Sitting
outside under a clear blue sky with rather too large an area of pink collectors
of solar UV laid out to supply energy for my vitamin D manufactories and aiding
my digestion of sugars, starch and a smidge of protein with sufficient ethanol
to aid the solubilization of lipids, I noticed how one could see for ever. At
least until the horizon haze introduces ambiguity.
All this is very different from the winter
scene with large rain drops or giant snow flakes leisurely falling and limiting
your vision as drive to the late night pharmacy for a cold cure.
The
problem with driving at night is that things in the air scatter the valuable
light from your beams sending a lot back to you and preventing the remainder
backscattering from objects in your path. The amount of light scattered is the
product of the size and number density of those pesky things in the air in
front of you. Large drifting snowflakes, for example, shut you off in a small
visual cocoon.
Now
Discovery News reviews a promising attack plan published by Carnegie-Mellon
(1,2). The work is from Narasimhan’s group and is providing a multi-beam smart
solution. The essence is that the headlight projects light beams and receives
images of backscattering from pesky things like snowflakes. This is a smart
solution remember, so a computer assesses the rate of fall of the flake and
projects other beams above or below it.
This
works fine with one or a few drops and simulation of higher numbers look very
promising for extending the range of vision when the skies are falling. Currently the response time is 13
milliseconds and that has to be speeded up as the larger number of bits falling
from the sky, the closer they are together and the shorter the time available
to process and flip the beam.
Hey,
who knows, maybe we’ll have smart flashlights connected to our smartphones.
- http://news.discovery.com/autos/smart-headlights-120705.html
- http://www.cs.cmu.edu/%7EILIM/projects/IL/smartHeadlight/
