Autonomy in the UK: Riding in Nissan's Driverless Leaf In East London

By James O Malley on at

Did you know that there are already driverless cars on London’s streets? No I’m not talking about a Tesla using “autopilot” mode, which only automates some driving functions, but actual, 100% autonomy? On a small stretch of East London roads, the future is starting to appear as Nissan has spent the last couple of months testing its autonomous technology. And a couple of weeks ago, they were kind enough to let me go for a ride in the passenger seat.

I started by zooming out, and speaking to Takao Asami, Nissan’s SVP of Advanced Engineering and Research. What’s the vision? What does Nissan see the future looking like? And curiously, unlike Ford with its relentless emphasis on “mobility”, the key message to push for Nissan is one of safety. “In 1995 we announced that we will reduce the number of the fatalities and the seriously injured by 50% by 2015 in the UK in the USA and in Japan. And actually we achieved it”, Asami explained. “So after that we made a second announcement we will reduce the number again by 2020. Now of course we have been taking care of the improvement of craft safety or active safety but it’s not enough to come closer to almost zero fatality. So autonomous driving is a natural extension of our effort based on this safety vision that we put in place more than ten years ago”

And there’s the obvious follow-up: Why is the company bothering to test autonomous vehicles in Beckton, near the Excel Centre? Can’t the company perfect its work in Tokyo?

“Every city, has its own difficulty. And London is not an exception”, he explains. He told me how on top of global control software, each region also requires its own software tweaks, to account for different rules and regulations - and perhaps most interestingly, their own cultural norms and what Asami calls “ethics” when it comes to driving.

I think this is a polite way of saying that software needs to account for the fact that Londoners tend to drive more aggressively than people who live in the country. Plus, say, people in Naples drive like utter maniacs compared to Northern Europeans.

So does Nissan’s software tell the car to be more aggressive about merging traffic in some places? “Err, we are not there yet. But we are aware of that need”, says Asami.

The Testing Setup

The test route is approximately 10km long and has been chosen because it allows Nissan to test autonomy on three different types of road. Starting my test “drive” (though, umm, no one was driving) from a hotel near the Excel Centre, after Tetsuya Iijima, Nissan’s General Manager of Autonomy pulled out of the car park, he took his hands off the wheel and the car was away.

We first went along the A1020 - negotiating roundabouts, in a test of a suburban link road, and followed it all of the way to Newham Way - a dual carriageway with slip-roads. Finally we ended up turning around Freemason’s Road, a residential street, where the modified Leaf would encounter traffic lights, pedestrians wandering out into the road and zebra crossings.

When it comes to actually rolling out autonomy full time, Nissan is planning a phased introduction. Its 2018 cars will be capable of driving autonomously on motorways and dual carriageways, a little bit like Teslas are today. This is because, despite the speeds involved, these roads are relatively easy. There are no pedestrians who can leap out suddenly, no bikes, and everyone is following a strict set of rules. 2020 is the company’s target date for a car that is fully autonomous everywhere.

As far as I can tell, this phased introduction is driven in part by the types of cameras available. The prototype leafs had 12 cameras on board, but arguably this is a bit overkill. The bigger problem is that they operate at relatively low resolution - because on-board computers cannot process the images fast enough otherwise (and fast processing is exactly what is required in an autonomous vehicle). The hope is, according to Iijima, that by 2020, better technology will be available at price where Nissan can still make affordable cars. Asami agrees with this assessment - he told me in a separate interview that the cost of sensors is one of the industry’s biggest challenges right now. He points to the Nissan Serena, a minivan that has so-called “first generation” autonomy - i.e. the ability to stay within two white lines on the motorway. If “second and third generation” autonomous vehicles are to become a reality, the cost of sensors industry-wide is going to have to come down.

Going For A Drive

The initial sensation of realising that the car is driving itself is a little freaky, but I’d already felt this when driving a Tesla week prior. Apparently the test drivers have received a few odd glances from lorry drivers, who have spotted the fact that though there is a human there, no one is controlling the wheel.

