The Researchers in Residence programme develops new collaborations through research residencies for university academics embedded within Connected Places Catapult’s teams to build stronger partnerships in the national innovation system and supporting the best environment for innovation.
Mauro Vallati is a UKRI Future Leaders Fellow and a Reader at the School of Computing and Engineering of the University of Huddersfield. He joined the University of Huddersfield, in October 2012 where he served as a Research Fellow till April 2016, and as a Senior Lecturer till July 2020. He is also a member of the PARK (Planning, Autonomy and Representation of Knowledge) research centre.
His main research interests are in Artificial Intelligence Planning and Argumentation. In particular, he is interested in the application of AI Planning techniques to Urban Traffic and Mobility (this is also the topic of his Fellowship). More generally, his work encompasses domain-independent planning, knowledge engineering, SATisfiability, machine learning, abstract argumentation, portfolio approaches, heuristic search, genetic algorithms, and automatic algorithm configuration. As well as this, his interests lie in the fields of innovative applications of AI in Medicine.
In the UK alone, the cost of congestion has reached nearly £8 billion in 2018 in lost time and fuel consumption and has become a major health threat. The arrival of Connected Autonomous Vehicles (CAVs) presents a unique opportunity for a fundamental change in urban mobility and urban traffic control. CAVs can communicate with other vehicles and with the infrastructure, via dedicated protocols, to take better-informed decisions.
AI4ME focuses on the use of Artificial Intelligence approaches, specifically deliberate AI planning, for maximising the exploitation of urban transport networks by generating real-time personalised routes for Autonomous Vehicles which are navigating or approaching the regional urban area controlled by the system. By increasing the exploitation of the urban network, AI4ME addresses the urban challenges of facilitating economic growth (by reducing traffic congestion and travel time unreliability) and improving health (by improving air quality).
The project uses novel applications of AI techniques to generate real-time personalised routes for Autonomous Vehicles which are navigating or approaching a controlled regional urban area.
It uses knowledge of the network and operational goals (such as improving safety, improving the environment, minimising delay, and reducing congestion) to optimise urban traffic control management.
Personalised routes are generated by taking into account traffic, incidents and other relevant inputs. These are balanced against the preferences of the AV and its occupants (for instance, are you in a hurry? Do you need to arrive somewhere at a specific time?) and the other operational goals set by the traffic controller. The result will ensure smooth-flowing traffic, reduced delays, and reduced investment in increasing capacity.
Our Researchers in Residence (RIR) activity was part of our Academic Engagement Programme where Mauro took up a residency with CPC between March 2018 and February 2021. The RIR programme develops new collaborations through research visits/residencies for university academics to spend time embedded within the Catapult teams to build stronger partnerships in the national innovation system and supporting the best environment for innovation, such as autonomous vehicles solutions.
Benefits & Results
The project consisted of 4 main work packages:
- Analysis of the state of the art and the research context.
- Identification of data sources and suitable scenarios.
- Design and development of Artificial Intelligence-based approaches to address the considered problem.
- Testing and Validation on the scenarios identified at 2.
The work was carried out in collaboration with Dr Zeyn Saigol, Principal Technologist, who supported all the stages of the project and suggested the Catapult teams to approach for dealing with issues and questions.
Future collaborations with Connected Places Catapult are being explored, particularly for networking and advertising opportunities, to disseminate the results of the project, and to connect with traffic authorities.
As a tangible output of the AI4ME project, 6 publications have been produced.
- A Principled Analysis of the Interrelation Between Vehicular Communication and Reasoning Capabilities of Autonomous Vehicles, In Proceedings of The 21st IEEE International Conference on Intelligent Transportation Systems
- Exploiting Automated Planning for Efficient Centralized Vehicle Routing and Mitigating Congestion in Urban Road Networks, In Proceedings of The 34th ACM/SIGAPP Symposium on Applied Computing
- How to Plan Roadworks in Urban Regions? A Principled Approach Based on AI Planning. In Proceedings of the 19th International Conference on Computational Science
- Reducing Traffic Congestion in Urban Areas via Real-Time Re-Routing: A Simulation Study, In Proceedings of 33rd Australasian Joint Conference on Artificial Intelligence
- Centralised versus Decentralised Traffic Optimisation of Urban Road Networks: A Simulation Study, In Proceedings of IEEE International Conference on Intelligent Transportation Engineering
- Effective Real-Time Urban Traffic Routing: An Automated Planning Approach, In Proceedings of the 7th International IEEE Conference on Models and Technologies for Intelligent Transportation Systems
Furthermore, the project has been described in two tutorials delivered to top-ranked conferences for the AI discipline.
- Planning and Scheduling Approaches for Urban Traffic Control. The 29th International Conference on Automated Planning and Scheduling
- Planning and Scheduling Approaches for Urban Traffic Control. The 33rd AAAI Conference on Artificial Intelligence
Finally, the AI4ME project also led to the organisation of a workshop focused on the use of AI techniques for urban mobility. The workshop was organised in collaboration with experts from the Carnegie Mellon University, the University of Toronto, and the Czech Technical University.
The PI and Simplifai Systems Ltd are actively engaging in enhancing the techniques developed during the AI4ME project for deployment. Discussions are ongoing with the Kirklees council traffic authority and Transport for Greater Manchester to identify suitable areas for field testing.
The AI4ME project allowed the PI to gain insights and expertise that lead to the award of a UKRI Future Leaders Fellowship on the topic of Artificial Intelligence applications to traffic control and management (https://www.ai4utmc.info/home). The ambitious aim of the fellowship is to lay the foundations of a fully autonomic AI-based system for urban traffic control.
The Future Leaders Fellowships support talented people in universities, businesses, and other research and innovation environments. They also allow universities and businesses to develop their most talented early career researchers and innovators or to attract new people to their organisations.
The Fellowship, from UK Research and Innovation (UKRI) will enable Dr Vallati to further his work on providing a solution to an issue that affects the UK’s economy and the health of the nation, in partnership with Transport for Greater Manchester, Kirklees Council, and SimplifAI Systems.
The system will be designed to effectively manage congestion in specific areas by altering existing traffic light sequences and to communicate with vehicles to suggest that they drop speed, change routes to avoid congested areas or switch to electric power.
Keith McCabe, CEO of Simplifai Systems limited demonstrated interest in the outputs of the AI4ME project, and is exploring routes to increase the technology readiness level for commercialisation.
ICAV Cluster is Smart City and CAV news aggregator. To add your Smart City or CAV related news, please email our editorial team on firstname.lastname@example.org