Artificial Intelligence for Autonomic Urban Traffic Control

Over half of the world's population now lives in cities, and global urbanisation continues at a steady pace. As this trend continues, mobility is becoming an increasingly critical problem. In the UK alone, the cost of congestion reached nearly £8 billion in 2018 in lost time and fuel consumption, and has become a major health threat.

At the same time, new modes of transport, such as Connected Autonomous Vehicles, and new business models (e.g., Mobility as a Service) are disrupting the transportation sector. The traffic control industry has to reinvent itself to operate in a world of decentralised control (vehicles making decisions), ubiquitous sensor information, increasing urbanisation pressure, and large-scale interconnectivity.

Artificial Intelligence provides a range of approaches that can leverage the growing volume of available data, and the knowledge gained by traffic authorities in the past decades, to support urban mobility. In particular, the proposed line of research aims at designing and creating an autonomic urban traffic management and control framework. The autonomic framework will have the capability to self-manage and self-configure, and will have an holistic view of the condition of the controlled region to proactively act to prevent environmental and mobility issues, or react to mitigate observed problems.

Impact on Economy:
This Fellowship has the potential to significantly impact the UK economy: the direct cost of traffic congestion is estimated to be approximately £8B per year for the UK alone, due to fuel consumption and time lost. Furthermore, traffic congestion is also leading to an increasing pressure on the NHS, due to health issues - particularly in large cities.
The proposed autonomic framework can also provide the ideal platform to support the design of innovative techniques for managing traffic in urban regions; for instance, innovative traffic control approaches can be designed to exploit the expected transition to autonomous vehicles. This can provide a further boost to companies working on either traffic control systems or vehicles.
Further, the autonomic framework can provide a means for companies to better manage their commercial fleets, to optimise their use of the network to minimise journey times. Similarly, companies supporting the mobility-as-a-service (Maas) approach could benefit from the outcomes of this Fellowship: the Maas market is expected to be worth more than $106B globally by 2030.

Impact on Society:
Impact on Society is a key aspect of this Fellowship: improving urban traffic management has a direct beneficial impact on people's quality of life. The autonomic framework will enhance quality of life of citizens of UK urban areas by reducing the time wasted in vehicles (particularly while commuting) and improving local environmental conditions. The former would be reflected in more time available to people, that can be dedicated to families, hobbies, or other pursuits; the latter aspect would lead to better health conditions.
This Fellowship has also the potential to provide valuable input to Department for Transport policies: the use and study of the autonomic framework can help to identify future needs of the urban mobility field. Further, this Fellowship can support traffic management and control experts in understanding the potential of AI technologies, and in supporting the design of future AI-based approaches.

Impact on Knowledge:
This work is well-aligned with EPSRC cross-ICT priority areas of Future Intelligent Technologies, NERC Innovation activity on Environmental Data, and with the ESRC research priority on Climate Change.
This Fellowship will highlight the importance of AI approaches based on the combination of data-driven and model-based techniques. Further, it will lay the foundations of how to merge the two paradigms in order to obtain efficient systems, that can justify and explain their behaviour to human users.