What was really striking though was as the car approached the first roundabout junction. The car pulled up to the line and stopped. It then waited to see if there is anything coming and smoothly pulled out, following the lines of the roundabout. All by itself. I knew immediately that I was convinced by the technology.

Iijima told me that on the test route he’s only had to manually intervene - such as put his foot on the brake - around once in every three trips. And from what it sounds like this is only when behaving ultra-cautiously - lest Nissan upset TfL or the Metropolitan Police. During the trip I was on, he didn’t have to take control once. For all intents and purposes though, the cars are running just as they would under real conditions. There was no normal car leading or trailing in case there were problems, nor a man with a red flag walking ahead to warn that a new-fangled automobile was coming. The car was on its own.

Nissan's autonomous drive demonstration event – London[/caption]

What was really cool was that the test car had a couple of displays mounted on the dashboard so that I, the passenger, could see what the car was seeing. Every pedestrian on the pavement we passed was flagged up with a yellow blob - and a box appeared around traffic lights to indicate that the car was watching carefully. At one zebra crossing on Freemason’s Way a woman stepped out to cross the road, and the car immediately switched to a large graphic of a crossing, and it showed a yellow graphic crossing the road tracking what she did exactly. The second she had both feet back on the pavement at the other side, the car set off again with faster reactions than I’d probably have behind the wheel.

The biggest challenge comes in terms of mapping. Apparently once up and running maps will dynamically update with the road conditions. So if roadworks block a road or a new junction is added (and so on), passing autonomous vehicles will spot the changes, upload the changes to Nissan’s global map and the changes will be reflected worldwide within half an hour.

Nissan's autonomous drive demonstration event – London[/caption]

But for the time being, Nissan is stuck using this one small test route. This is because it is the only route Nissan has scanned in detail. Building a map of all of London, or all of Britain is going to take much longer, and Iijima calls it a “business issue” rather than a technological one.

Asami agrees, and suggests that mapping could be an area of cooperation across the car industry. “If many OEMs can get together and share the probe data, I think it’s a good thing. Then everyone will be helped to promote this kind of introduction of technologies”, he says. Coming up with a common data format for exchanging this data could also be a crucial point of cooperation, and is something where everyone can win.

Big Money and Big Data

So the technology appears to be coming along rather nicely - but can Nissan make any money out of it? When I asked Ford the same thing at Mobile World Congress, the executives I spoke to essentially argued that it was too early to tell exactly what the car industry of the future will be, and what Ford’s role in it will be. Asami was similar reticent to comment, saying that they are in the “early phase of study”.

I also tried asking my brilliant “gotcha” question that I also put to Ford: All of these new autonomous vehicles will be sharing their location in real time with the car companies. What does this mean for privacy? What if, for example, a Trump-led National Security Agency demanded access to our data just as it has snooped on our internet activity in the past? Would Nissan hand it over? It’s the sort of question that is good for a journalist like me, as if they say they will share, that’s a headline - “NISSAN SAYS IT WILL GIVE YOUR DATA TO TRUMP”. And if they say they won’t, that’s a headline too: “NISSAN SAYS IT WON’T COOPERATE WITH TRUMP”.

Annoyingly for me then, Asami managed to weave between the two positions in such a way to make his answer completely un-newsworthy. When asked if Nissan would cooperate, “That I don’t know”, he said. “First of all we have to comply with the regulation and law of each country. So we have no intention of to violate any law. And also we have a contract with our customers. And when we have telematic service, we promise that we will not release the personal information to any third party - unless we have an official request from the government, for example. So we are trying to be transparent and fair, but it’s a little bit beyond my responsibility to say yes or no, to be frank with you.”

Dammit Nissan.

Still, that’s perhaps an enormous argument for us all to have in 2020. In the meantime, it appears that Nissan is fully focused on getting us there - and on the strength of my brief ride I’m ready to believe that driverless cars are finally, almost here.