This project set out to produce a methodology to assess the suitability of the road network for autonomous vehicles. It will bridge the gap between vehicle capability, infrastructure quality and the resultant implications for CAV safety to ensure that they are operating in optimally safe environments as they are introduced to the road network.
FTVG supported the team with funding to purchase key datasets which fed into the infrastructure assessment methodology. It also allowed them to build a website to publicise the project and to give it a sense of legitimacy, something which proved very useful when approaching experts. Finally, it allowed the team to host 2 interactive workshops in which attendees voted to inform the weightings of the index parameters which constituted the index. A total of 41km of road network comprising of 699 links was assessed across 11 different parameter categories.
The project demonstrated that the infrastructure quality was generally of reasonably high suitability for CAVs across the urban area of Leeds assessed (see map below). However, there were key locations where quality dropped significantly, indicating that the study area had potential to be challenging for CAVs to navigate safely. Furthermore, when testing the impacts of inclement weather conditions, the suitability dropped across the network but also revealed new areas of concern.
The road environment is mostly quite legible and intuitive to human drivers, as it is designed to be, but CAVs perceive the road quite differently. This project demonstrated that by confirming the areas that were least suitable were not necessarily those you would expect from a human perspective – particularly where the digital infrastructure was lacking in quality. Furthermore, the index was calibrated to be quite cautious in terms of safety, so more segregated routes away from vulnerable road users tended to score relatively well.
Lawrence said, “This project was a fantastic opportunity to dive deep into a technical specialism and to explore the possible implications of a highly anticipated technology. Through the development of the project, it also offered the chance to speak with some key players in the field of connected and autonomous mobility who really challenged our thinking and helped to shape the project into something that could be of value to the industry.”
Ramit added “The project gave me an opportunity to explore a visionary domain of autonomous vehicles. The most valuable insights were from the interactions with the industry experts and the knowledge sharing sessions that were organized during the project. The key take-away from the project is that the autonomous vehicle industry has immense potential to improve road safety and efficiency of shared transport.”
The team faced a number of challenges throughout the project, most notably through the pandemic. Ramit, based in India, in particular felt the pressures;
“The Covid-19 second wave in India was the biggest challenge faced by me during the project. I would like to take this opportunity to thank Lawrence for being very patient and helping me in all possible ways.”, he said.
Lawrence also felt the ongoing pressures of the pandemic throughout the project and also reflected on some of the technical challenges he faced;
“Our project required the bringing together of many traditionally very separate areas to produce our insights. As a novel approach, data was rarely collected for purposes similar to our investigation, so it often required a degree of manual manipulation which proved to be very time consuming.”
However, the team saw several useful applications for their work in the longer-term;
“Having proven the concept of our methodology, there is also opportunity for car manufacturers to develop it further and input their known vehicle capabilities to produce a more bespoke analysis. As well as assessing the readiness of their vehicles, it could also be a diagnostic tool where driverless technologies fail to correctly navigate the road environment”, said Lawrence, “By bridging the gap between our current understanding of infrastructure needs, and those of CAVs, the necessary remedial actions can be taken over the course of upcoming maintenance cycles to ensure safe operation. Perhaps ironically, our ambition for this project is to make itself redundant within 10 years as CAVs will be mostly operating within environments that are the most suitable.”
Ramit was grateful for the opportunity to investigate this field and expand the debate around CAV’s in the urban environment;
“I would like to thank FTVG for this amazing opportunity and guiding the project throughout its progress. The lasting legacy is providing the proof of concept that operation of CAVs on urban roads is a reality with given infrastructure and technological support.”
View the group’s website here: www.autonomousvehicleindex.com
Download the report here: CAV Road Scoring Index