Why does fog cloud driver judgement?

What happens when you drive in fog? Often low visibility is blamed for accidents in these conditions, but how fast are we really driving?
16 June 2013

Interview with 

Paolo Pretto, Max Planck, Tubingen




What happens when you drive in fog?  Paolo Pretto from the Max Planck Institute for Biological Cybernetics in Tubingen, Germany,  clarified a few issues related to driver's perceptions of their speed in foggy conditions. He explains to Martha Henriques...

Paolo -   So, there are a lot of accidents in fog, right?  We blame low visibility condition and there were a couple of papers in the past that just said we can't explain why this happens and they provided an explanation which was quite comfortable when you drive in fog, because fog is simply a natural way to reduce visual contrast.  And we know that a low visual contrast leads to lower perceived speed then what happens to the driver is they perceive their own speed being lower and therefore, you accelerate to match what the actual speed will look like in clear visibility conditions.  So, what this paper did was they brought this to a situation of self-motion.  So, when there is no object moving, but the observer itself is moving in the environment.  So, they considered this for the first time. And, probably, the way they reduce their contrast was the same way that used in the laboratory conditions.  So, they reduced the contrast of the visual seeing of the driver in a driving simulation uniformly, independently of depth, independently of the distance from the observer.  And so, that's what was missing in the previous study and that's what we did in our new study.

Martha -   And in those previous studies, their simulations were more like driving with a misted up or dirty windscreen whereas you adjusted contrast so that it seemed more like actually looking at fog outside through a clear windscreen.  Why is that an important difference and how does that affect human perception of speed when driving?

Paolo -   The difference is quite simple.  So, no one ever tried before to simulate a gradient of contrast reduction.  So, a contrast which is not reduced uniformly all over the visual field.  But different regions have different contrast reductions and that's basically the new part.

Martha -   So, you created a more realistic simulation of driving in foggy conditions and the initial models used were essentially just too simple and you've added a few more factors to make it...

Paolo -   So, we're not appropriate.  We're simulating a realistic fog.

Martha -   Once you have this better idea of the way fogs should be modelled, what happened when you got some drivers into these simulators, the testing?

Paolo -   Well, we used different methods.  We did some study on the perceptual side.  So, studies which are basically passive for our subjects so they don't have to do a driving path.  But with this kind of methodologies, we can somehow quantify their perception.  We can relate their internal representation of their speed to the actual current physical speed that they're driving at.  So, we measured what is technically called the point of subjective equality.  So, when two different speeds are met to different scenes, moving scenes.  With that, we derive the perceived velocity, the perceived speed of the driver.  And now, the method that we use was more like a behavioural neuroscience method.  So, we asked our participants, our subjects to drive, to do some driving tasks and by trying to match a given target speed.  They were trained to reproduce before the real experiment.  With that long, we could measure the other side.  So, we could measure the production of speed.  By that, we could relay the effect of perceived speed that we measured in the psychophysical experiments with the observable changes in the behaviour of the drivers that we measured in the second experiment.

Martha -   So, you found that the perceived speed was significantly higher than the speed produced by the drivers in the simulator.  Meaning, they would actually tend to drive more slowly in foggier conditions.  So in fact, our instincts seem to be telling us to do the sensible thing and slow down.  So in fact, speed perception doesn't seem to be a factor that would make us more prone to traffic accidents in fog after all.

Paolo -   What we did in our study was to concentrate on the perceptual effects, on the visual motion that was processed by the drivers.  But we know that when we drive that it's not only perception going on.  We have also our own motivations.  For example, if we are in a hurry, if we have an appointment, we will speed up.  We will drive faster.  So, we simply will ignore what our perception, what our brain will suggest that.  there are a lot of higher order cognitive factors that overcome the perceptual effects.  If we try to simplify and we try to analyse separately the perception, that's what we found.  So, the brain would kind of suggest that to drive safely.  We don't do it obviously.

Martha -   And so, how do you think these findings might influence say, road safety campaigns or policy?  What do you think might be the practical applications that your work could contribute to?

Paolo -   At the time we conceived the study, we were not really concerned about practical applications.  But after seeing the results, we actually realised that this could be kind of useful for many different fields.  For example, we know that in the driving simulation community, there is always a problem, and the problem is that in the simulation, all the drivers over estimate their own speed.  With our findings, we could think of some application in the future where playing around with the contrast of the visual scene, we could somehow adjust more or less efficiently towards higher or lower levels of perceived speed in the simulation.


Add a comment