• The safe running of self-drive vehicles will rely on vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communication
• The University of Michigan’s Mcity facility is testing the technology in a 32-acre artificial smart city
• Car makers are already integrating semi-autonomous technologies in their latest models, while some cities are trialling smart street lighting and Traffic Light Information systems

The automotive industry is currently very excited by the possibilities of connected cars. The promise of so many science fiction films is about to be fulfilled.

Manufacturers are regularly announcing either semi-autonomous technology in their latest models (not just Tesla, but also the likes of Volvo, BMW and Volkswagen) or plans for the next generation of cars, due at the end of the decade, with self-driving capabilities.

Much has already been written about how autonomous cars will improve our driving lives, make us all more relaxed or productive in the vehicles of the future, improve safety, increase efficiency and be better for the environment, but it will take more than just clever self-driving cars to deliver those benefits. Because in order to make the most of this brave new world of motoring, it’s not just the vehicles that will need to communicate with each other to share vital information; the entire road network infrastructure will have to play a part, too.

The basis for the safe running of electronics-laden self-driving cars is vehicle-to-vehicle (V2V) communication. The information about the road around them – gathered by the likes of sensors, cameras and Lidar systems (which use lasers) – will be shared with other connected cars on the road around them. These autonomous cars will rely on dedicated short-range communications (DSRC) to send and receive information. As the main information shared by DSRC units relates to a vehicle’s location, its speed and the direction in which it is travelling, the focus of V2V technologies is collision avoidance. This means, of course, that all other vehicles on the road also have to be connected. If not, vehicles could be invisible to connected vehicles without other technology such as radar.

Also vital will be vehicle-to-infrastructure (V2I) communication. Static features such as intelligent road signs and traffic lights will need to be in place as consumers begin to accept and adopt self-driving cars.

An indication of how V2I could work is already available, albeit on a proving ground at the University of Michigan’s Mobility Transformation Center (MTC), recently rebranded as Mcity. The facility includes a 32-acre artificial town opened in 2015 to test autonomous cars in real-life traffic situations.

Huei Peng, the centre’s director, has said that information shared between autonomous cars and the kinds of smart road infrastructure being tested at Mcity will be crucial to the development of a self-driving future.

“Communication will help cars or autonomous vehicles to see more clearly and further,” he said. “Future road infrastructure needs to be designed to support those human drivers and robot drivers.

“We think we need to have a very systematic way of understanding how these cameras, radar and Lidar see the environment, and we design the infrastructure for them to be driving on the road safely.”

Among the smart-infrastructure features being examined at Mcity are road signs, which need to be not only visible at 80mph to the human eye, but also readable by vehicle cameras fitted with the traffic-sign recognition features now available on many modern cars, from low-slung sports cars to high-riding SUVs.

Sensors on self-driving cars also rely on the paint used for road markings to keep them in one lane, so Mcity is using special metal mesh on the lines to help sensors of an autonomous car.

Roadside units (RSU) fitted to traffic lights at junctions can tell from 300 metres when a vehicle fitted with DSRC is approaching, how fast it is travelling and other data. These units can then use signal phase and timing (SPAT) to change the phasing of the lights to control the traffic flow.

As Debby Bezzina, senior programme manager at the University of Michigan Transportation Research Institute, told a group of reporters recently: “What better to tell the vehicle to go into the intersection than… the intersection?”

Indeed, data is already being passed the other way in real-life situations in Las Vegas, with Audis sold in the city being fitted with a Traffic Light Information system that tells the driver, via a display in the instrument cluster, how much time there is before the lights change. The theory behind this is that providing the driver with this additional information helps reduce stress and allows them to relax.

Other pilot programmes for smart infrastructure are currently in operation in Copenhagen. Smart street lighting, for example, brightens when cyclists are crossing busy junctions, to help make them more visible to motorists, while wifi access points can detect road users’ mobile phones as they pass and triangulate their position, to enable streets to be mapped to optimise mobility and safety and minimise CO2 emissions.

One key element underpinning all this technology is a reliable, high-speed radio communication system. Car manufacturers are using the 802.11 wifi standard, but when 5G is introduced at the end of the decade it could prove crucial, as it will be able to send data over greater distances and enable different radio networks to work together.

As Andy Nix, professor of wireless communications at the University of Bristol, told E&T magazine: “We see ultimately the challenge being handled by a range of radios. You have 802.11p, LTE and millimetre-wave possibilities, ideally all being coordinated through software-defined networks. With multiple technologies, hopefully as one is weakening, another is kicking in.”

So when does all this technology cease being theoretical and we start to see it being implemented on our streets and highways?

The end of the decade seems to provide a nexus for autonomous-capable cars and 5G communication, but wider acceptance and uptake is likely to be another decade away, with complete saturation taking perhaps another 10 to 20 years.

The smart infrastructure is arguably easier to put in place, with a rolling programme of roadside upgrades possible once testing is completed and standards established.

However, this process relies on public money. In an era when funding for large-scale infrastructure projects – especially those that might be considered non-urgent – is hard to come by, it’s difficult to see governments prioritising spending on smart road technology. It’s possible that the technology will be ready and workable before there’s money to implement it.

The way the spread of electric vehicle infrastructure has stuttered could prove an indication of the difficulties ahead for smart roads. But they will arrive at some point in the coming decades and, when they do, science fiction will finally become fact.