What is the cost of implementing synchronized traffic lights in a mid-sized city?
10.06.2025 12:22
Let's add $10,000 to half the signals for upgrading sensors. $5M.
But you know who has to stop at every light? ….The people travelling perpendicular to that one way street. On a standard city grid, it is mathematically impossible to coordinate wonderful “green bands” (technical term for coordination because we actually are more interested in getting platoons through on the whole green not just one car) along the East-West one-way streets, and the North-South streets at the same time. So all of our money spent on “synchronization” still ensures that people not traveling on these perfect ideal East-West one way routes will encounter red lights at every single intersection. Optimization does not mean you won't encounter red lights. It depends which way you are travelling.
These are rough numbers off the top of my head, it's been awhile since I've done such high level capital planning, so my numbers might be out of date, but they’ll give you a rough ballpark and some insight into the cost estimation process. For simplicity I'm going to assume you mean capital costs and not operating or maintenance cost.
If you can't tell, I'm a traffic nerd. I'm not trying to mock your question, I'm genuinely enjoying working through the hypothetical question. I hope my answer was educational and interesting to read.
In real life, a well run city will only upgrade and synchronize one corridor at a time. They spend money only on the upgrades that will provide the most bang for their buck, in a local area, where there are the most problems. They will only do traffic studies, when there are problems in an area. They will only reprogram and optimize the synchronization of a corridor on an as-needed basis, to save on time and cost.
This isn't actually a great way to do it, though, because clocks drift. And even a small drift in the internal clock will cause loss of coordination. You programmed in a 26.5 second offset (i.e. this light turns green 26.5 seconds after the proceeding light turns green) but over time, the internal clock could drift by a second or two or three… and this could completely throw off your coordination. I've seen a major corridor have vehicle delays of 20 *minutes* simply due to a clock drift over a few months.
DNA from 9,000-year-old skeletons disproves migratory waves theory from South Africa - Earth.com
What is most interesting though, is that this will likely not achieve what you think it is going to achieve. When people ask for “Synchronization” what they really want is for Synchronization in *their* direction of travel. What I have listed above is for Optimization for all directions of travel.
First, we should be clear what you mean by synchronized.
Better “synchronization” traditionally needs communication. A cheap solution in modern day traffic signal controllers would just be a GPS receiver, to receive GPS time, which is extremely accurate, and would keep all of these traffic signal controllers perfectly in sync. Say $1000 for each traffic signal, so $1M for that upgrade.
Let's add $10,000/signal for upgrades to communication (maybe some cheap cell modems instead of fiber or copper, though that adds operating costs, which we will ignore). $10M.
Well… despite many people's beliefs, haven't coordination between traffic signals is not always the best thing to do. Many signals run in uncoordinated mode on purpose, because in some cases, allowing that traffic signal to work independently from all other traffic signals allows for it to be more flexible, and adapt to local traffic more easily. If a traffic signal is isolated, with no other traffic signals in sight, then what is the point in coordinating it? That just restricts it to a coordination schedule when it could adapt to traffic better if it was operating in a free uncoordinated mode (you asked me for the cost of synchronizing all the traffic signals, you didn't ask if it was a good idea). Coordination works best along a corridor, of multiple traffic signals, ideally spaced equally apart, with major traffic flows in one or two directions along the path of that corridor. So buy synchronizing all the traffic signals of our fictitious city, in many areas traffic will actually be worse.
But at least spending $30+ million to “synchronize” all the signals has improved life for vehicles in general, right?…
Two way streets presents him even more interesting problem, because even in the most ideal circumstances, we can only get 50% of the green band in each direction. So even though we could get both directions to appear “synchronized”, only half the number of vehicles will get through as on the one-way Street.
But in real life, traffic is typically going in *twelve* directions at each 4–way intersection. This is the four directions of travel for straight through movements, but also 4 left turns and 4 right turns. Let's ignore the four right turn movements, since many jurisdictions allow right turn on red. That still leaves vehicles going in eight directions of travel. So when you say “synchronized” you should be clear as to which of the eight directions you'd like me to synchronize for, leaving all the other seven directions of travel to frustratingly hit red lights, or have poor green bands (small numbers of vehicles making it through that green).
But just clock coordination isn't really optimal. Much better would be to have actual communication to the signals, this way a central operator could implement broad scale changes to the system, and even change timing plans to accomodate different times of day and different travel patterns.
Most traffic signals actually have the ability to be coordinated by offset. They may not have any sensors, or communication, but they can be synchronized simply by their internal clocks. This doesn't cost much (relatively), just the time of an engineer or technician to calculate a good offset, and program it. This level of synchronization for, say 1000 signals, would cost around $250,000 (a very rough ballpark) to do some modeling to find optimal offsets for synchronization. But for excellent modeing, we also need traffic data, say $10,000 per intersection. So, that's $1M for the data. $1.25M in total, for a very rough estimate.
To add it up (I'll skip the GPS clock updates, since I have communications, and can send out a citywide clock update every night): $31.25M.
Add another $5,000 but only to half the signals, for upgrades to Controllers (many cities have various controllers of different generations, as they grow, so only some of them need to be upgraded in my scenario). This is pretty optimistic, but we will say $5M.
When people visualize synchronization, they always remember a one-way Street where they were traveling and got a green, followed by another green, followed by another green, etc all the way down a corridor. This is extremely easy to achieve IF the street is one-way. So they think, why don't we do this for the rest of the entire city?
I'll also need a new traffic management center, servers, terminals, and a Centralized advanced traffic signal management software system. Say $10M, if I put this together with my own team on a shoestring budget with the cheapest ATSM software I can find and bargain basement used equipment.
So perhaps what you mean by synchronization is a city wide adaptive and responsive system with automation?
1st measles case of the year in South Dakota as CDC updates travel guidance - ABC News
In this scenario, we are also going to ignore all the pedestrians and bicyclists, and emergency vehicle pre-emptions, and train crossings, just to keep things simple. But in real life, traffic engineers absolutely do need to accommodate all of those in the desired quest for synchronization. If I don't allow the correct clearance time for pedestrians, I would be violating national standards, and probably lose my license or get sued.