Some time ago, Richard Hopkins posted the following comment:
I had a great, but simple idea for headlights.
Polarise them. That way, when it’s wet, they won’t reflect off the road surface and dazzle everyone. Oncoming traffic will be able to see the road markings.
I thought I would verify that it works.
Light polarisation is, to me at least, a fairly unintuitive phenomenon. I thought it was worth testing that the water was reflecting light based on its polarisation and not absorbing it and re-emitting polarised light. To this end I shone a torch through a polarising filter and bounced it off a tray of water to hit my eye: as I rotated the polarising filter, the amount of reflected light changed. Having also verified that the torchlight was not polarised to start with, I am happy to accept that Richard’s proposal could theoretically work.
How well it would work depends on the angle at which the light beam hits the water; there is a nice graph on this page. Note though that even at the worst case when the angle of incidence is close to 90º, parallel-polarised light will be reflected less than randomly or elliptically polarised light.
So, this does look rather promising. The power consumption would be at least twice as high, assuming they are to function by generating randomly or elliptically polarised light which is passed through a linear polarising filter to make polarised light. Of course if you replace the conventional bulbs with something that naturally emits linearly polarised light, the power consumption need not go up for that reason.
I haven’t yet convinced myself as to what the effect of these would be in fog: in particular would the polarisation chosen to limit reflection from puddles towards other drivers make the amount of light reflected back towards the driver from fog droplets greater or lesser? It would be very convenient if it solved that problem as well…