A New AI Traffic Light Could Help Shorten Your Commute Times (jalopnik.com) 82
A new study out of Germany says having traffic lights use AI technology may keep traffic flowing faster and smoother. Jalopnik reports: One of the partners in the study with an aggressively German name -- the Fraunhofer Institute for Optronics, System Technologies and Image Exploitation -- recently installed high-resolution cameras and radar sensors at a busy intersection with a traffic light in the city of Lemgo, according to New Atlas. The setup recorded the number of vehicles waiting for the light to change, the amount of time each of them had to wait and the average speed a vehicle drove through the intersection. Science wizardry was then used to train a machine-learning based computer algorithm. It experimented with different light-changing patterns. They would continuously adapt to real time traffic conditions and see which ones worked best to keep wait times down.
According to the simulations, the best artificial intelligence patterns could improve traffic flow by 10 to 15 percent. That may not sound like a ton, but add up all the time you spend white-knuckled at a long traffic light, and chop 15 percent off. Not too bad. The algorithm will be used to run the traffic lights at actual intersections in Germany for the next few months, and can only get better. The study is also looking to find ways to reduce waiting times at crosswalks for pedestrians. They're using LiDAR sensors among other things to assess the walking speed of pedestrians to make sure they have enough time to cross before the light turns on them.
According to the simulations, the best artificial intelligence patterns could improve traffic flow by 10 to 15 percent. That may not sound like a ton, but add up all the time you spend white-knuckled at a long traffic light, and chop 15 percent off. Not too bad. The algorithm will be used to run the traffic lights at actual intersections in Germany for the next few months, and can only get better. The study is also looking to find ways to reduce waiting times at crosswalks for pedestrians. They're using LiDAR sensors among other things to assess the walking speed of pedestrians to make sure they have enough time to cross before the light turns on them.
mp3 (Score:2)
Fraunhofer, the main developer of the MP3 compression?
Re:mp3 (Score:5, Informative)
The main developer of MP3 was a guy named Karlheinz Brandenburg. He worked at another Fraunhofer Institute. The Fraunhofer institutes are independent research facilities in Germany operated by the Fraunhofer Gesellschaft (Fraunhofer Society), which is a community of about 29,000 researchers in Germany.
Re: (Score:3)
Fraunhofer, the main developer of the MP3 compression?
Yep. The exact same Fraunhofer.
For those who don't know, Fraunhofer is a big research institute in Germany who research just about everything. Audio compression, traffic lights... you name it, they're researching it.
Or if the AI take a disliking to you (Score:2)
Re: (Score:3)
Aggressively German name??? (Score:5, Insightful)
...an aggressively German name -- the Fraunhofer Institute for Optronics, System Technologies and Image Exploitation
WTF? Why is this name "aggressively German"? Note that almost the entire name is English. FWIW, this weird tidbit comes from other articles about the study [msn.com], which TFS didn't bother to link, so why quote from it?
Journalists apparently don't bother to think about what they write, editors no longer exist, and Slashdot editors...well...
Re: (Score:1)
Or "aggressively black". You'd be lynched in an instant if you said that.
(Oh, wait, lynchings aren't for white people...)
Re: Aggressively German name??? (Score:3)
Or "aggressively black".
Like Goldberg.
Re: (Score:2)
Re: Aggressively German name??? (Score:2, Interesting)
Jalopnik has been showing a lot of casual racism lately. My pet peeve is they continue to refer to Volvo Cars as a Chinese company, despite being located in Sweden, run by Swedes, employing Swedes, building flagship cars in Sweden, paying the employees in Swedish money, and being traded on the Stockholm NASDAQ. But heaven forbid you call Jeep a Dutch brand despite its parent Stellantis being based in Amsterdam.
Re: (Score:2)
Well, I think there is enough precedence to keep in mind where the Chinese have their puppets.
That being said, Volve is indeed an example that seems pretty hands off.
Zeitgeist, I guess. You can't trust anybody anymore. Not that you ever could but it seems more obvious these days.
Re: (Score:1)
Only assholes self censor.
Re: (Score:2)
It's alright. Some editors are simply ignorant and prefer to make fun of themselves instead of validating basic facts.
Working from home (Score:3)
That really cuts the commute time down in my experience. If you have a proper home office settup rather than working from the kitchen table and a good fast internet connection then it's as good if not better than the office.
