An autonomous Nissan Leaf that drives like a real person: it is a reality and we have ridden in it

Until a couple of days ago, I thought autonomous driving was something of the future. But a future so distant that he could not be seen or appear. And, to be honest, it’s not like I was very interested in the subject either: I’m more into testing cars and driving myself.

But something has changed in my perception of the subject. And the fault lies with a Nissan Leaf prototype that is part of the ServCity project, the most serious attempt that I have been able to meet so far.

What’s more: if it will be serious that the project participates, apart from Nissan, Hitachi, which is in charge of the development of artificial intelligence and other partners such as Catapulta (geographic information), TRL (for the integration of all systems), SBD ( development of a mobile app) and the University of Nottingham, which is in charge of studying experiences and reactions with pedestrians.

The figures of the project are also important: in three years (they began in a pandemic, where field tests could not be carried out) they have traveled 2,575 km of validation, acquired 5.45 million gigabytes of driving data and spent 10.7 million pounds (about 12 million euros) in its development. And that until December, because the project is currently going ahead.

Not just a car: also an infrastructure

During these three years, the 116 people involved have developed a network that helps the car drive itself and make “smart” decisions. But let’s go in parts.

During this “non-test” around Greenwich, an area east of London, I sit in the back seats. But I am not going alone: ​​behind the wheel is a driver “just in case”, who keeps his hands close to the wheel all the time, but without touching it. And also Nirav Shah, an engineer who explains to me what is happening.

Fast version: The Leaf starts up as soon as the “non-driver” (sorry I don’t remember his name) presses the Pro Pilot button on the steering wheel and starts driving like a normal car. Traffic lights, pedestrians, crazy vans, a motorbike, bikes… The chaos of analog circulation cannot cope with the virtual and low system of the car, simply hallucinating.

Medium version: the four LIDARs, nine cameras, a radar, a V2I antenna (which connects with external infrastructure), six electronic control units for autonomous driving and another two for managing the car itself, are in charge of taking me on a tour of about 5.5 km surrounded by real traffic.

And that real traffic includes a traffic light that turns amber, a pedestrian who crosses the wrong way, a van that suddenly brakes and parks invading the lane in which we are driving, a motorbike zigzagging and a bike in a small and narrow roundabout. .

Again: the chaos of analog circulation can not with the virtual and low system of the car simply hallucinated.

Decide what you have to do naturally

And now the extended version. You already know this: the four LIDARs, nine cameras, a radar, a V2L antenna, six electronic control units for autonomous driving and another two for managing the car itself, are in charge of taking me through this journey of about 5, 5 km surrounded by real traffic.

I’m going to go around that ‘loop’ twice. Well, “they’re going to give me two laps.” The first is the car that circulates by itself. And what you just read happens. Take decisions. For example, we are approaching a traffic light that turns amber and the Leaf stops.

And what happens if you are too close to brake, for example? “The system calculates if he has time to stop. If he can’t, he keeps going like any driver would,” Nirav assures me. I think that this would need a little more development in the cities of Spain…

I hoped that the autonomous car would remain in the unperturbed lane, in the middle, whoever fell fell. Come on, like many of those who go through the middle on the highway, but with no one to look at with rancor.

Arriving at the van, the self-driving car slows down a bit, separates a bit, and does what anyone would do, circle it carefully. “Half lane change,” they tell me, though it doesn’t have to be that way: it just leaves the safety space it’s programmed for.

The hallucination of this type of black magic is maintained when I see how it moves naturally in its own lane: if a motorcycle passes, it moves away a bit. A pedestrian crosses where he shouldn’t. The control screen marks a red rectangle and the Leaf downshifts a bit.

Nothing I wasn’t going to do myself, I think. The same happens at the entrance of a roundabout that is a bit complicated because it has two lanes, buses circulate and it is small, which requires a tight turn. The system seems somewhat hesitant to join the circulation, but it enters without problem when it detects a gap and continues as if nothing had happened.

The good part: coupled with infrastructure, go a little further

Before I mentioned that I was going to do two laps of the circuit. In the second, the infrastructure created by the SMLL (Smart Mobility Living Lab) is used, which consists of 270 cameras in a circuit of just over 30 km, although in my case there will be 12 that watch over my integrity.

They are located at traffic lights and streetlights and have sensors that detect objects, process, classify and locate them down to the centimeter (there are 500,000 lines of code): stopped buses about to join, but we still can’t see them. People crossing at a traffic light… All this reaches the car, which processes it and makes decisions.

Like, for example, changing lanes. Nirav Shah makes me notice that we are going to go through that bus zone. The car, a couple of hundred meters before, before even driving down the street in question, changes lanes safely and smoothly when it sees a gap.

Nothing forced, and as if the “non-driver” is the one actually driving. I reach the end of the journey and get out of the car impressed. As I said previously: the chaos of analog circulation cannot cope with the virtual and low system of the car, simply hallucinated

David Moss, Senior Vice President of R&D at Nissan Europe walks up to me and excitedly asks, “What’s up?” Incredible. He tells me that despite everything it is a level 4 of autonomous driving (check here the six levels of autonomous driving).

That means that it is “highly automated”, but not completely: for example, in the event of a pothole in the asphalt, the driver should correct the trajectory… although little else. Without a doubt, the future is amazing. Not necessarily better or more fun, at least when it comes to driving.

But amazing yes. And a lot…