The safe, seamless intersection: smart infrastructure will save lives and keep cities moving

Continental tells Xavier Boucherat why smart infrastructure will be essential to achieve Vision Zero, particularly in complex driving situations such as intersections and mid-block crossings

Intelligent infrastructure will be pivotal to Vision Zero, the industry-wide ambition to make fatalities and serious injuries on the road a thing of the past.

That’s according to Jeremy McClain, Director, Systems and Technology at Continental. The mega-supplier has long supported the Vision Zero initiative, and is now pushing deeper into the connected space with developments such as smart intersections. Object detection sensors paired with Dedicated Short-Range Communication (DSRC), to deliver data, can link intersection infrastructure to vehicles. In turn, this could help semi-automated and fully-automated vehicles to deal with so-called ‘corner cases’, thus maximising their potential to eradicate driving-related deaths.

McClain is confident that in time, on-board sensors and in-vehicle technology will become sufficiently advanced that driverless vehicles can handle any situation, but autonomy is arriving at a slow pace, and whilst advanced driver assistance systems (ADAS) are finding their way into new models, humans will remain in the driving seat for some time. Meanwhile, intersections are a problem that need solving. In the US, McClain suggests around 43% of all traffic collisions, and 30% of traffic-related fatalities, take place on or near intersections. The main issue is visibility – quite often, neither drivers nor sensors can see beyond obstacles such as crowds of people, or large trucks.

We can always put more sensors in the car… If we fitted them into the corners of a bumper, that vehicle would be able to see a little better. But it would still have to behave very hesitantly at intersections, the same way we do, to make sure it can see everything

But even if sensing were better than it is today, the solution it could provide at intersections would still be inefficient, explains McClain. “We can always put more sensors in the car,” he says. “For example, if we fitted them into the corners of a bumper, that vehicle would be able to see a little better. But it would still have to behave very hesitantly at intersections, the same way we do, to make sure it can see everything.” Indeed, with OEMs or operators likely to shoulder liability for self-driving vehicles, they might act even more cautiously. Tentative behaviour from every vehicle at every intersection could quickly lead to traffic congestion.

The solution is believed to be V2X connectivity, capable of delivering information on intersections to a car in advance of its arrival. At its most basic, data from the intersection could be used to alert a driver to any dangers ahead. This data could then also be used to pre-condition forward-collision warning systems, or even autonomous emergency braking (AEB) systems. Typically, these require confirmation from a sensor to come into effect, but pre-conditioning the system can improve its performance.

Beyond intersections, there will be other applications. One is mid-block crossings. “Mid-block crossings tend to be very dangerous situations too,” says McClain, “because pedestrians and cyclists are often occluded, perhaps by buses, or by vehicles which have yielded to the pedestrian and are blocking the view of oncoming vehicles in the adjacent lane.” Sensors could be built into infrastructure like traffic lights and streetlamps. Some cities have already taken a lead on this – Portland, for example, has fitted 200 safety sensors provided by Current, part of power generation giant General Electric, on its three deadliest streets.

However, challenges remain for the general rollout of V2X. To begin with, virtually no vehicles are fitted with the required technology, and McClain believes it could take as long as 25 years before V2X technology makes it into 95% of vehicles on the road today. As a result, there are few vehicles for the technology to talk to.

The biggest challenge is creating a business model for smart infrastructure technologies which can provide a justifiable return on investment. It is difficult to justify investments on safety alone – there has to be additional value

Then there’s the question of the money to deploy the tech. “Deployment in vehicles themselves isn’t so much the issue, because the customer will pay,” says McClain. “What’s more a potential obstacle is fitting the infrastructure, for which governments and municipalities will have to pay. And so the biggest challenge will be creating a business model for these technologies which can provide a justifiable return on investment. It is difficult to justify investments on safety alone – there has to be additional value.”

McClain believes the solution will lie in leveraging the data collected by infrastructure. Cities and municipalities, he suggests, will be able to make use of their investments to better understand how people move, improve infrastructure maintenance schedules, where and when it can expect crowds, and how to improve the overall traffic flow. “I’m not suggesting safety doesn’t sell,” he adds, “but additional value is important, particularly at this pre-deployment stage when we don’t have all the proof.”

The other challenge is how to leverage V2X to protect vulnerable road-users. As it stands, there is no way to communicate with cyclists and pedestrians. Research and development continues, but virtually all proposals would require some sort of device to be carried on that person. A smartphone could work, and that person’s location and route could be broadcast to all nearby vehicles, but this would not only require people to have their smartphone on them at all times: it would require them to consent to their data being used. In an age of heightened concern around data privacy, will inner-city residents really sign up for systems which could effectively track their movements city-wide?

“This could present a problem if personally identifiable information were to be used,” says McClain, “but from our perspective, there’s no need for personally identifiable information – we should be able to anonymise it all. From a safety perspective, there’s absolutely no need for that, and even from an added-value data perspective, I still don’t see the need.” There may be a need to identify certain types of information such as vehicle types to improve the quality of the data, but this still doesn’t make personal identification necessary.

Projects such as connected intersections play a role in Continental’s vision for seamless mobility in cities, as the supplier prepares itself to meet the requirements of mobility service providers, and not simply OEMs. Work continues to improve the business proposition of connected infrastructure, and to educate potential customers worldwide. “That’s the most important thing we can show them,” concludes McClain: “that this technology can improve their overall way of life.”

This article appeared in the Q4 2018 issue of M:bility | Magazine. Follow this link to download the full issue