TransiT Engagement Findings:

In conjunction with the consultation, the TransiT team conducted a survey of academic and grey literature on digital twins in the transport sector to better understand what has been already published.  The team also undertook an investment landscape review to understand what UKRI-funded projects are still underway related to digital twins, transport and decarbonisation.  Additionally, TransiT engaged CENSIS (Scotland’s Innovation Centre for sensing, imaging and Internet of Things (IoT) technologies) to do a more general technology landscape review, examining the role, maturity, capability and future opportunities for digital twins in the transport sector.  These three reviews  (i.e., literature, investment and technology landscape reviews) mapped a clear picture of the challenges, gaps and barriers to providing a transport digital twin for decarbonisation applicable across the UK to inform further research.

The challenges, gaps and barriers revolved around the following:

• Many different types of digital tools, i.e., digital models, digital shadows, digital twins, exist.
• Most research papers have made assertions about their work as being digital twins, whereas, in the context of the definitions provided above, the majority of the projects would fall under the digital model or digital shadow categories.
• There are limited applications of DTs in the transport sector and very few with a particular focus on decarbonisation
• There are no cross-modal projects with a focus on using digital capabilities for transport decarbonisation – most projects are at siloed digital model and digital shadow levels.
• No autonomous digital twins were found to exist.
• There is a greater level of activity in road transport decarbonisation, mostly models and shadows. Aviation and Maritime sectors are advanced in terms of digital maturity but lacking in decarbonisation efforts
• Digital twins predominantly focus on operational optimising with an incidental benefit of decarbonisation, not the main intent.
• Only 7 projects of 158 ongoing projects on transport decarbonisation have a focus on digital twinning, many have the potential to inform digital twinning.

The CENSIS report reviewed the requirements for digital twinning and establish the research gaps as evident from the implementation of digital twins.  The report identified and described the three main components, hardware, data management middleware and software used in digital twins.  Challenges in establishing each of these core components were outlined including reliability of sensor hardware over time, the choice of platforms required to handle the vast quantities of data generated by transport IoT and the development of data analysis techniques and visualisation methods used.

The report identified a number of research agenda topics that spanned socio-technical, cyber security and federation challenges. Of the research topics mentioned, the following issues have cross-over with the concerns of the stakeholders:

• how human behaviour can be captured in data models
• how change management can support the digital twin transformation
• help organisations define problems they could address with digital twins
• mapping of the skills required
• methods to incentivise data sharing
• mapping of standards and regulations
• cyber-physical security and resilience
• assessing and preserving the veracity of data
• integrating and leveraging heterogeneous communication sources
• how the integration and leveraging of heterogeneous communication sources through digital twin technology can enhance real-time decision-making
• methods to enhance data quality and accuracy
• robust data security methods
• Develop standardised data formats and protocols

Additionally, there is a clear technical gap in validation of success, responsibility across multi-domain cyber-physical systems, the protocol of dealing with emergency events, templates for systems and network architecture allowing for technical development, distributed active defence functions, integrity of outputs over time, addressing sensor reliability through AI and ML optimising.