Re: (Score:2)
It does not take much to improve upon the sometimes lapidated office furniture at my (multinational luxury brand, ironically a German car manufacturer) offices. Also, I have healthy and tasty food, coffee and water at home. And quiet. And sufficient parking for my Toyota. Temperature to my liking, and a wardrobe to quickly change if it isn't. No mask wearing, foggy glasses, and alcohol hand sanitizer (OR greasy smudges left by colleagues). I can go on and on, but I'd argue this is WAY better than any office
Have to use cameras (Score:3)
With all the vehicles stopping a car length or more behind the white line, the magnetic coils aren't being tripped.
Of course a better option might be to use a directional antenna to see how many people are on their cell phones at a light. If > 1, turn light green.
Re: (Score:2)
Do you think that directional antenna would distinguish between active use and an idle phone receiving a texted photo? Someone reading a story from someone using voice controls? Drivers from passengers?
Re: (Score:2)
I was being somewhat facetious. Considering the number of people who have their faces buried in their phones at red lights who then don't drive when the light turns green, having some form of measurement to determine who's looking at their phone would keep the lights green longer and keep them off their phones and paying attention to their driving.
Obviously you are correct. There would be no way to determine one use from another.
Re: (Score:3)
In which polite city do you live? In LA most people stop with two wheels in the pedestrian crosswalk - and sometimes that'll be the rear wheels!
Re: (Score:2)
Re: (Score:2)
With all the vehicles stopping a car length or more behind the white line, the magnetic coils aren't being tripped.
Wait what? Your magnetic coil is only one car length long? Ours are several car lengths long and on any partially busy intersection there are additional coils some 25m back so the traffic system can count the cars queuing at the light to prioritise how to move the most amount of people. This goes for cyclists too, every bike lane has coils in it, many multiple coils.
No news here - Move along (Score:3)
Re: (Score:2)
In San Antonio, we have gotten rid of several turnarounds on the access (frontage) road. They never remember to update the light logic so once you make your left turn, your immediately hit a red light which quickly blocks all the traffic behind you. The red light made sense before removing the turnaround as there was no real reason you would skip the turnaround (which skipped the lights).
Taking away the turnaround is bad enough but the light situation makes it a real pain.
"Science wizardry was then used" (Score:5, Insightful)
Wow, the details of this summary are breathtaking.
Re: (Score:3)
We used to have a non-AI dumb version of this. Traffic lights had a sensor that could detect when a car was waiting and shorten the delay time, or keep one direction open until traffic wanted to use the other one.
I don't know what we stopped doing it. Then again the people in charge of traffic lights around here seem to be idiots. Often you hear that they accidentally left the lights on the default settings for a decade and that's why traffic was so bad. Maybe they only employ midnight flashers.
Re: (Score:2)
The problem I have with that is, that even in light traffic, I have to stop at every stoplight and wait for the sensor to kick in and turn the light green. If it could see the cars coming and turn the light green in time for them to keep cruising, it would work much better.
Real life getting TMPE? (Score:2)
I only need to look at Cities Skylines's excellent Traffic Manager; President Edition to know what must be done.
Isn't it shocking that in 2022 we still have traffic lights that run by a set timer regardless or traffic? -You know that one light that needlessly takes ages while the other green lanes have no cars?! -Well hopefully smarter traffic systems can rid us of rotting away in rolling cages as our lives idle by.
Imagine, a traffic light system that knows which lights to prioritise and which lights to
Re: (Score:2)
Isn't it shocking that in 2022 we still have traffic lights that run by a set timer regardless or traffic? -You know that one light that needlessly takes ages while the other green lanes have no cars?!
Sometimes that's laziness, cheapness, or stupidity, but other times it makes sense. If you have a major intersection that works well on a simple timer, then you may need other intersections around it to be on the same timer (with whatever offset) so that the pulses or surges of traffic coming through them reach the major intersection at convenient times.
Faster commutes ... or save the planet (Score:1)
Re: (Score:2)
Do busses where you live run red lights?
Re: (Score:1)
Kettering, Ohio was WAY ahead of Lemgo ... (Score:3)
Back in 1976, the city of Kettering, Ohio (named after Charles Kettering, inventor of - among many other things - the automobile self-starter), got a Department of Transportation grant [bts.gov] (the article, written by Kettering's then-chief traffic engineer, starts on page 62 of the PDF in the link) to design and build a city-wide traffic light control system called KARTS (Kettering Area Responsive Traffic System). If I recall correctly, it ran on a PDP 6, and it included CRT graphical displays of the traffic light grid and sensor data. KARTS enabled Kettering's street grid to handle more than 120% of its theoretically maximum traffic load without any new construction (other than installing the twisted-pair grid that linked the traffic light and road sensor network to the minicomputer which controlled it).
