As we slog through this first year of a new decade, it's tempting to look around and wonder where all the robots are. If we go by classic science fiction dates, then by now a HAL 9000 computer should have gone rogue on a mission to Jupiter, killing all the space craft's crew, while just last year, November 2019, to be exact, Harrison Ford should have been making a healthy living hunting down and 'retiring' rogue humanoid robots around a rainy, desolate Los Angeles.
You’ll note that neither of these has happened. No robot brought me my dinner last evening. No robot fetched me the TV remote. And, crucially, no robot drove me, or you, or pretty much anyone to work today.
So, the initial promise of a world of robotic, autonomous cars has not materialised, and unless there are some staggeringly fast and widespread developments in the next 12 months, we're going to miss the date originally promulgated for the arrival of widespread autonomous vehicle services. Even though the Irish Government is now laying out plans to allow the use of such cars on Irish roads (for development purposes) and even though there are tentative plans for a centre of autonomous development excellence adjacent to Shannon airport, you're almost certainly not going to see massed ranks of robotic cars on the road this time next year.
So, when? Well, have you got your piece of string? Have you got a ruler? Then let’s begin.
This foundation has to be very solid. You have to cover all situations because there's no driver as a fallback option
2030 is now being touted as the next most likely date for the widespread use of autonomous vehicles (AVs). Certainly, Continental (which makes, on a huge scale, electronic components for cars as well as tyres) reckons that's the right date. Andree Hohm is Continental's chief of autonomous driving, and he predicts it'll be the end of the decade before the tech starts to take hold.
“People always ask me when driverless vehicles will be on the road and I tell them the answer is ‘today’,” says Hohm. “If you have a very specific application area, for example like a private road, and want to travel at low velocity, you can buy such a vehicle. Urban situations may be more complex than highways, but they have one key advantage – slower speeds give you more opportunities when systems fail, or you get erroneous readings from sensors.
‘Very promising’
“We at Continental consider driverless mobility to be very promising. It will have a significant share of the mobility of the future.”
But he goes on to say the key is to create a bedrock of safety and reliability in the systems, so that people are convinced the car is safe to use. “This foundation has to be very solid,” he says. “You have to cover all situations because there’s no driver as a fallback option. It still sounds a little bit creepy, if you think about it – going into a car where there is no one at the steering wheel. This is for us a clear signal that we have to introduce those functionalities step by step. We need to clearly show what we are introducing. We need to involve people in pilot projects, so they can actually experience how exciting the technology is.”
That has become a recurring theme in the debate over AVs. The robustness (or otherwise) of the software and sensors involved is a major issue, both for those developing such systems, and for those considering becoming eventual end-users. Right now, there’s a distinct lack of trust involved, certainly from the end-user point of view, and the recent case of Uber’s apparent lack of safety culture, and that culture’s involvement in the death of a pedestrian, struck by a prototype AV in Arizona, will not have helped matters.
There are generally two types of responses from people heavily involved in AV development when it comes to this aspect of trust and safety. One is to, cavalier-style, dismiss the worries as those of eggs facing a future as omelettes – some are going to get broken. Others, more realistic and more sensible, acknowledge the size of the problem and duly adjust their expected arrival dates for mass AVs.
A recent survey by Neckermann Strategic Advisors, dubbed Being Driven, found this trust aspect is going to be a major hurdle for companies pushing autonomous tech to climb over. According to the report, we're happy to judge a conventional new car safe once it's been prototyped for three years and passed the EuroNCAP crash test. For AVs, though, we expect 100 per cent reliability (having, in fairness, been sold autonomous driving by the likes of Google and Uber) yet: "AVs are built by humans that, by definition, cannot reach a total zero-defect product."
The Neckermann report goes on to suggest that AV technology should be sold to the public in terms of it being similar to aviation technology – that we already fly in vehicles that are, for the most part, autonomous and electronically controlled. That does, though, seem to rather miss the fact that aeroplanes have dedicated, highly trained, and well-paid pilots and co-pilots to take control in an emergency, and that there are rather fewer objects to hit at 38,0000-feet.
