We’ve all seen where the power might go out in a section of town, resulting in all four directions of an intersection having the flashing red lights, forcing all directions to treat it as a 4-way stop sign.
Anyone know if there are 100%-reliable mechanisms that keep traffic lights from ever indicating a green light in all directions, leading to sure disaster?
No there aren’t. If the proper sequence of wires short out, the lights will go green in all directions. It’s incredibly unlikely to happen. Modern traffic lights are made to withstand very hazardous conditions, and designed to shut down or go into flashing red mode if a problem is detected. However a shutdown is a more dangerous situation than all lights turning green. When all lights are green, drivers at least know they are approaching an intersection and can use due caution. When the traffic light is turned off, drivers can collide in the intersection without ever having a clue that there was an intersection.
I haven’t. When the power goes out in my town, the lights go out, and there are just dead lights hanging over the street. If there’s no power, how would they flash? (Does just flashing use so much less power so that it can be powered by a battery or something?)
Drivers in Chicago are pretty good about this too. Last summer the traffic lights were out at a busy intersection (Foster and Cicero) and we kind of self directed traffic by alternating one car in each direction east/west then one car in each direction north/south. All while a police officer sat in his squad car watching! :mad:
This is going back a long way in my memory, but in conversation with some engineers who knew a bit about traffic light systems it came up that at least some designs do have a fail safe. Basically they have a “crow bar” circuit on the greens so that if both sets light up, it blows the fuses/breakers on the lights. This renders the traffic lights inoperable, or possibly sets them back to flashing amber mode. But the logic is that anything is safer (including totally inoperative lights) than a situation where crossing traffic both see green.
Adding to the above. A quick search turns up a more sophisticated Conflict Detector. So rather than just blowing the breakers, it acts as a parallel system monitor. It has enough control to bypass the main traffic controller and directly drive the lights.
Traffic lights have a “conflict monitor” separate from the controller. The wires for each light is fed into it and if the controller sends signals that it’s programmed to watch for (like two conflicting greens) the monitor will make the lights go to flashing red.
I think 4 way green would be worse than blackout. During the day, a blackout isn’t too much of a problem because you see intersections and dead lights. At night, 95% of the locals will know what’s going on and avoid you or their actions will tip you off that something’s up. You can see most intersections (especially on major roads) and will look for a stop sign or something, realize the power is out and proceed with caution. If there are cars approaching perpendicular, their headlights will get your attention of it looks like they’re going to intersect you. Not always, but most of the time.
With a 4 way green you know there’s an intersection, ignore everyone else and just blow through at full speed. Not always, but most of the time.
Yes. This is the primary protection, and it works on multiple light systems. I don’t if the crowbar circuit is used outside of single units.
This could be correct. I think there are many more cases of total blackouts which have contributed to accidents because 4 way greens are so unlikely to occur. But it’s fair to assume that many drivers will not approach a green light with caution, and do the opposite, as you have said, blow through the light.
Back in the early days of microprocessors (ca 1974), I worked for a company that was building an intersection computer. A key specification was to implement a fail-safe method not involving the microprocessor to cause flashing red if it detected all green.
Traffic lights were an early commercial success for microprocessors. I’d heard the same concerns you mention. There had been numerous problems from earlier systems with simple problems like frozen solenoids. Keeping multiple light systems synched was also a complex problem. But like any control system, it requires a variety of fail-safe mechanisms to ensure proper operation.
In one of my more interesting liability claims as an Insurco (fake name for a company you have heard of) adjuster, I had a 100% confirmed claim where each driver on intersecting streets was presented with a simultaneous solid green. You can imagine what happened. Multiple witnesses, located on each leg of the intersection, stayed and provided statements to the cop who ultimately placed the proximate cause on the malfunctioning light.
I denied the claimant’s claim as both drivers each had, and simultaneously did not, have the right of way. They met front corner to front corner in the middle of the intersection so neither was able to assert control. Claimant’s insurance co. filed arbitration against us, looking for a 50/50 decision. I successfully defended my 0% decision on the simple fact that there was no proven negligence shown on my driver (and incidentally, my argument was that neither party was able to meet the burden of proof). I wonder if my quip, formally presented but done to make writing my contention at least not unfun, along the lines that society will fall into complete disrepair if citizens of New Jersey can’t trust their traffic folks to ensure safe and proper controls at an intersection, swung the arbitrator.
I’m not engineer enough to know if this is a true fail-safe or not, but here is a 1954 illustration by Marie Neurath, an early pioneer in information design, showing the workings of a mechanical traffic signal. Notice that it’s called: The Lights Which Must Not Make a Mistake.
I worked on a non-microprocessor discrete logic traffic light controller chip in 1976. It read the “car detectors” embedded in the road, and also had external timer control. The way the chip was designed, perpendicular green lights couldn’t happen. I did the logic simulations to test that.
Atlantic City used to have fail-safe lights. They had only three bulbs. If you going in the long direction (NW to SE), you saw a red on top, yellow in the middle and green on the bottom. But if you were going crosstown, you saw the opposite, green on top, red on the bottom. I think you have to go back at least 50 or 60 years to find this. There was simply no way it could show green in both directions. I guess you could see green and red at the same time, but then it would show that way in both directions.
I have been driving for 58 years and I have never seen lights fail in the way the OP described. I’ve seen plenty of situations in which the lights failed.
So apparently the light changed to yellow both between green and red and between red and green. Get ready to make those jack-rabbit starts folks. I think maybe I see why it was changed.