Autonomous driving will never succeed without the development of largescale underlying infrastructure, including physical components such as parking and charging points, as well as mobile networks and regulatory aspects determining vehicle behavior. Surprisingly, little thought has been given to developing a coherent infrastructure integrating these diverse aspects, never mind how this may change the way people travel and interact with public transportation including rail.
As a result, there is little consensus over what the intended consequences of autonomous driving should be, nor the unintended consequences that might result. A key point is that as autonomous driving unfolds it is quite likely that the infrastructure will encourage centralized control, under pressure for greater speed without compromising safety alongside efficient routing of traffic with congestion avoidance.
With a likely trend away from individual ownership to shared use, autonomous vehicles could morph into a combination of train and taxi, arriving to pick people up, but then driving entirely under control of the infrastructure as they slot in and out of traffic during their journey, maintaining closer distances to vehicles in front and behind than is safely possible with human drivers. Some of these vehicles would be dedicated to “last mile” transport ferrying people to and from work, airports and rail stations, while others would be long distance carriers following trunk routes, quite possibly at greater speeds than are allowed today.
This hope of speed limits being raised significantly has led to some speculation that autonomous driving could compete with high speed rail as well as air travel over intermediate distances between 100 and 500 miles, or perhaps even further in some regions such as the USA.
Such speculation may be premature, but governments and planners do need to factor these considerations into planning of major transport investments in rail networks, roads and airports given the long time scales involved. They need to make some projections about autonomous driving, and then assess how that should alter how they cater for these in a flexible way that will allow for changes as these projects advance.
Layouts for parking could be very different if, say at airports and major stations, autonomous vehicles drop their passengers off like taxis and then go and park themselves. Especially if many such vehicles are pooled or even owned by some rental or taxi firm along the kind of model the self-hailing app companies like Uber espouse, then there may not be need for much parking at all at airports, stations and places where many people drive at present. This could free up a lot of space for other activities.
Another quite different long-term question concerns control over the vehicles themselves, and what role centralized management will play over routing and driving style. It is widely assumed at present that vehicles will be as the word autonomous suggests in control of their own destiny as servants of their owners, but already there are signs that their speed, acceleration, braking and steering could be controlled by the infrastructure, even if the car has the capability to take over. If the communications infrastructure is sufficiently robust, reliable and fast, it would eventually be safer and more efficient, from a traffic management perspective, to have vehicles totally under its control for driving and routing, with the owner reduced just to entering the destination as when hailing a taxi via an app today.
Such longer-term considerations are very much in the back seat when it comes to emerging infrastructure initiatives. This was the case for the US hearing held in June 2018, by the Senate Committee on the Environment and Public Works, entitled Innovation and America’s Infrastructure: Examining the Effects of Emerging Autonomous Technologies on America’s Roads and Bridges.
It is true that Senator Barrasso did stress the importance of flexibility in America’s infrastructure as autonomous vehicles are introduced, during his opening address. But the session ultimately disappointed anyone seeking concrete details, being confined to rather anodyne statements of intent, and need for a sea change in the way the country’s roads and bridges were funded, essentially just calling for more money.
What did emerge from the hearing was a tension between the desire to give as much freedom as possible for innovation while exerting some regulatory control over development of the infrastructure. There was also a tension between the needs of public and private transport in cities, where the US has traditional favored the car over buses and trains but with some kickback in recent years because of congestion, pollution and the specter of carbon emissions.
Inevitably politics featured here with Polly Trottenberg, Commissioner of the New York City Department of Transportation, arguing that solving the congestion problem of metropolitan areas must begin with “robust investment in our mass transit systems”, insisting that the benefits of these dwarfed those of autonomous vehicles.
But the message should have been that autonomous vehicles can be a part of the grand plan, and could shape the infrastructure accordingly. They can operate as the end points of public transportation systems while still enhancing the autonomy of the individual. The EU has picked up on this better, with its drive for harmonized development of smart cities, smart grids and the urban connected car infrastructure.
The EU’s white paper called ‘On the road to automated mobility: An EU strategy for mobility of the future,’ published in May 2018, advocates infrastructure that caters equally for private cars and public autonomous vehicles, such as self-driving buses following set routes, as well as commercial trucks.
The EU paper does not address the potential for integration between autonomous cars and public transportation at all, but is stronger on the wireless communication front. Here it highlighted three significant European projects, one being the connected vehicle eco-system Autopilot that began in January 2017, focusing on open IT platforms for connected ecosystems. Then 5GCar project started in June 2017 is evaluating existing and future spectrum usage for the ecosystem contributing to standardization efforts.
Finally, the Ensemble project starting around now is focusing specifically on standards for multi-brand platooning for trucks. The aim is to enable any vehicle in principle to join platoons rather than being confined to trucks from their own stable, to enjoy fuel savings and reduce congestion by travelling more closely together.
One point that has emerged from early projects and consultations is the old conundrum of balancing rural and urban needs in the infrastructure. The problems of poor fixed and mobile communications with blackspots in remote areas is legendary, and still not fully solved, with clear implications for autonomous driving if vehicles lose connections with the infrastructure. There is also the issue of communication in tunnels and within structures such as underground car parks, where both cellular and GPS signals can fail to penetrate. That is an argument for ensuring that vehicles can operate safely in genuine autonomy, given that even if some form of wireless communications is enabled in such places, signals cannot be guaranteed everywhere.