AI and big data to make the worlds’ roads safer

Article by i4 conect

Launching a start-up company at the age of 60 could be too daunting a challenge for many people, but for civil engineer and big data visionary Peter Jamieson, his business prospects have never looked more exciting.

His company, Newcastle-based Anditi, is embedding artificial intelligence (AI) into the planning, design, inspection and maintenance of the world’s streets and roads. He and his team have already established working relationships with road authorities in Australia, the United States and Europe. They anticipate improved maintenance planning, processes and subsequent road safety will save billions of dollars.

Anditi’s ‘Automated Road Asset and Safety Star Rating Assessment System’ is also regarded as a key enabling innovation that will help the United Nations achieve one of its key Sustainable Development Goals (SDGs) – safer roads. As Jamieson points out: “The global cost of deaths and injury on roads is 2.6 times the annual expenditure on constructing and maintaining the road infrastructure.”

Australia is one of the 100 countries that have signed up to the UN’s SDGs. As part of the safer roads goal, all roads that carry 75% of the world’s traffic must have a three-star or higher rating by 2030.

“This means about 25 million kilometres of roads have to be assessed by then. The existing manual inspections – driving a camera car and looking at its video footage – has only managed about three million kilometres in the past 10 years.”

The star rating is part of the International Road Assessment Programme (iRAP), which is tied to the UN’s goals to improve the safety of the world’s roads, especially in developing countries. The task is herculean. Jamieson explains that the current system requires an operator to code every 100 metres of videoed roadway according to 58 attributes and each attribute has a further 300 categories.

AI decision support

‍
“So, for example, in one 100-metre stretch there might be a roadside pole, or a tree, and you have to decide the degree of hazard this presents, plus there might be buildings or fences or other factors present. So for every kilometre of video that someone is inspecting, they are making about 3000 decisions. Typically, a human brain is good for about 8000 decisions a day, so you can see what’s going to happen to the accuracy of coding after a few kilometres. So what we’ve done is to develop the AI to automate this whole process.”

The technology that Jamieson and the Anditi team have developed combines the vast accumulation of data and 3D vision collected globally by companies such as TomTom with gaming technology. Jamieson’s own background is in managing and using large-scale environmental data. Their system turns slow, costly, manual road and traffic inspection into a web portal that allows users to remotely analyse and interact with road corridors almost anywhere in the world.

Jamieson also notes mounting interest from insurance companies that are looking at weighting insurance cover according to road safety ratings. The risk of driving on a one-star road is estimated to be four times greater than the risk of driving on a three-star road.

Government leg-up

‍
However, as promising and exciting as the outlook for Anditi is, Jamieson says the business and its innovations have only reached this stage through timely government support. It has received a $790,598 Accelerating Commercialisation Grant under the Australian Government’s Entrepreneurs’ Programme and guidance by facilitator Larry Weng from i4 Connect, the service delivery partner for the Entrepreneurs’ Programme Accelerating Commercialisation.

This support enabled Anditi to advance the technology from a prototype (Technology Readiness Level 5) to a suite of commercial services (TRL 8) within its overarching product, RoadViewer. It also helped cover the initial TomTom licensing fees.

Jamieson explains that the data on which Anditi has built its system is produced in the first instance by TomTom for its navigation products and for potential use by automated vehicles. The data comprises multiple formats – GPS, 360-degree video imagery and LiDAR (Light Detection And Ranging, laser beams that detect and measure surrounding structures such as buildings, poles, fences, roadside furniture, vegetation, etc). Anditi engineers incorporate all of this into a virtual reality road inspection system.

The back end behind much of this work utilises the additional expertise Anditi’s engineers have in gaming technology to facilitate both realism and user interaction. “When you combine all of this you end up with a rich, realistic view of a road, all of its components including its condition, and its surrounds,” explains Jamieson.

“And we can measure all of this … the width of roads and traffic lanes, line markings, the height of surrounding structures, curvatures and gradients, etc.”

As part of a pilot study in the United States, the Mississippi Department of Transportation used RoadViewer to update its database of highway overpass heights. Traditionally this would have been done manually with a measuring pole. This is no longer realistic without closing the highway – a prohibitive step given the density of today’s traffic flows. RoadViewer, however, enabled the authority to undertake the measurements virtually, and accurately.

In another example, RoadViewer was used to map a route for a prime mover to ferry a 90-metre wind turbine through the streets of Perth. The program was able to quickly identify a suitable route – one with wide corners and no obstructions.

Game changer

‍
The Anditi system is considered by many as a ‘game changer’ in helping countries cost-effectively meet their commitments to UN SDGs and iRAP and its aim to improve road infrastructure to save lives.

The Anditi system is the only one that uses both LiDAR and 360-degree imagery and is the only system accredited globally by iRAP, says Jamieson.

In assisting Anditi in its commercialisation journey, i4 Connect Commercialisation Facilitator Larry Weng says the first step was establishing a good working relationship. In the case of Anditi, he also helped to sharpen the value proposition that risked being lost in the vastness and scope of the global opportunity that was present.

“Peter is a serial entrepreneur and an experienced data analyst, but the numbers involved in this project were, and are, immense. Tens of millions of kilometres of roads worldwide, billions of dollars in potential for local, regional and national governments, countless lives saved. It’s a dataset almost beyond comprehension. And rationalising all of this is Peter’s novel AI tools and algorithms.

“So we had to drill deeper and deeper into the business plan to put real figures against real gains and make the value proposition really stand out,” says Weng.

“I also saw a networking opportunity and introduced Peter to a Brisbane-based AI company that provides predictive analytics to governments. It says, for example, ‘this stretch of this road needs this upgrade to prevent these accidents’ – predicting what will happen before it happens. I saw the two businesses as complementary. Working together these businesses can provide an even greater capability in this field.”

Like most facilitators in the Accelerating Commercialisation program, Weng brings his own experiences as a start-up founder to the table. It’s this experience that allows him to provide real-world insights into navigating the commercialisation journey. Finding synergies with other companies can be a valuable part of this process, especially for small Australian businesses with global ambitions.

“Facilitators also bring an open mindset to the relationship, which can be important because the start-up creator will tend to be very focused on the innovation and not necessarily aware of all the challenges or opportunities.

“And we are independent. We have no financial interest, but we want to see the business succeed. So we will be blunt, sometimes, in the questions we ask and the answers we need. And that’s also why a good working relationship is essential.”

Jamieson is unhesitatingly in agreement. “The funding from the Accelerating Commercialisation program and the support from people like Larry Weng allowed us to take our prototype into the commercial arena. Our inspection system, the only one in the world accredited by iRAP, was a direct consequence of receiving the commercialisation grant. We wouldn’t have got there without the impetus this provided.”

‍