Roads Research Alliance (RRA)

Digital twin systems: Junxiang Zhu

There are many parts of the digital roads platform including the autonomous vehicles and other elements such as sensors, but for my part it is about the integration of the information. 

The initial digital twin is the big framework. There is information input, analysis, management and output so we need the static data. That includes the low side data and environmental data. We need to combine information and data to create the initial digital twin. We also need sensors to make the digital twin real time. 

We also have artificial intelligence (AI) and high performance computing (HPC). AI makes the system smarter while HPC makes the system stronger. Lastly, we have the information output module so the screens, VR and 3D printing. 

My focus is the information part and how to put all the kinds of information together with modelling and geospacial data. I work on how geometry can be read, interpreted, converted and reused. Essentially, I am building models.

Also, in a computer-based system, we also need to use semantic information as well as geometric information.  I am also testing a new form of implementation for growth. We are seeing if this new information can improve the performance of digital insistence for roads, buildings and other assets. 

My fellowship is about roads and digital training roads. What I am focussing on is the digital twin system itself – specifically on the data side so how this data can be integrated in a more flexible and efficient way for information, knowledge, discovery and better decision making. 

Cycling environments and infrastructure: Khashayar Kazemzadeh

I joined the University of Cambridge in January 2024 as a Marie Skłodowska-Curie Fellow, and continued as a Research Collaborator until March 2026. My research focuses on how we can design transport systems and infrastructure that are more inclusive and appealing, particularly for underrepresented groups, to encourage greater use of active modes such as cycling.

This is also a personal passion of mine. In many cities, car use is not only costly for individuals but also has broader impacts on public health and the environment. By improving the quality and inclusivity of infrastructure, we can make cycling a more realistic and comfortable option for a wider range of people.

How has the Roads Research Alliance helped pave the way for your future in the industry?

The Roads Research Alliance provided valuable opportunities to engage with industry partners and better understand real-world challenges. These collaborations helped me connect academic research with practical applications, particularly in terms of how transport solutions are designed and implemented.

The programme also supported my academic development, including preparing for my current role at The University of Manchester. I continue to build on this experience through my research and teaching. For example, I am currently involved in a project on adaptive cycling in Manchester, which supports individuals with physical and cognitive disabilities to access cycling. This work draws directly on the knowledge and experience I developed during my fellowship.

What were your objectives when you began at Cambridge University?

My previous research in Sweden focused on improving cycling conditions and understanding how to design infrastructure that feels safe and comfortable for all users. At Cambridge, my objective was to expand this work at a larger scale, particularly in the UK context, by studying how people perceive and experience cycling environments.

Initially, I explored naturalistic approaches, including field observations and video recordings in cities such as London and Cambridge. This later evolved into the use of digital twin models, developed in collaboration with TRL, which allowed us to test different scenarios in a controlled and systematic way.

The aim of this work is to provide evidence that can support planners and policymakers in making cities more inclusive, liveable, and bicycle-friendly.

Digital twin modelling: Shirin Malihi

My background is in engineering, with focus on photogrammetry, 3D computer vision, remote sensing, spatial data science, and their applications in surveying and civil engineering.

I am currently undertaking a MSCA fellowship focused on enhancing digital twin models of road infrastructure, with a particular emphasis on the geometry. In this context, accurate digital representation of road geometrics is fundamental, because geometry forms the base layer upon which other  information is built.

A key part of my research is technical  innovation, while another important part involves working with industrial partners to identify and solve real-world infrastructure problems. Roads are a critical component of national infrastructure, and their digital models must be not only accurate, but also reliable.

My work focuses on uncertainty analysis, quality control of digital road geometry. I investigate whether the geometric information used in digital twin models can be trusted before it is integrated into final products or decision-making systems. This involves analysing different types of data, particularly remote sensing data, laser scanning, imagery, and AI-derived outputs.

In general, my role is to verify the quality, consistency, and reliability of spatial knowledge before it is presented or used in downstream applications. One of the most important data sources for digital twin models is laser scanning, while another is imagery captured using professional cameras. These data sources are closely related to photogrammetry, surveying engineering, and 3D computer vision.

My research involves big data analysing, classifying, and validating spatial data, especially in challenging edge cases where existing AI methods are not fully reliable. This is particularly important in aerial data, where boundaries, object edges, and road features may not always be accurately detected or reconstructed. I am working on methods to identify areas where the data or AI-generated results may be unreliable.

A major focus of my work is the detection of problematic or untrustworthy regions in 3D data and imagery. These are cases where current AI or computer vision methods may fail to produce results with sufficient accuracy or confidence. In the next stages of my research, I aim to address these limitations by incorporating additional data modalities and improving the robustness of the analysis.

One specific phenomenon I am investigating is known as the ghost effect. This can occur when there is a mismatch between laser scanning data, image data, and the real-world position of objects. For example, moving vehicles may appear in different positions across datasets because of time delays between laser measurements and image capture. This creates inconsistencies in the digital representation of the road asset. 

My research aims to detect, understand, and eliminate these types of anomalies so that the final digital road geometry is accurate, reliable, and suitable for use in digital twin applications. Overall, my work contributes to improving the quality control of road infrastructure knowledge through photogrammetry engineering, remote sensing, AI, big data analysis and 3D computer vision.

Autonomous driving: Yue Xie

Former Fellow Yue Xie started a lecturing role in robotics and automation at Loughborough University and continues to seek out research opportunities. 

“There is not too much time to pursue new research but we have had a new paper accepted and that has been published online. I continue to collaborate with fellows in digital twins and, in the meantime, I am working with the fellows in sustainability where we are doing some resilience research.”

Yue Xie was one of XX applicants for the National Highways Open Call which was launched by the Innovation team in November. The Open Call provides the company to source submissions from SMEs, individuals and contractors for solutions to key challenges associated with safety and customer. 

“The Open Call asked us to submit our research idea so I submitted my application with Galliford Try who were also my industry partners when I was on the Future Roads programme. I think the biggest benefit I have had from the programme is that I have built these connections with industrial partners. I hope I can get selected and, if not, we are still keeping our eye on that and hoping to build some connections with National Highways in the future research and applications.”

Work as a lecturer

We are considering in the near future on the roads with autonomous driving. There will be autonomous vehicles, electronic vehicles and human driving vehicles so we are working on how to coordinate all of them. We want to take the information on the infrastructure like the road, traffic lights, speed limit and see how we can coordinate them to make the system more intelligent and adapt it to different environments or emergencies. It is a very big plan and we haven’t finished that completely, but we are still working on it.

Continued connections with the fellows and Future Roads programme.

I still have connections with people. Before I joined Future Roads, my research was purely theoretic computer science and I didn’t have the application areas. After I joined the Future Roads Programme, I have had several publications and, as an independent researcher, I have expanded my research areas from theoretical to real world applications with a focus on transportation and traffic management. 

Surface performance of pavements: Zhaojie Sun

The service performance of road pavements is closely related to the quality of our life. However, pavements deteriorate over time because of the influence of traffic loads and environmental factors. 

For example, after a few years of use, there could be some potholes on pavement surfaces and some cracks in pavement structures. In general, the damage in pavement structures is more difficult to detect and repair, and it can lead to more serious consequences. Therefore, the evaluation of the structural condition of pavements is important. 

To achieve this goal, I am developing a technique for pavement structural evaluation by combining physics-based modelling and machine learning. This technique can predict the remaining service life of existing pavements based on non-destructive testing results, which helps formulate accurate maintenance and rehabilitation strategies. 

This technique can provide structural health information to a digital twin of roads to achieve more effective management of pavements. This work contributes to the digital and intelligent transformation of the road sector.