We have all heard of traffic jam, but what about traffic signal pie? This week I will try and explain how traffic signals work at stand-alone crossings (bear with me as I am bound to get something wrong!)
Traffic signal engineering is a whole sub-set of highway engineering and the people who make it work are clever people. Much of the technology is beyond me, but I will do my best to give you some insight over the next few weeks (I am not sure how many posts this will take and if it will be signals every week for a bit yet - yes, making it up as I go along). When talking about "drivers", this is defined in legislation for the purposes of traffic signals and will apply to those using the carriageway - including cyclists! (it is weird, but rules is rules).
Some other useful lingo is that we refer to traffic signals rather than traffic lights. When we talk about each individual "light", we call them "aspects". For example, a basic traffic signal aimed at "drivers" has three aspects; red, amber and green. The posts on which the traffic signals sit are called poles and the cluster of signals is known as a head.
Some other useful lingo is that we refer to traffic signals rather than traffic lights. When we talk about each individual "light", we call them "aspects". For example, a basic traffic signal aimed at "drivers" has three aspects; red, amber and green. The posts on which the traffic signals sit are called poles and the cluster of signals is known as a head.
![]() |
A Pelican crossing (red man way in the distance) and the ubiquitous push button to give pedestrian demand. |
After a bit drivers get an amber signal and then quickly, a red traffic signal. Shortly after pedestrians get a steady "green man" which appears on the traffic signals on the other side of the carriageway. After a while the green man flashes and then just after, drivers get a flashing amber and then a green.
The amber and then red signals to drivers is an absolute instruction to stop. The flashing amber is an instruction to allow people crossing to finish. The green is an indication that drivers may proceed beyond the stop line. The red man is essentially a warning that it is not safe to cross. The steady green man suggests that pedestrians may cross. The flashing green man means that a pedestrian may finish crossing, but shouldn't start to cross if not already started.
A Toucan crossing with far side aspects. Under the red man is the green man and to the right of the green man is the green bike. |
Puffins continue the avian theme and this is shorthand for Pedestrian User-Friendly Intelligent Crossing. The big difference is that the pedestrian aspects are displayed on the push-button (near side aspects) whereas for a Pelican, they are far-sided. The pedestrian aspects are the simple red/ green men. When the button is pressed, a little red rectangle lights up or a ring of red lights around the push button lights up. The signal phasing is the same for a Toucan (no flashing amber/ green).
![]() |
The "Countdown" aspect. This one is actually outside TfL HQ in Southwark, London on a junction, but they are being used at standalone crossings too. |
Pushed by Transport for London (and now with UK-wide authority for use by the Department for Transport), this system adds another pedestrian aspect on Pelican crossings (and junctions which I will cover another time). PC@TS essentially creates a hybrid Pelican where the flashing amber/ green man is replaced. Drivers will get the red, red/ amber, green sequence and pedestrians will see a green man and the the countdown aspect which is an amber signal which counts down the number of seconds in which they have left to cross. There is a very small gap between the end of the countdown and traffic getting a green.
The final type of signalised crossing is a Pegasus which has nothing to do with birds, it is for people on horseback. From memory, there are 4 in London and they are basically Toucans for riders and horses with a high level push button for the rider. They can have far or near sided aspects and red/ green pictures of someone on a horse is used.
I have already mentioned pedestrians and cyclists pushing a button to indicate crossing demand, but what other detection is used with crossings? For traffic, the basic arrangement is the use of microwave vehicle detectors (MVDs) which are the little camera-like boxes on top of the traffic signals. They essentially emit a beam which a moving vehicle causes to reflect back to the unit and is so detected. The other method of detection is magnetic induction loops cut into the carriageway surface. Where general traffic speeds are above 35mph, then additional sets of loops are provided further away from the crossing to track the approach speed over distance which is a safety feature.
For Puffin crossings and "Puffin-style" Toucans (near sided aspects), "people" detection is used. First, sensors can "see" people waiting to cross and so if the demand button is pressed and someone wanders off, the demand can be cancelled, thus not triggering a red signal to traffic. They also have sensors "watching" the crossing area and if there are people taking longer to cross (because there are lots of them, or someone is moving more slowly), then traffic will be held a little longer.
I have no idea how these sensors work and so if you do, then please add a comment! I do know that some of the older kit was a bit temperamental and used to pick up things other than people which messed up the timings. There also used to be pressure mats in the footway to detect someone waiting, but I don't think these are used any more because again, they were temperamental.
![]() |
Not the Vogons as suggested by some on Twitter, but a tactile cone. A little out of focus, but you get the point! |
Second, there is the "tactile cone" which is a little upside down cone placed under the push button of the crossing. The cone rotates where then green man shows and again, is to help blind and partially sighted people, but will be of assistance to anyone with reduced hearing and eyesight.
So, what about the timings and the set-up of a signalised crossing. If you are really interested in the detail there is loads in The Design of Pedestrian Crossings, but I want to keep things a little more simple.
In the diagram, I have amalgamated some of the information from the design guide linked above, but added little pictures of who sees what because I think it is easier to follow.