It didn't employ an AI (pause for laughter), but it was completely operational, providing real-time control over traffic light timing, by late 1978, when I was given a tour of the place. And the entire system was designed and built for approximately $1M in constant-value dollars (not counting ongoing operational costs, of course) ...
Re: (Score:2)
Many people are way ahead of Lemgo and you don't need AI to do it. Our traffic lights aren't timed or sequenced, they are prioritised based on time of day, wait time, and in many cases the number of cars counted approaching the intersection. Our magnetic coils read an entire lane at once and can tell the difference between a car, and 5 cars, and we have additional coils up the road. Heck some intersections even adjust the priority based on weather so if its raining pedestrians and cyclists get priority. Als
Induced demand (Score:2)
Re: (Score:2)
Ah, finally someone mentioning it. Improving traffic means nothing because of induced demand. You'd have to wait for a year at least to confirm the gain.
That said, well-done AI can be quite interesting. For example, if AI could control lights of an entire city (or even multiple cities), it could decide to favor traffic going to non-contended areas and block traffic coming in, to keep traffic fluid in all conditions.
Even at the intersection level, there are cases where it could help, especially in certain
Re: (Score:2)
Potential to shave the *most obnoxious* 10-15 (Score:2)
Beyond 10-15, it's likely the worst parts.
-Someone not stopping over the inductive loop causing the light to not know to change
-Sitting at a cross-street waiting for a light despite no cross-traffic
-Starting a switchover sooner by detecting a car approaching a red at an idle intersection faster.
So much less time stopped for no good reason.
And look, cities love it because it's even more justified to have cameras that happen to be able to double-duty as red light cameras!
If only (Score:2)
If only everywhere was as dedicated to providing a good driving experience as Germany. In Massachusetts I've seen signs that say "lights timed to require frequent stops". It blows my mind when I'm in another city coasting through 5 or 6 lights that are properly timed.
I wonder if my city participate in the study (Score:2)
Re: (Score:2)
That's due to a change in the design and meaning of crosswalk signs. A 20-30 second green walk sign (actually white) would catch a bunch of people in mid-crosswalk when the crossing traffic started moving. So they changed to short 'walk' signs and longer (usually blinking) 'dont walk'. Which means start crossing on the white, but you have sufficient time to get across on the blinking red. In fact, our city* has count-down timers on the crosswalks. You can judge for yourself whether you can still make it acr
Stereotypical Goldfish (Score:2)
There are absolutely zero ways to reduce traffic. Time and time again, it's proven over and over.
It doesn't matter whether we're talking about bandwidth of roads, of telephony, of data.
The reason that your commute takes as long as it takes is because if it were faster, more people would drive, and if it were slower, fewer people would drive.
Chopping 15% off of your commute time will simply result in more people choosing to drive alongside you, growing the very same commute time right back up again.
Welcome
Re: (Score:2)
So when traffic improved for a decade after they widened the freeways around here and traffic only got worse when the county grew significantly in population that was all in my head right?
Your entire post seems to be built on the idea that we somehow have an infinite number of potential drivers. I would say that given the shoddy state of US mass transit in most parts of the country everyone who can drive is already driving regardless of traffic. There just arent very good alternatives for the vast majority
Re: (Score:2)
Well, a) you proved my point. Roads got bigger, more people came. Of course there is a delay.
b) if your town was small, it didn't have infinite drivers. my megalopolis is huge. it does have infinite drivers. we choose when to drive, and how far to drive. less traffic makes us choose to drive more and farther.
c) your town had freeways that were too small. period. modern cities are designed to lock traffic within freeways. we can't grow the freeway because doing so would just bottleneck the traffic on
Re: (Score:2)
Well, a) you proved my point. Roads got bigger, more people came. Of course there is a delay.
No I didn't. Larger highways does not equal future population growth at all.
b) if your town was small, it didn't have infinite drivers. my megalopolis is huge. it does have infinite drivers. we choose when to drive, and how far to drive. less traffic makes us choose to drive more and farther.
The vast majority of Americans don't live in major cities https://www.nationalpopularvot... [nationalpopularvote.com] .
c) your town had freeways that were too small. period. modern cities are designed to lock traffic within freeways. we can't grow the freeway because doing so would just bottleneck the traffic onto the smaller streets. So we intentionally design our freeways to be too-narrow to carry enough traffic that would over-burden the main roads. if you were able to grow your freeways at all, then they were never properly sized in the first place. so that's not growing the freeways, that's fixing the original design flaw.
I don't think anything you said there is true at all. Site a source.