‘Leap of faith’
The trust aspect rears its head again when the Neckermann report asks people to rate how much they trust those working on AV technology to “do business fairly”. Some 58 per cent said they trusted car manufacturers and dealers either “only a little” or “not at all”. That suggests existing car makers may have a great deal of difficulty convincing people to take the leap of faith into driverless cars.
There’s also a disconnect between ownership and actual likely usage. Most expect AVs to be primarily the preserve of vast fleets, belonging to ride-hailing firms or new versions of public transport companies. The expectation is that these robotic hire cars will be legislated for by the vast cities in which the 10-billion-strong human population of 2050 will mostly live. Yet, according to the Neckermann report, 71 per cent of the 16-24-year-olds surveyed would rather buy and own their own autonomous vehicle than hire one journey-by-journey.
Do we believe that a shared, electric, autonomous mobility environment is possible? Yes, we do. But probably later, and in a different way than expected
Indeed, the report notes that the switch to AVs, if it’s going to happen, requires humanity, en masse, to make a conceptual double-leap, putting their faith in the computers needed to control the cars, as well as distancing themselves from the long-hold emotional connection to cars (a factor generally ignored by most other reports into the subject.
"We can't think of any examples of technology adoption, where humans – well known for their resistance to change – were able to embrace two similarly disruptive, linked innovations, at the same time. Is shared autonomy a 'double-leap' too great for people to accept?" asked the Neckermann report's authors, Lukas Neckermann and Frederic John.
Legislation will, of course, play an especially weighty role in all of this, and not just decisions about how safe an AV has to be to get its licence. Concerns about increases in traffic congestion and urban pollution (the very things AVs are supposed to play a part in solving) could well put a stop to the entire process of large-scale AV usage, certainly by private operators, before it gets started. According to a new report by environmental think-tank Transport & Environment (T&E), the arrival of Uber in cities such as Paris and London have seen numbers of registered taxis almost double. In London, the number of journeys by taxi or hackney has increased by a quarter since Uber started operating there.
Emissions
T&E’s analysis estimates that in London and Paris alone, the emissions of Uber taxi services could be as high as half a megatonne of CO2 – 515 kilotonnes of CO2. This is equivalent to adding the CO2 emissions of an extra 250,000 privately owned cars to the road.
These kilometres have predominantly been driven by petrol and diesel cars, exacerbating the air pollution crisis in European cities. Worryingly, French government data from 2017 showed that 90 per cent of the registered private hire vehicles, which includes Uber, were diesel cars. The high share of diesels in the private hire vehicle (PHV) fleet can be also found in the traditional taxi market. It’s thought that other European cities where Uber operates have similar levels of diesels in their fleet.
Yoann Le Petit, new mobility expert with T&E, says: "Uber's CEO tells us they 'do the right thing, period'. But the reality is that Uber is part of the traffic and pollution problem, adding car trips in our cities and adding to the climate and pollution crisis. If it wants to become part of the solution Uber needs to stop using petrol and diesel cars and rapidly shift to 100 per cent electric rides. That's the right thing to do, full stop."
Obviously, legislating for electric vehicles will reduce or eliminate the local pollution, but it won’t help when it comes to congestion, and many have predicted a future of total gridlock if vast fleets of robotic cars take to the streets.
So, back to our original question – when can I have an autonomous car? The answer, for now, seems to be a relatively imprecise “eventually”. If a global set of safety regulations can be agreed upon; if concerns over pollution and gridlock can be allayed; if a sufficient number of potential buyers can be lined up – if all of that comes to pass, then you will be able to have your autonomous car. As with the 2021 date looming and likely coming and going, there is not much point of making a more solid prediction than “eventually”. As the Neckermann report puts it: “Do we believe that a shared, electric, autonomous mobility environment is possible? Yes, we do. But probably later, and in a different way than expected.”
