I live right next to a tunnel which is about a half-mile or a little less than a kilometer long. The tunnel is only wide enough for one lane and one-way traffic. At each end of the tunnel is a stop light (several actually - and still people occasionally run the light!) with signs that say “5 minute red light.”
There are obviously sensors that detect the approach of a car. When I’m driving home late at night, and go through the tunnel, the light turns green for me as I approach.
What puzzles me is: how do we know it’s safe to be in the tunnel? Couldn’t a car still be in the tunnel as I drive up? Then the light would turn green and we’d be left with a potential collision on our hands. What would avoid this?
I’m also curious about what sort of traffic sensors would be used to tell the light that a car was approaching.
Usually there are loops in the road to sense your car. I’m sure you can find plenty of sources of information on how car sensors work, but the jist is that they sense the metal in your car. Drive over one, and they know you are coming.
Let’s say the tunnel is north-south, and you approach from the south. The sensor detects you, and you get a green light. The north side light knows the other light is green, and won’t change for a car approaching from the north.
At this point, it is likely one of two things will allow the north side light to change to green. It is possible there is a sensor to tell when your car has cleared the tunnel. I find this unlikely, as then it would need to keep track of how many cars have passed through.
The more likely idea is that the other light will be allowed to change after a given amount of time since the last car was detected, or after 5 minutes, whichever is sooner.
If you want to test this theory, the next time you pass through the tunnel, emerge on the other side and then approach the tunnel from the other direction. If there is a sensor detecting your exit, the other light will change immediately (assuming no other cars entered the tunnel after you). If it doesn’t change, pull out your watch and see how long it takes to change. Go through the tunnel and do the same thing again. Time it again. I’m guessing the times will be the same.
If you take longer to traverse the tunnel than the time the lots allot you, this test will fail. I’m guessing you will have more than enough time, though.
In my experience with a single lane bridge, the reason the light can turn green for you immediately as you approach is that as long as there hasn’t been traffic for awhile, the lights on both ends turn red and stay that way. Then, the first approaching car to be detected automatically gets a green light, because the signal knows all (non-existent) traffic has been stopped for the last half hour.
When traffic is heavier, long green lights in one direction are followed by mutual reds sufficiently long to clear traffic from the bridge, followed by green lights in the other direction.
I wouldn’t be so sure. They certainly have an incentive to know that the tunnel is clear, because when a school bus hits an armored car…
It’s pretty easy to tell which direction a car is going with pretty dumb sensors. They could have sensors at each end of the tunnel and have those and the lights connected to a central processor to coordinate the lights. Then the sensors can count how many cars enter and how many exit in each end. If a car stalls, or throws it into reverse and backs out the way it came in, then the sensors would be able to tell that.
I counted cars for three years. You know those rubber hoses in the road? Well, if you have two six feet apart connected to a good automatic traffic recorder, it can count two lanes of traffic (going in opposite directions) and tell you the speed, direction, time of passing and class (i.e. truck or motorcycle or pick-up w/ a trailer, etc.) with a high degree of accuracy on a fairly busy road.
That’s not to say that such precautions are taken at the tunnel in question. However, it would be pretty easy to do–given the equipment and expertise available.
Finally I can apply some knowledge I attained interning with the Ohio Dept of Transportation last summer in the traffic office. The detector loops are really simple:
You put a loop of wire in the pavement (cut a slot in the road, lay down your wire, seal it up). This wire acts as any coil of wire does, and produces a small current when you pass a magnetic object over it (such as a 2500 pound iron car). The loop is connected to a computer either on a nearby electrical pole, or in a traffic box close to the signal. In that box, there is a signal amplifier, a loop detector, a computer to keep track of when a loop is triggered and control the signal itself, and a backup computer in case the first one fails. There is a traffic engineers handbook that has all the standard signal timings for standard intersections, and I would bet that somewhere in it there is a section on how to time a one way tunnel. Just program the timing into the signal computer, and it does all the work.
Thank you all for your very helpful answers! I’m going to try the little experiment proposed by wagnoid.
There are bike lanes on both sides of the tunnel, but you’d be really hard-pressed to manuver two cars alongside each other. If there’s a bus in there (we often have school buses, delivery trucks, etc), then it’s hopeless.
So if a car died or if there were an accident in the tunnel, the signals wouldn’t know and would keep operating as usual?
They might be able to sense it and adjust for that, if they’ve done the proper engineering. Suppose the tunnel has two entrances (duh), X and Y. X has the green light. A sensor at X can tell how many cars pass over it and which direction they were going. So if five cars enter the tunnel at X, that is logged into the computer. Now, one of the cars stalls. The sensor at Y can tell how many cars pass and which direction they’re going. It counts four cars exiting at Y. Now it knows that there is one car unaccounted for, and that it was travelling in the X->Y direction.
So a car could stall with the lights & sensors system knowing that something is wrong.
Suppose that instead of stalling, the car stops and reverses out of the tunnel. Now the sensor at Y detects 4 cars exiting at Y, while the sensor at X detects one car exiting at X, and then concludes that the tunnel is clear. It can now give the Y signal the green light.
That’s not to say that such a system is set up. Call the city/county engineer and report back. You’ve gotten me really curious.