Re: (Score:2)
Oops, "cite". Old habit from childhood...
Re: (Score:2)
The vast majority of Americans don't live in major cities https://www.nationalpopularvot... [nationalpopularvote.com]
That seems to be a cherry-picked number, based upon 100 of the largest US cities, from a site that may have an agenda toward cherry-picked numbers.
From the Census Bureau website https://www.census.gov/program... [census.gov] (as of 2016, I'm sure there's more recent numbers):
Urban Areas ...
80.7%: Percent of the US population living within urban areas
12.1%: Urban population growth between 2000 and 2010.
9.7%: U.S. total population growth between 2000 and 2010
Urbanized Areas
486: Number of urbanized areas
16: Number of urban
Re: (Score:2)
"Urban area" is an incredibly broad term and does not apply to this conversation as the above poster is clearly referring to major urban areas, hence the validity of my own link.
I mean, if he's actually talking about areas smaller than 200,000 people (right around where my top 100 list finishes) then he's absolutely insane to claim that there's "infinite" potential drivers even if he's saying it figuratively. The largest city in my county is right around that size and I assure you that everyone who can affo
Re: (Score:2)
I live in a megalopolis that spans about fifteen cities. over 10 million people drive between those fifteen cities each and every day.
Sure, one city is only 150'000 "residents". Means nothing. 10 million people use that city, pass through that city, shop in that city, play in that city, work in that city.
Stop thinking about boundaries and numbers. Laboratory-level controls are meaningless. Start thinking cradle-to-grade.
So sure, we can exclude all areas where populations are isolated from neighbouring
Re: (Score:2)
Word of the day, pretty-much-every-american-that-can-afford-to-drive-already-drives. https://policyadvice.net/insur... [policyadvice.net].
Re: (Score:2)
umm, your new word of the day is: when.
Re: (Score:2)
How about, most Americans don't live in your hyper dense region of the country so why are you using it as a model to claim norms for the country.
Re: (Score:2)
I don't think I ever mentioned your country. But thanks.
And once again, we've already discussed your definition of "most" is arbitrarily chosen to make your point. In reality, "most" residents of your country travel through one or more hyper dense regions on a daily basis -- that's what makes them hyper dense.
My region actually isn't hyper dense at all. We're ~12 million people sprawled out over ~100 miles by ~50 miles. Again, your word of the was "morphological" area. Yes, millions of those ~12 travel
Re: (Score:2)
I live in a megalopolis that spans about fifteen cities.
My region actually isn't hyper dense at all. We're ~12 million people sprawled out over ~100 miles by ~50 miles. Again, your word of the was "morphological" area. Yes, millions of those ~12 travel more than 100 miles every day. Arguably, the majority of those ~12 travel more than 100 miles in at least two routine trips every week.
Man, you'll say anything just so you arent wrong. Either you live in a megalopolis https://www.merriam-webster.co... [merriam-webster.com] and therefore be definition do live in a very densely packed part of the country or you don't live in a megalopolis. Which is it?
Re: (Score:2)
It's nice that you can cherry-pick your definitions from an internet full of common usage.
The definition of a megalopolis has absolutely nothing to do with density. It is purely: two or more cities (of any shape, size, or density), which have no meaningful space between them.
The definition of an arbitrary word matters not to anyone's argument. My thesis (which exists independent of word choice, you may help me to find the correct words, instead of arguing my word choice, which may be influenced by cultura
Re: (Score:2)
It's nice that you can cherry-pick your definitions from an internet full of common usage.
Wtf? Cherry pick? I used the Merriam Webster definition, not some fringe site. They're probably the largest American dictionary maker still in existence. I literally don't know a better place to look up deffinitions of
I think I'll go with their definition rather than your home brew one.
The definition of an arbitrary word matters not to anyone's argument.
So now words dont have any meanings. You're a classic!
Everything you've said up until a couple posts ago is that you live in a high density area. Now all of a sudden you don't. Why don't we settle this since all we're doing i
Re: (Score:2)
First off, words don't have meaning here, thoughts and theses do. We use words to express thoughts -- which requires both the writer and the reader to share on a wealth of knowledge. So if you want to argue my words, we can convert this conversation. On the other hand, if you want to argue my point, then we can work on understanding my meaning, and I can apologize for using words differently than you do.
Second, Merriam Webster is a great source. And it's only one. There are countless others. So that's
Re: (Score:2)
Oh, Fourth: "The definition of a megalopolis has absolutely nothing to do with density. It is purely: two or more cities (of any shape, size, or density), which have no meaningful space between them."
