This article is more than 1 year old

I survived a head-on crash with driverless cars – and dummies

Drove to the Transport Research Lab, ended in France

Geek's Guide I’m driving along the French Riviera, and it’s a challenge. It’s a manual car, but for some reason I can’t hear the engine, which after ten years driving an automatic makes it difficult knowing when to switch gears.

Suddenly the pedals start moving themselves, the steering wheel gyrating wildly as it avoids oncoming pedestrians while making sharp turns onto the picturesque wharf of this typically picturesque town.

Welcome to the world of autonomous vehicles. And welcome to Wokingham. Because you’ll be relieved to hear that rather than this sometime driver being unleashed onto unsuspecting paysans, I’m actually in a simulator in the bowels of the UK’s Transport Research Laboratory, the home of transport innovations as varied as the bouncing bomb (from its previous incarnation as the Road Research Laboratory) and the mini-roundabout – though which has caused more human misery is open to debate.

Does that summon up visions of chaps in brown lab coats testing samples of road? You wouldn’t be wrong. They do that there. But they also conduct research into electric vehicles, and are leading the UK’s first trials into driverless cars, in Greenwich.

I visited on a typical summer’s day in August; grey, and drizzly, and seemingly colder the closer I got to the mini-roundabout at the entrance to the car park. Yes, they’re justifiably proud of their achievements here, though I didn’t actually see any bouncing bombs.

Child sized crash test dummies at the TRL

You don't have to be a dummy to work here...

As I got out of the car, it was hard not to peruse the other vehicles in the car park. They were a mix, just like any other office in the South of England, but with an higher-than-usual smattering of hi vis markings – the TRL operates vehicles that inspect the road surfaces of the UK, as well as providing accident analysis for the courts and other agencies.

Sadly, I didn’t see any vehicles operating under their own control

The first people I met were [upcoming Reg lecturer] Professor Nick Reed, the TRL’s Academy Director and Dr Alan Stevens, research director for transportation.

If you’ve paid any attention to the development of automated vehicles, you’ll have twigged that firstly there’s an awful lot of hype, and that once you start to think about it from a technology point of view, that the issues of integration, data handling comms, and back-end analytics seem sure to be problems that will inevitably be a drag in the real world rollout.

But, Reed and Stevens were both reassuringly un-evangelical. This is the UK TRL after all. And so after your writer bemoaned the potential headaches of automated vehicles and integrated traffic systems, Stevens gently pointed out, “but we’ve done it already. We did it at the Olympics".

“Because we do the core software (for the traffic lights system) one of the things we did was essentially publish the traffic light timings real time, to the web.” This meant the vehicles ferrying the athletes along those cursed special lanes could pickup the data, and pace themselves accordingly.

In fact, both Reed and Stevens are clear that we are evidently on the brink of widespread autonomous driving; not necessarily driverless cars, but certainly self-driving cars. As Stevens puts it, “Some of it is now, some of it is 20 years, some of it is never".

With motorways, we are pretty close, less than five years, he suggests, including platooning. Stop and start with interactive traffic lights, that’s five to ten. However, no-one seems willing to commit on urban and rural driving. Or, looked up at another way, "up to 20 and over 50 can be done".

The key roadblocks are less to do with technology, and more with us, the humans. OK, humans, and lawyers and insurance types.

Taking the latter first, Reed said: “The different situations you might encounter are essentially infinite – route, traffic, weather, pedestrians, dogs, plastic bags. Then you’ve got the legal issues around operation of these vehicles and who’s responsible.”

“If a collision occurs there should be process to interrogate that data and liability determined in a more objective manner than it is currently because you know exactly what the vehicle was doing and how it acted," said Reed.

As well as “interrogating” the driver, investigators will be able to interrogate the data, ascertain if the code and sensors worked as intended, making pinpointing fault a little more objective. “I’m not going to say it's easy [but] a surmountable challenge," he added.

As for the rest of us, it’s a matter of getting humans to trust the technology, and then, to not over-trust the technology. “At the moment people’s reaction is based on what they’ve seen on the news, essentially ‘I’m not going to be in control? No, I don’t want it’ is an instant reaction,” said Reed.

Then, adds Stevens: “Typically, people won’t believe the tech will work until they’ve tried it. It was the same with adaptive cruise control. They you sit them in the car, and they say ‘wow, it works.”

At that point, Reed suggests that people will adapt to the technology very quickly, as they have with every other motoring advance.

More about

TIP US OFF

Send us news


Other stories you might like