Re: Traffic Sign Bord and Lighting

[Jan Hollan <jhollan@amper....muni.cz>]

Shigemi,

> About 3M Diamond Grade retro reflective sheet
> -At 20 degrees (observing from driver) its reflectance is 0.48 with white,
> 0.05 with green

you speak about retroreflective surfaces, giving some ``reflectances''.
I don't know what a numbers like 0.48 for white could possibly mean.  I
can imagine it is 48 cd/(lx*m2) when the angle driver--sign--headlamps is
below 1 degree.

Thanks to your posting I browsed some literature on retroreflective signs.
There is a study of a US governmental agency
   FHWA RD-03-081 Updated Minimum Retroreflectivity Levels for Traffic Signs
   http://www.tfhrc.gov/safety/pubs/03081/
 giving some minimum values for various types of such surfaces which
differ in the sharpness of distribution of luminous intensity
around the backscatter direction.

Excellent slides from a related lecture are at
      Workshop on Nighttime Visibility of Traffic Signs Soliciting
      http://tcd.tamu.edu/Documents/MinRetro/presentations.MD.pdf
 -- they explain the metrics graphically as well.

The basis of all the recommendations are the needed luminances of the sign
surfaces. This is, for me, the most interesting part of the study. For
vast majority of drivers, one, two or three candela per square metre is
enough, and this can be achieved with low headlight beams. Remember that
low beams should not shine there in fact, just the light dispersed on
reflectors and lenses gets to the signs. (I should emphasise that the
study neglects the light reflected from the road, which can produce
stronger illumination of the sign than direct light itself, but as it
results in a larger source--sign--driver angle, the gain from
retroreflection is lower, and a contribution to the sign luminance may be
often negligible, who knows.)

The needed luminances are (for me) the most important message of the
study. Even for stationary lighting there is hardly any need to have a
luminance over 10 cd/m2, 3 cd/m2 is enough mostly (at least another
1 cd/m2 will be added by headlights, with retroreflective signs).

Without any retroreflection, having a perfectly diffusing surface instead
and a white (just 13 per cent absorbing) surface (i.e., giving 0.3
cd/m2/lx), 3 cd/m2 can be achieved by 10 lx. A green surface would need
some 40 lx. With a giant 10 m2 sign, it translates to 400 lm. This can be
produced by a 50 W halogen incandescent bulb in a suitably tailored
luminaire (letting almost no light outside the target). It is imaginable
that with some remaining retroreflection using a distant beamer (LED
maybe) to achieve angle luminaire--sign--eyes below 5 degrees, even much
lower emitted luminous flux would do.

Last, I don't know what you mean by

> -Regulation requires brightness of white at 60cd, green 8cd
> -Average brightness must be more than 20cd

Brightness is measured, if at all, in lm/m2. Candela applies to luminous
intensity (10 m2 at 3 cd/m2 has 30 cd).

Do you mean luminances, i.e.  cd/m2? (They can be written as ``nt'', if
needed, using a symbol for the old name of this very unit, ``nit'' -- this
name was not included into SI.)

Then demanding such high values is a nonsense, for drivers.  Such very
high values would be relevant for visually impaired people who cannot
drive any more. Twice lower values should do even for very old drivers,
but four times lower values for vast majority of real nighttime drivers.
Even such values are impossible to achieve by retroreflection from low
headlights only.

But retroreflection (with low luminances but high contrasts) suffices in
practice. There is an Ohio DOT study recommending it instead of stationary
lighting,

   Unlighted Overhead Guide Sign Feasibility Study
         Authors:  Helmut T. Zwahlen, Andrew Russ, and Sahika Vatan
   http://webce.ent.ohiou.edu/orite/OU%20CE%20-%20ORITE%20-%20Human%20Factors_files/Exec%20Summary%20-%20Abstract/executive%20summaryUOGS.htm

and a subsequent one, based on older drivers

   Field Evaluation of Unlighted Overhead Guide Signs Using Older Drivers
         Authors:  Helmut T. Zwahlen, Andrew Russ, and Sahika Vatan
   http://webce.ent.ohiou.edu/orite/OU%20CE%20-%20ORITE%20-%20Human%20Factors_files/Exec%20Summary%20-%20Abstract/older%20driver%20UOGS.htm

A report on measured luminances of signs with and without stationary
lighting, by the same authors, is

   Photometric Evaluation of Unlighted Overhead Guide Signs
   http://www.ltrc.lsu.edu/TRB_82/TRB2003-000936.pdf

What pleases me: I've written luminance limits into a proposed bill for an
amendment of Clean Air Act here in Czech Republic. They are given
indirectly, through luminous intensities: 500 cd for billboards and signs
over 30 m2 (what implies 17 cd/m2 at most), 300 cd for those over 5 m2
(implies <= 60 nt) and 200 cd for smaller ones. Even ``glaring'' Japanese
recommendations fit easily in these generous limits. If road safety would
be the target (and not just the first small step in any protection of
night environment), for commercial billboards the limits could be halved,
so that they would not be much more conspicuous than messages essential
for safety.

jenik