2025AD Homepage

A tough act to follow: the excellence of human drivers

Difficult driving conditions - but we still perform (Photo: Rocky89 / iStock)

Article Publication Meta Data

Gareth Watson
Gareth Watson
Show Author Information

Article Interactions

7 shares
Rate this article on a scale of 1 to 5.
5 4 votes
0 comments
comment
0 views

We like to have machines do things for us – but only if they perform better than we do. Therein lies the real challenge for automated cars. Humans are good at driving. In fact, we are much better than we think we are, as Rob Gray, Associate Professor of Human Systems Engineering at Arizona State University, reveals.

Professor Rob Gray: driving expert

We hear much about when driving goes wrong but little about when it goes right. Think about what a driving scenario entails: traveling at over 70 mph (110 km/h); holding a conversation with a passenger; music playing; cars whose movements you can't predict are surrounding you; you're listening religiously to your GPS while also keeping an eye on the screen, the road and exit signs, yet you successfully reach your destination. How exactly do we manage this? Professor Rob Gray, an expert in the psychology of perception and driving was kind enough to elaborate.

"I think we can take for granted some of our abilities in driving. We do so many things almost automatically that we forget just how complex it is. That is until you try and teach someone - then you get a handle on it!" he says.

Playing to our strengths

"Where we're really good, and we don't realize it, is hazard perception. It will be difficult for automated vehicles to match us on this." Professor Gray explains. "When we approach an intersection or merge onto a highway, we are able to pick up very subtle cues about what somebody else might do. These are small pieces of information that we recognize as potential hazards." To suggest we have a sixth sense for this is maybe a little unscientific, but it is often difficult to put your finger on what exactly gives something away - as Professor Gray points out with an example we can all relate to:

"We can almost tell on the road when someone is on their phone. There are just very subtle changes in their driving behavior that give it away".

Subtle cues tell us when a driver is on their phone (Photo: splitshire.com)

So we are good at anticipating situations - but what about the actual driving? It turns out we are pretty remarkable there too: "Another area where we are exceptional is steering in difficult conditions," Professor Gray explains. We've all been there: dark; driving sleet; glare from on-coming headlights and snow covered lane markings. Conditions that will heavily challenge even the best sensor system of an automated vehicle. "In the worst conditions we are able to guide the car around bends successfully. Our visual perceptual system allows us to 'fill in the gaps' where information is lacking."

Hardwired to perform

This perceptual system of ours seems pretty impressive - and it is! But what makes it so good? The hardware for one.

"Our peripheral vision system is amazing at picking up the motion of objects passing by like the road edges - we're very sensitive to that. And it doesn't need bright or clear images, it performs in poor conditions." Professor Gray explains.

Recognizing a situation is one thing, doing something about it is another. We must react to what we see. Luckily we are pretty well equipped in this department too.

"Our nervous system allows us to react very quickly to stimuli. Think of when you veer off-lane slightly and hit the rumble strips. That vibration causes an almost reflexive reaction to adjust the steering. There is no interpretation involved. It's almost as if we bypass the brain."

When accidents happen, is it our hardware failing us?

For a car accident to happen, something has to go wrong. As it turns out, it is rarely the perceptual system. Perception is only one part of driving; attention is the other.

"By far, research and accident reports suggest that causes are much more attentional. The perceptual information is there, people should have seen it. Attention and distraction are the problems."

So when it comes to the driving itself, machines are hardly a match for us yet. Our problem is maintaining attention. As Professor Gray explains: "Considering all the driving being done every second across the world, the rate at which there is a fault is actually so low that we let our minds wander. That's why we are really bad at keeping an attentional focus." It doesn't seem to be a case of are we good enough at driving but rather, are we too good?

Truly exceptional cases

If bettering your average driver will be a tough task for automated driving, then bettering the exceptional ones will be even tougher.

"Think about a formula one driver pushing a vehicle to the limits. I find it hard to believe that you could ever program a machine to perform to this level, never mind avoid a crash. You could simply never predict it given those speeds. When those drivers escape serious injury by pulling some maneuver, they are doing so with instinct. Pure instinct."

Acting instinctively: a tough ask for machines (Photo: Fotolia)

Although not a driving example, Professor Gray points to the Hudson River plane landing in 2009 to illustrate our prowess. "Can you see a machine ever matching us for coming up with creative solutions like that? Our ability to see opportunities for action is exceptional. I can't imagine what an automated system would have done in that scenario."

Moving up another gear

In an ideal world, where we could maintain attention at all times, it seems our innate abilities would serve us well. Unfortunately we can't. So the more we can hand over the tedious parts of driving to an automated system, the better chance we have to maintain attention when it really matters. So, until the driverless car is actually here, the technologies leading to it will be a great support to our already impressive driving skill set!

About our expert:

Dr Rob Gray has worked in the automotive industry as a research scientist before being appointed as an Assistant Professor in the University's newly formed Applied Psychology Program. His research focuses on perceptual-motor control with a particular emphasis on driving. He applies his research to the improvement of training, simulation, accident prevention and human-machine interface development. He discusses these issues on a podcast he hosts and produces, The Perception & Action Podcast which can be found at perceptionaction.com

Do you consider yourself a tough act to follow when it comes to driving? Have machines got their work cut out?

Article Interactions

7 shares
5 4 votes
Rate this article on a scale of 1 to 5.

Article Publication Meta Data

Gareth Watson
Gareth Watson
Show Author Information

Related Content