That's not my "home brew one" as you suggested. See, I can cherry-pick my sources too. Arguably, my source is better than Merriam Webster.
Re: (Score:2)
First off, words don't have meaning here, thoughts and theses do. We use words to express thoughts -- which requires both the writer and the reader to share on a wealth of knowledge. So if you want to argue my words, we can convert this conversation. On the other hand, if you want to argue my point, then we can work on understanding my meaning, and I can apologize for using words differently than you do.
Yeah, you can take that to your local community college's philosophy department and have fun with that there. Words have meanings and I have no interest in talking to some one who is content to just arbitrarily define words as they see fit. I mean you've already been doing that with your own arguments. First you live in a high density area with "infinite potential drivers" and then all of a sudden you live in the most normal place in America.
Second, Merriam Webster is a great source. And it's only one.There are countless others. So that's what we call cherry-picking.
No it's not. it's called going to the most trusted source first. C
Re: (Score:2)
I like that you presumed I lived in your country. Typical US resident. Have you ever traveled outside of it?
And yes I've been to 4 continents thank you very much.
Your conceit here is absolutely amazing.
Re: (Score:2)
You dont even provide a link.... nice.
Probably because any link you provided would at least have a secondary definition that matched Webster's.
Re: (Score:2)
That is the very definition of cherry-picking. You chose a source, instead of listing many sources. You had no reason to believe that Merriam Webster's definition was at all typical, let alone not simply an error. And again, you chose a common-usage source, not an etymological source.
Infinite drivers doesn't say anything about high density. Not once did I say I lived in high density anything. You'll need to learn that density requires two metrics, not just one.
You clearly haven't left your country. You'
Re: (Score:2)
I don't need a link. Had you actually searched for multiple sources, you'd have found this one, I promise. You don't care about any of it. You aren't actually considering the subject being discussed. You care only about syntax. I'm not here for that. I'm here only to share my life-lived experience, with those who may have none or different experience.
Re: (Score:2)
Oooh, that's more than I've seen. I've only been to 2, 2.5 if you count the way some do.
Alas, I think I'll take a moment and pick at your words.
"been to" has zero meaning. Have you lived in any other cultures? No one cares that you traveled to a business convention in an airport somewhere. All international airports are the same, as are international city cores -- which was our original discussion, but I digress. I don't know about war zones, but I'd suspect that being a soldier deployed to a war zone,
Re: (Score:2)
You don't care about any of it. You aren't actually considering the subject being discussed. You care only about syntax. I'm not here for that. I'm here only to share my life-lived experience, with those who may have none or different experience.
What a novel way to characterize our conversation. I would say that I've been talking about your initial claim of "There are absolutely zero ways to reduce traffic." the entire time where as you've fallen back onto word games, complete revisions of past statements, and the changing of topics.
Re: (Score:2)
Exactly. You've been talking about MY initial claim. You've brought zero experiences of your own. You have, however, kept repeating your incorrect recollection of my words.
And you've just done it all again. I asked for your opinions, I gave you mine. You contributed nothing.
You're just picking a paragraph of mine, quoting it -- as though it isn't available otherwise -- and addressing individual elements from within my larger discourse.
You offer nothing. Not to me, and not to any other readers of this
Re: (Score:2)
All you've done is try to steer the conversation towards a targets you feel you can actually support rather than support your initial claim. We've fallen all the way down this hole we're in due to your side tracking and revisionism. I'm done participating in that as it's become tiresome and repetitive. Goodbye.
Re: (Score:2)
Still offered zero of your own experience. Way to contribute. You didn't talk about your travels. You didn't talk about your own personal experience. You didn't talk about your own community/city. All you did was talk about my opinion. And now you're leaving, still having added nothing.
Re: (Score:2)
why use AI when you can solve it with algebra? (Score:1)
seems like a pretty simple problem with very few variables.
just build a model and fit it empirically to the data. Then run a simple optimization algorithm. If you can't solve it by taking the second derivative, just use newton's method or something.
so much less computational power. so much less computational complexity. and you get a more accurate answer.
We've had similar in US (Score:2)
Many don't seem to realize we've had responsive traffic lights in the US for years now, and they are getting so widespread they are the majority around here now, no longer using the old electromagnetic loops or timing. They also have optimized to need fewer cameras, many served by just one overhead fisheye.
They work on a greedy system, anticipate approaching traffic, and adjust green/yellow timing based on approaches. (And respond to horses, bicyclists, non-standard vehicles that don't trip old electromag
Squirrel (Score:2)