Ordnance Survey (OS) is using advanced imaging technology and artificial intelligence to help utilities firms reduce the cost of street works and to help private businesses benefit from emerging technologies, such as 5G and autonomous vehicles.
At a media event in the London Stadium at Queen Elizabeth Olympic Park, acting CEO Neil Ackroyd explained how OS is running a pilot project to create real-time maps of UK streets. These maps will be used by OS and its customers, including utility firms, local authorities and the providers of emerging technologies. Ackroyd said OS, as the national mapping agency of Great Britain, has a responsibility to ensure its data is fit for purpose:
We need to think about how we can produce new information that we can use to improve, not just how our assets are managed, but also information that captures the general environment when it comes to mapping the urban environment.
Ackroyd said the answer to this challenge is a roadside infrastructure dataset. Working alongside technology specialist Mobileye, OS is retro-fitting utility vans and cars with Mobileye 8 Connect imaging technology to detect road-side features, such as network boxes, traffic lights, road signs and man-hole covers. Dashboard-mounted cameras record information in real time as vehicles travel along the road. Ackroyd said the system could provide big benefits:
We need to get the roads that have been built to work much more efficiently and effectively. We’ve been working in this area for many years but we’ve not managed to substantively reduce inefficiencies and impacts. There’s a significant opportunity for cost savings.
Greater detail on road-side infrastructure will help utility firms and local authorities deal with the logistics of street works, which cost the UK about £5.5bn a year. About 4.5 million holes are drug annually in the UK, with 1.2 million alone in Greater London. These street works produce six million man-hours of disruption, costing as much as £1bn annually in lost productivity. Emerging technology creates additional headaches. Ackroyd said the guidance systems that control autonomous vehicles will need real-time data on roads. New 5G networks will require more infrastructure, too. Ackroyd said public and private organisations engaged in emerging technology deployments will benefit from detailed knowledge of street works:
For 5G to be successful, we have to invest in fast fibre roll out. We have standards to help us share data but they haven’t solved the problem of the lack of motivation to share information and undertake asset mapping. Putting in this new layer of capabilities, with much higher currency at a much lower cost, is close to the magic butler that utilities have been asking for. We’re really at an exciting point of change, where we can get more out of existing infrastructure to reduce the risk of accidents and create new productivity in terms of how street works are completed.
Creating the next-generation mapping system
Professor Amnon Shashua, president and CEO of Mobileye, described the announcement as an “exciting moment and the dawn of AI mapping.” He explained how Mobileye technology detects road-side features in real time before sending this data to the cloud and onto OS for enrichment:
It’s not a one-time system – it’s continuous. If something changes in the infrastructure and a vehicle passes, OS knows automatically. This is a world-first and this is why it’s exciting. You equip existing vehicles with low-cost cameras – and then the only cost is compute in order to create something very deep and important for the mapping industry.
The real-time information on street works in the cloud is analysed by OS using artificial intelligence (AI). Paul Cruddace, technical change and innovation manager at OS, is leading the team that is developing the next-generation computing stack that will power this evolution in AI mapping. His team aggregates the real-time data and then considers how this information matches a historical database of features held by OS:
The exciting thing for us is we’re creating a new geospatial layer. The technology challenge we have is to take the data from Mobileye systems to a cloud-based environment and then apply algorithms so we can serve data out to our customers, such as utility firms.
Future developments to the mapping system could focus on attribution, so OS is able to certify who owns a road-side feature and to quantify the quality of the asset. As more and more cars are fitted with cameras, some of this data could be sent in by the public. Cruddace said relationships with third parties will be crucial to shaping these developments:
We’ve very interested in getting other organisations involved in trials so we can learn what customers want from the system and to develop the feature set. Once we have an assets database, we’re looking to create a notification system, so we can tell people when new assets appear. We’re putting a lot of effort into this and we’re very excited about how it will develop going forwards.
Eleven vehicles in the OS fleet are currently using the Mobileye system in a cross-country trial. OS is talking with utility organisations and local authorities now about how this initiative might develop. Acting CEO Ackroyd said OS will scale-up to thousands of vehicles in the next few months. Northumbrian Water has become the first utility company in the UK to join the project. Operations solutions manager Clive Surman-Wells said Northumbrian Water is using the trial to explore how to serve customers more effectively:
At Northumbria, we operate two large networks: a waste-water network and a water network. Those assets are largely buried and out of sight. But when you look around, there are little clues on the ground about where buried assets exist, such as man-hole covers and fire hydrants. Most of those we have mapped, as do the other utilities – but this technology provides a step-change in mapping precision.
This helps us respond to customer needs when things go wrong and to be able to identify where those assets are quickly. Every time we dig a hole, there’s a risk we can hit someone else’s assets – knowing the precise location of an asset can help us reduce the risk of strikes. What we want to explore during the trial is how far we can go with this technology.