The road ahead for autonomous vehicles is longer yet more promising than many believe
Scarcely a week goes by without another major announcement in the world of autonomous vehicles (AVs).
British transport authorities are pushing for them to be on roads within as little as two years. BMW and Daimler recently unveiled a long-term, strategic cooperation on AVs, and investment into US-based self-driving startups hit as much as $4.5 billion last year.
Yet for every technology optimist dreaming of a utopian transport system that eliminates accidents and congestion at a stroke, there are other commentators pointing out emerging roadblocks: cost and the technological, social and regulatory challenges we need to overcome to ensure AVs do everything they are designed to do safely.
So what does a realistic pathway look like for the development of this ground-breaking technology?
“Most commentators are being a bit too aggressive about when we are likely to see AVs fully adopted on the streets, and also too conservative when it comes to imagining their true transformative potential,” says Rich Parkin, a partner at Strategy&, PwC’s strategy consulting house.
While automotive companies have announced plans to produce “mainstream” vehicles that are fully autonomous in certain applications by 2026/27, the economics of producing them in significant numbers obviously needs to be viable for that to happen. These companies are currently earning returns that are in many cases below their cost of capital – at around 4.5 percent.
Yet AV investments are huge and will continue to demand a great deal of capital.
Under currently announced plans, Strategy& estimates that return on capital will drop to between 1.7 and 3.1 percent for automotive companies between 2025 and 2030. These are not attractive levels for shareholders, being significantly below other sectors, and the cost of capital.
Moreover, the economics of personal AV ownership means that we are unlikely to see any notable adoption of AVs for personal use before the early 2030s. Even then, which is when we anticipate the industry reaching scale, an AV will still cost about £5,000 (US$6,500) on top of what a standard vehicle would cost.
We expect that – as with the early predictions for electric vehicles – timelines may well slip.
“We expect that – as with the early predictions for electric vehicles – timelines may well slip,” says Parkin. Moreover, many of the investments will not work out as intended.
“This means that different stakeholders - public sector, technology suppliers, and automotive manufacturers - need to have a fundamental rethink of the way forward, with a clearly identified path to achieve their goals, be they profitability or public policy outcomes, and based on a clear-headed view of likely developments.”
Adoption of AVs in major cities for “high-utilisation” applications will be the first segments where the economics will make sense. One early example is Drive.ai, a Texas-based start-up that has deployed vans painted in eye-catching orange which run along routes using designated stops, to and from popular locations like office parks and retail outlets. Commercial truck “platooning” will be another area, with major savings in driver labour costs.
Then there is the issue of technology – and safety. The Society of Motor Manufacturers and Traders in the UK says 25,000 serious accidents could be avoided every year thanks to technology which reduces driver error. That’s good news, of course.
Yet the technological challenge has been starkly illustrated by a series of crashes in the U.S. last year. First, Uber halted its AV programme after one of its cars was involved in a fatality in Arizona, then another fatal crash happened in California when a Tesla in autopilot mode collided with a roadside barrier and caught fire.
Data published earlier this year show that Waymo, the Alphabet subsidiary that started working on driverless cars a decade ago, led the pack in terms of how few times a human had to take control of a vehicle to prevent unwanted outcomes in tests. But there was a fairly large gap between Waymo and the next batch of companies. This highlights that there is still some way to go before we can declare AVs safe.
“Their impact on safety has yet to be proven out by the data,” says Michelle Avary, head of Autonomous and Urban Mobility for the World Economic Forum. “How do you decide what constitutes a safe autonomous vehicle? And what data do you need to prove it?
The data is another challenge – in particular the complexity of managing the huge volumes of mapping data that will be needed to allow AVs to operate safely in cities.
Developing technology the right way will require close cooperation between public and private sectors and a close eye on the right regulatory frameworks to achieve this.
What’s certain is that the move to full autonomy in personal mobility through AVs will be truly transformational. It will free up time for drivers, providing economic benefits. It will eliminate a lot of the location-based activity that has long dominated vehicle purchase, ownership and use; cars will drive themselves to the garage for servicing, for example.
The societal benefits are also broad. Driverless vehicles could help reduce social exclusion of the mobility-impaired as ageing takes its toll on populations. “Just think about the people who lose their driving licence as their eyesight deteriorates with age and whose lives could be transformed by AVs,” says Parkin.
The potential of AVs to widen access to mobility starts to open up huge horizons for nations with demographic and other challenges such as Japan, where there are transportation “deserts” that disproportionately hit the elderly and single mothers.
“Access to mobility makes a very dramatic difference, not only in individuals’ lives but in nation states,” says Avary.