John Rowland wrote:
"Tom Anderson" wrote in message
.li...

All that said, isn't the real solution to make the lights
(or rather, the junction) sensor-controlled, or perhaps
better-sensor-controlled? If the junction knew there
was a queue of cars waiting to go one way, and cars
were only a few a minute the other way, it could just
change its lights to let them through.

I think bicycles fail to activate the sensors, which is one reason
why cyclists habitually jump red lights. Why are these sensors so
rare anyway? Are they very expensive?

Good question. Back in the dark ages, when traffic levels were low,
many if not all traffic lights were controlled by rubber detector strips
set into metal frames on the road surface. Nowadays you can have either
movement detectors mounted on the lights, using presumably the same
cheap technology as in burglar alarm PIRs, or sensor wires embedded in
the top layer of the road surface. But most lights now seem to work on
a fixed time sequence.

When I lived just outside Reading, a busy cross roads near us was the
subject of Transport Research Laboratory investigations into different
phasing and sensor strategies. The final system was excellent in
responding very quickly to the actual traffic levels on each road. If
all the traffic queue in the green direction was cleared, the lights
immediately switched to the next phase. At quiet periods at night, all
lights were set to red, and an approaching vehicle would immediately
trigger a red+amber/green sequence for that direction. That was 15-20
years ago.

I'm sure the technology is cheap, and it shouldn't cause any more
disruption to install it as laying anti-skid coatings around signalled
junctions.

We also seem to have gone backwards on area control of a set of lights.
Do these schemes still exist in London? I keep being stopped,
particularly at light-controlled pedestrian crossings, in a way that sug
gests that each set of lights functions independently.
--
Richard J.
(to e-mail me, swap uk and yon in address)