I saw a good article on c/upliftingnews about AI improving traffic signal controllers. It’s good and all, I just can’t help but think of the “look at what they need to have a fraction of our power” meme while reading it
If everyone was biking, people would deliberately bump and swerve over each other and they’d still block the streets. Cars aren’t the problem, people are.
Improving traffic lights doesn’t require AI, you just need sensors and some basic code to respond accordingly.
Most lights in the us run on a cycle without accounting for traffic at all. Most don’t even take into account the time of day.
Car dependent design is bad. But the us can’t even do car dependency well. You have to constantly wait at a light to leave the intersection clear for no one.
The solution is not AI the solution is having people responsible who care at least somewhat.
So I do reinforcement learning research at my university, and the coworker I sit next to everyday does traffic signal optimization using multi agent reinforcement learning and simulation. (E.g. his reseach is on stuff like this paper)
And we literally agree with you; sensors are THE problem for 90% of the inefficiency. Its rare to even know how many cars pass through in a day, or whether its 1 or 500 cars waiting at a light. However, Google knows (or can approximate), which is partially why they and they alone can get something like 30% improvement.
The other 10% inefficiemcy is coordination stuff though, which can be more difficult than you might think to fix.
in the rush hours it gets tricky because of effects like a light turning green, but traffic being jammed from a red light before. For these you need a network model and it is crazy complicated to adequately model and optimise even just a small street network.
So yeah, best solution is to reduce car traffic as a whole.
what is ahead? for that you need to find out which are the main routes people take. But you also cant just give the dominant route alle the passage, because the other routes are important too. With that you get a complex network you need to optimise, where a central control uses the sensor input from the individual lights, but local contral is not sufficent.
And this is what the original comment stated, with his colleagues using reinforcement learning as one possible approach.
For a big road/street, whatever the main flow of traffic is following. So for a north-south street that’s busier than the east-west street intersecting with it, optimise the flow for traffic going north-south, including the intersections ahead. A “green wave” or “groene golf” in Dutch would work wonders. Stick to the advised speed on the digital signs and you get a wave of green lights for x amount of upcoming intersections. I’ve had them for up to 9 in a row. For the streets crossing the main road, you get some sensors, probably inductive loops to check if there are cars waiting. If there are, periodically give them green and turn the main road to red. If there are no cars on the main road (e.g., at night), you could have an extra induction loop ahead of the crossing so that the light turns green for the crossing road whenever someone approaches, before even having to stop at the light.
Sure, you could use reinforcement learning there. But you really don’t have to. Analyse the traffic for a while, and it’ll stay pretty much the same for a long, long time. Just optimise the cycles according to the time of day and day of the week and you should be good.
Roundabouts are great but it’s expensive to convert a crossing to a roundabout, not to mention difficult since the traffic needs to be rerouted during construction. In the most congested areas it’s almost impossible, which is ironically where it is needed the most
Additionally (I still love roundabouts) there can be a max-wait-time problem when there is heavy traffic in one direction.
If a basketball game ends there can be 20,000 cars bumper-to-bumper trying to leave. Let’s say (looking at a map) they’re going left-to-right through an intersection.
If there’s 1 car trying to go top-to-bottom…
If the intersection is a stoplight it doesn’t matter. Even if there were 20 million left-to-right cars; it’s still a 5 or 10min wait for the top-to-bottom car.
If the intersection was a stop sign it also doesn’t matter; it’ll be the left-to-right cars turn then the top-to-bottom cars turn
At a roundabout though (at least in the US), vehicles entering on the left always take priority over vehicles entering from the bottom. So the top-to-bottom guy could be there all night
Game days on my campus can cause a 2 hour wait on a 1 mile road. My campus is unusual, but just FYI absolutely insane wait times do happen regularly in some cities.
Yeah I think scribbling out the 30% with a 100% and saying “roundabouts” would make for a pretty good punchline. I figured I’d get complaints about AI being quick and low cost compared to road construction, which is why I ended up going with the “bikes” punchline instead.
I think this is a US problem. In europe roundabouts seem much more popular - drivers are aware of the rules and bikes seem to go along fine.
