducing congestion and environmental footprint. To advance the technology, the machine learning tools used by the solution’s various models are now being tweaked to refine the accuracy of modelling predictions of key metrics, such as traffic patterns, congestion hotspots and energy consumption. While the team hasn’t faced many technical challenges so far, privacy and security legislation has sometimes made it difficult to access real-time vehicle data necessary for the simulations. Validation tests are currently under way to ultimately ensure a robust system capable of seamlessly connecting with the transport operators of any European city.
Keywords
CONDUCTOR, connected, cooperative and automated mobility CCAM, congestion, traffic, transport, bus, road, goods, simulation, machine learning
