[Mike Garrard]

Mike Garrard asked the Naked Scientists:
   
Just listening to Dave and Meera's piece on cats eyes.  Driving along the A12 I think between Chelmsford and Colchester, I noticed the cats eyes were leaving dot trails as they passed just out of my periphery vision.  I spent a while trying to figure out what was discrete about the process: my eyes (not raster scan I think , the headlamps (tungsten, continuous), the retro-reflectors, and couldn't come up with a solution.  I was going to ask yourselves, but a bit of internet research turned up LED cats eyes.            

http://cleantechnica.com/2009/01/17/luna-road-unveils-solar-powered-led-cat-eyes/
 
The advantage is claimed as continuous illumination, along with more intelligent features such as turning blue for sub-zero temperatures.  
 
Regards,
 
Mike Garrard

What do you think?

[graham.d]

Yes, the A12 has active (rather than passive) cats eyes. At least in places.

[SeanB]

The LED units inside give high brightness by being driven at high current. This would both made the battery flat in seconds and burn out the chip if the current was not delivered as a very short pulse, with a longer time between pulses to enable the average heating to be within the limits of the chip.

The short pulses are integrated by the eye to appear as a bright spot that is always on, at least in the center of your visual field. At the edges of your vision the eye is much more sensitive to motion, and as such the individual pulses are visible with the car in motion, with the eye yet again integrating the result to appear as a row of dots, with a spacing dependent on both your speed and the distance between you and the emitter, in place of separate pulses.

[RD]

At the edges of your vision the eye is much more sensitive to motion

Isn't this phenomenon a consequence of the geometry of the situation rather a peculiarity of the eye ? :
The angular motion of an object passing (at constant speed) is greatest when it is closest to the observer,
the LEDs which are closest to the driver are alongside the car, i.e. those in peripheral vision when driver is looking ahead.
 
The stroboscopic nature of the LEDs only becoming apparent at the highest angular velocity, (when they just happen to be in peripheral vision in this example as the driver is thankfully looking where he is going. Passangers in the car could safely observe the phenomenon in the centre of their vision by looking out of a side window, provided the car was travelling fast enough).

[graham.d]

RD, I believe Sean is right about the periphery of the eye being more sensitive to motion. I think it has evolved this way as it is an importent survival technique to be able to sense movement away from the central field of view, in order to alert you to a predator for example.

http://en.wikipedia.org/wiki/Peripheral_vision

[RD]

I stand corrected ...

... flicker can be sensed in peripheral vision at higher frequencies than in foveal vision.

http://en.wikipedia.org/wiki/Flicker_fusion_threshold

[SeanB]

The ones installed here ( 5 years or so now and still around 90% still surviving, they are really tough units) only show flicker when passing them with the car, though I can sense the flicker from straight ahead  as well, but it is most pronounced from the corner of your eye.

[Geezer]

The ones installed here ( 5 years or so now and still around 90% still surviving, they are really tough units) only show flicker when passing them with the car, though I can sense the flicker from straight ahead  as well, but it is most pronounced from the corner of your eye.

If they used them here, they'd last about ten minutes  []

I suspect the snowplows would make short work of them.

[SeanB]

These are buried into the roadbed, and have no projecting parts, just a flat really tough glass top with capscrews holding the unit in the socket buried into the roadway. Trucks regularly drive over them with no problems to the units, and in a car they make no noise. I think a snowplough will break something on itself before it damages these, as nothing is higher than the road. I think they would be difficult to install on a concrete roadway, as you need to core out a hole to install the socket. These had a section removed for road repairs, and they were reinstated after the roadway was milled, reformulated and relaid.

[Geezer]

You could be right, although they actually bury the lane marker lines in recesses so they are below the road surface, and the plows still manage to rip 'em out!

[Mike G]

Whilst I agree that non-fovial vision is more sensitive, I'd have to say I agree with RD that geometry was the dominant effect.  What I noticed was not flickering, but the string of dots, which allowed me to deduce they were flashing.

How does the energy balance work out?  The mark/space of the dots suggests maybe 5% lit time although that's a bit of a guess.  I've got a couple of those garden lights, with solar cells about 3cm square, and a good day's sun lets them feebly light (say 5mA) a single 0.2in LED for about three of hours [edit - perhaps that's battery limit not indcident energy.  But in winter they quit earlier].  The website I linked to makes it look as if the illumination is all night, not just when headlights are detected.

Q2: What non-linearity makes the pulsing worthwhile?  If it were percieved illumination then even the simplest of LED torches would pulse, let alone the high tech bike lights.  The same applies to battery life: if pulsing helped it would be used.  Pulsing makes sense for say an IR remote control because the receiver can detect and respond to short pulses, so you can pulse 1A through a 30mA continuous LED and increase range.  If the eye simply integrates the numbers of photons then pulsing an average 30mA would look the same as a continuous 30mA.

Mike G

[Geezer]

I found this. Perhaps more than you really wanted to know. Also, this information seems to be about ten years old, but there are a lot of good points that were news to me.

http://members.misty.com/don/ledp.html

[Mike G]

Thanks. This supports my observations that pulsing does not help the perception of brightness.  He does say that human vision is non-linear but does not elucidate.  I am sure there is a good engineering reason why LED cats eyes pulse and bicycle headlamps don't.

I drove Chelmsford to Colchester recently and did not observe LED characteristics so I think I have the location on the A12 wrong.  Anyone know specifically?

[RD]

Q2: What non-linearity makes the pulsing worthwhile? 

Pulsing does allow lights to be dimmed by varying the "duty cycle".

Perhaps the LEDs dim when it is darkest to avoid their glare blinding drivers, and/or to save energy.

BTW ...

BBC reported that the devices, which flash almost imperceptibly at 100 times a second

http://en.wikipedia.org/wiki/Cat%27s_eye_%28road%29#United_Kingdom