However, roundabouts are only efficient if all connected roads are about equally frequently taken. If one main road and three small connectors are on the same roundabout, the small ones may end up being softlocked.
Yes, but you don’t need lights if there are only bikes. Lights are there to prevent heavy vehicles from colliding. If there are no heavy vehicles, then the lights aren’t needed.
So you’re ok with getting hit by another bike (or several) when you go through an intersection.
Unless you live in a small town, if everyone used bikes, city centre intersections would be mostly mountains of crashed bikes and people trying to get out of that mess while more bikes continue to pile on.
Just reload and watch the video. Humans evolved for that speed, which is the reason we don’t need traffic lights to avoid people bouncing into each other in busy pedestrian areas. The same translates well for bicycles.
Only once we start putting tons of metal around us, allow for super fast acceleration, and isolate ourselves from our surroundings by blocking view and and sounds, this becomes an issue.
You’re right. I was just imagining crossing an intersection at the same speed I normally cycle at, but of course I’d slow down at the intersection, the same way I stop at the red light.
Edit: Why the downvotes? I’m conceding I was wrong.
Edit 2:
Alright, this is actually funny.
I made a few off the cuff comments without thinking it through, while I was doing something else, and was shown that I was wrong, so I said as much, because why wouldn’t I? I don’t mind learning, right? Who cares?
Well, it seems you care. You care very much. To the point of toxicity. Confirming your reputation as a community.
So from today, I’ll make it my mission to troll you. I’ve never done this, and I’m even a cyclist for fuck’s sake, but you’re so easily triggered and so toxic… It’s going to be fun! :D
Yes, after you repeatedly spewed bullshit about things you have never experienced in your life, and when people helpfully handed you videos that explain things you insisted it wouldn’t show what it shows (and yes, that is an intersection in the second video).
Just stop trolling and talking out of your ass and you won’t get downvoted. Easy, huh?
Dude, seriously: are you ok? Is something going on in your life that makes you so upset at such a trivial thing? I’m here if you need someone to talk to.
Honestly. I’ve been through some shit, I know how it is. Forget this discussion, and talk to me.
The average travelling speed in Copenhagen is 15.5 km/h for cyclists and 27 km/hour for cars. In places with green wave for cyclist the average speed is 20.72 km/h.
I live in the Netherlands. There are so many bikes there is traffic jams out of bikes, there are piles of bikes everywhere.
No, you don’t need traffic lights for bikes, only if there are high speed heavy vehicles. I wouldn’t even say it’s just the heaviness, it’s the heaviness coupled with speed that makes them necessary.
Ok, I need to experience this myself, but I’ll take your word for it.
I was just thinking of the normal speed I cycle at when going to work, which is 25 - 30 km/h, and can’t imagine that not causing issues on intersections if there were no red lights.
Of course the answer is to slow down at intersections :)
The trivial point was; car wait times are reduced when there are less cars.
The main point is; even from a bike perspective its not about stopping/not-stopping, it’s about wait time. I have NEVER had so many bikes in front of me that I missed the cross-walk signal and had to wait a whole other red-light cycle. Comparatively I regularly have that happen to me in a car. Idk if its a 30% improvement but its less time waiting at red lights.
Finally, technically no, bikes don’t always have to (legally) wait at red lights. This is only a technicallity but some crosswalks, like several in my town (or the iconic one in Japan), we get the walk signal on red. My town is also unusual by officially allowing bikes on pedestrian paths. So bikes can legally cross on red.
No one mention on how this AI would treat peatons or if they even know they exists at all. Stop measuring traffic by speed and throughput and start measuring safety.
Absolutely. After moving from a roundabout-less country to Ireland, I wish every red light was replaced with one. The only drawback is that they’re more complicated than red lights and many people don’t know how to use them properly (or don’t care).
Sao Paulo have a lot of them on the middle of some high density neighborhoods, they work great as traffic calm measures because you can’t blast though a street because they are a roundabout on each intersection.
I don’t know where you live, but training and the test in Ireland are quite rigorous already. My son failed his first test because of a roundabout issue where anyone in their right mind would have done exactly the same as he did. But the rules are what they are and the test is strict.
Add comment