The reason I am calling these posts "traffic signal pie" is because it is a useful way to think of how time is allocated to who gets which signal. For a Pelican crossing, there are 7 slices which can vary in time, but for any given situation will add up to the whole pie - the whole cycle of the signals. For a pelican crossing, of course, the flashing amber is for the pedestrians, but if they are finished crossing, the drivers can have what is left.
The time ranges for the periods in the diagram depend on site conditions and the design guide gives full details. (A) is all to do with traffic flow, speed and method of detection. If there is no pedestrian demand, the signal to traffic stays green. When the button is pressed, traffic might be stopped quickly if it has had a green for some time. A road with faster traffic (actual speeds are used which is a debate in itself) will have a longer period before stopping so approach speeds can be checked and loops are used.
(B) is fixed and allows a driver to clear the crossing on the amber signal if they are too close to stop. (C) is set to 3 seconds where the speed is above 35mph (read the guidance for more). (D) varies by crossing width (it is in the guidance and for example is set to 4 seconds for a crossing of up to 7.5 metres). The plus 2 is used if it is observed that people turn back thinking they cannot cross in time.
(E) can be used to extend the red signal to drivers if there is concern or evidence that drivers are perhaps bullying their way through on a flashing amber. (F) is normally set to 6 seconds, plus 1 second for each 1.2 metres in crossing width above 6 metres. Periods (D), (E) and (F) are essentially the part of the cycle which campaigners for longer crossing times are targeting. Period (G) is set at 1 second for crossings under 10.5 metres wide and 2 for wider crossings.
The other types of crossing have similar set ups, but obviously without the flashing amber. This is a long post and so I will leave you to look up those in the guidance. One point to mention is that we haven't made out minds up on what is the best form of signalised crossing. In London, TfL did studies on Puffins and Countdown and both were generally positive. In fact, TfL are pushing for Countdown (with far-sided signals) everywhere now. There has been controversy as TfL has used the opportunity to reduce the time given to the green man period.
Personally as a user, I like far-sided signals with Countdown (i.e. no flashing amber) on the basis that I can see when I can cross and I know how long I have left and I hope authorisation is given to allow it at Toucans too. The timing issue is important, but not the point of this post. The issue I have with Puffins is that once you start to cross, you have lost the "comfort" of seeing a signal. The flip side was that Puffins were set up to you look towards traffic when watching for a green man and the detection gave you longer if needed.
A far-sided set up with detection is theoretically possible in that walking speed can be measured during the green man period and a slower person would trigger a higher starting number for Countdown. I don't think anyone is looking at this unfortunately. It is worth mentioning the tricky subject of high speed roads and that is anything 40mph and higher. Detection using loops is absolutely vital for speeds (actual) above 35mph and up 50mph. Beyond that, serious consideration should be given to reducing traffic speeds (see page 3 in the linked document) because of the clear safety risks to pedestrians if a driver ignores the signal (OK, bad enough at lower speeds). Actually, a bridge or tunnel might be a better idea.
That's it for this post. I will debate time allocation in a future post, but hopefully this gives you a start in the tricky issue of signals.
(B) is fixed and allows a driver to clear the crossing on the amber signal if they are too close to stop. (C) is set to 3 seconds where the speed is above 35mph (read the guidance for more). (D) varies by crossing width (it is in the guidance and for example is set to 4 seconds for a crossing of up to 7.5 metres). The plus 2 is used if it is observed that people turn back thinking they cannot cross in time.
(E) can be used to extend the red signal to drivers if there is concern or evidence that drivers are perhaps bullying their way through on a flashing amber. (F) is normally set to 6 seconds, plus 1 second for each 1.2 metres in crossing width above 6 metres. Periods (D), (E) and (F) are essentially the part of the cycle which campaigners for longer crossing times are targeting. Period (G) is set at 1 second for crossings under 10.5 metres wide and 2 for wider crossings.
The other types of crossing have similar set ups, but obviously without the flashing amber. This is a long post and so I will leave you to look up those in the guidance. One point to mention is that we haven't made out minds up on what is the best form of signalised crossing. In London, TfL did studies on Puffins and Countdown and both were generally positive. In fact, TfL are pushing for Countdown (with far-sided signals) everywhere now. There has been controversy as TfL has used the opportunity to reduce the time given to the green man period.
Personally as a user, I like far-sided signals with Countdown (i.e. no flashing amber) on the basis that I can see when I can cross and I know how long I have left and I hope authorisation is given to allow it at Toucans too. The timing issue is important, but not the point of this post. The issue I have with Puffins is that once you start to cross, you have lost the "comfort" of seeing a signal. The flip side was that Puffins were set up to you look towards traffic when watching for a green man and the detection gave you longer if needed.
A far-sided set up with detection is theoretically possible in that walking speed can be measured during the green man period and a slower person would trigger a higher starting number for Countdown. I don't think anyone is looking at this unfortunately. It is worth mentioning the tricky subject of high speed roads and that is anything 40mph and higher. Detection using loops is absolutely vital for speeds (actual) above 35mph and up 50mph. Beyond that, serious consideration should be given to reducing traffic speeds (see page 3 in the linked document) because of the clear safety risks to pedestrians if a driver ignores the signal (OK, bad enough at lower speeds). Actually, a bridge or tunnel might be a better idea.
That's it for this post. I will debate time allocation in a future post, but hopefully this gives you a start in the tricky issue of signals.