Nissan Teaching Cars Human Behaviour

It was an enlightening performance today at CES in Las Vegas from Dr Melissa Cefkin who’s the Principle Researcher, Design Anthropologist for Nissan. She’s based in Silicon Valley and is the go to person when it comes to how autonomous cars of the future will operate in the real world.

She explained the almost infinite complexities when it comes to how such vehicles will operate in even the most basic of scenarios. As you’d expect this is going to be a real minefield for driverless cars to become socially acceptable.

Dr Cefkin presented a series of slideshows showing just how complex even simple intersections can be.  It’s not until this kind of scenario is broken down that you realise just how far off this way of driving is. Take for example a four-way intersection, you have pedestrians, other cars turning different directions, drivers being more assertive than others, pedestrians ‘piggy backing’ onto the back of others as they try to beat the crossing light. People in general at an intersection are performing socially acceptable acts but are often kind of stretching the rules. To the autonomous car this is a real headache.

On another note think about how humans utilise hand gestures or other body movements as they go about their daily drive or stroll on the footpath. How many times have you waved a pedestrian across the road because they left their run a bit late and you were in a generous mood? Or perhaps just a nod of the head has made another driver aware that you’re extending another courtesy because they may be stuck in the middle of an intersection. Then there are the jerks and cowboys, how can an autonomous car deal with that?

Then there are regional and cultural differences in driver behaviour. A mad house roundabout in Iran with multiple lanes and pedestrians going every which way is completely different to a narrow laneway in Japan.

The challenge for Nissan and all manufacturers is how to replicate the human brain’s ability to sense, perceive, decide and then act.

So this is where, as Nissan puts it, the “rocket science of autonomy” comes in. Artificial Intelligence is going to bring about teamwork between humans and robots. Traffic flow needs to be maintained, there needs to a level of confidence between man and machine and the cars need to act in accordance with the local driving culture. Put simply obstacle avoidance is the number 1 priority.

Nissan has been working on new forms of signalling on the outside of the car, way beyond indicators and brake lights. There may be a message board or what they call “intention indicators” with a colour spectrum that indicates different motives. For example a strip of light down the side may acknowledge that the car has seen the cyclist alongside.

The car of the future that I still believe is many, many years away will be a complete rethink in automotive design inside and out. Nissan, like the others, is working at a furious pace to make this happen but the science behind it is frankly mind boggling.

[ces]