Yes. Trains. Metro. Trans. Regular and affordable. Plus cycling infrastructure to get even more people out of their cars by making cycling safer. Making driving easier and more pleasant will just create more drivers.
I'd love to see this idea expanded upon by applying it to public transit and biking as well: - Analyze specific intersections and streets along bus routes to identify good candidates for new bus lanes and signal priority. - Analyze specific routes that are underserved that could significantly reduce traffic if a bus were added. - Analyze specific streets where adding a bike lane would improve traffic flow. - Analyze areas where drivers consistently speed to identify spots where traffic calming measures would be useful. Cities are always struggling when it comes to funding public and active transit, so being armed with that kind of data would give them the insight they need to prioritize the right projects.
5:46 - a lot of the US was already designed around pedestrians and streetcars in the 1920s and 1930s and yet we found the willpower to destroy them and rebuild them to cater exclusively to the automobile. "It's already built for cars" is poor rationale for why we can't make major changes for sustainable cities.
Us urbanists in seattle are very wary of this solution. They are implementing it on 15th avenue, which is a dangerous stroad, but it is congested a lot so it brings down the speed of cars. I worry that this plan will therefore increase average speeds and make it more dangerous for cyclists and pedestrians. And what happens to the beg buttons for crosswalks when a pedestrian wants to cross? Well, that would create congestion so we shouldn’t let them cross… or what about the cutting of the traffic lights- how would that affect the timing of bike lights? This could really be bad for safe streets for people outside of cars.
I think, this becomes really interesting, if it is used in combination with a road diet. Many US cities tend to widen roads or just overbuild them in the beginning without even having traffic issues. Imagine the AI saying "this 6 lane stroad would be more efficient with 2 drive lanes, 1 center turning lane, bus lanes and bike lanes on each side" :)
Another potential negative: as cars and especially pedestrians/cyclists become familiar with the specific timings of the lights they pass through frequently, they sometimes kind of… cheat the system by starting to move before they have a green. So widescale light retimings have some potential safety implications as well. Obviously that would go away over time and can probably be somewhat mitigated with proper communication about the retimings.
I’m glad you acknowledged that your title is basically clickbait by admitting that induced demand would probably ensure traffic to some degree at peak times at least.
The only critique that I'd give on your discussion of induced demand, it is a bit more nuanced than "oh, I might as well start driving now that there is less traffic." It is "Oh, I might as well start driving there now that there is less traffic." The first sentence almost makes it sound like a shift in mode share, people who previously walked, biked, or took transit deciding to drive for a trip instead of what they were doing before, when in reality these are trips that previously never would have happened that are now happening and it is specifically trips by people choosing to drive. A good example would be someone who currently drives 5 miles for their commute, there is a job that pays slightly better 15 miles away, but it just isn't worth the inconvenience of driving that far... now that there has been that highway expansion though, those extra 10 miles don't seem like such a big deal, so they go ahead and apply for that job. There could be some amount of mode share shift, but it will almost always be part of moving to a longer trip. I used to walk to the grocery store a few blocks away, but now that the highway has been expanded, I can drive to the one across town that I prefer. It's worth also adding, you can get the same induced demand can work on high quality transit as well. My in laws live in Ogden (about 40 miles north of Salt Lake City), they previously would never consider going to Temple Square to see the Christmas lights, but now that Frontrunner has service between Ogden and Salt Lake, with a station convenient to them in Ogden and timed connections to the Trax light rail in downtown to get them literally to the front entrance of Temple Square, it is now an annual tradition. Why pay to go to a Utah Jazz game and deal with the hassle of getting to and from whatever the hell that arena is called this year and deal with parking while there, when we can watch it on ESPN for free (well, included in what they already pay for ESPN), but now that they can take Frontrunner and there is a Trax station literally at the front door of the arena (that they have wisely just named Arena to avoid having to change maps every time the naming rights change), you know what, maybe it is worth occasionally splurging and buying tickets to see them in person. It's like a lot of urbanists are now saying, we need to build infrastructure for the cities we want, not for the cities we have, because our built environment determines our behavior more than our behavior influences how an environment is built.
More straightforward examples: Build a bigger road, and you will soon see a shopping center next to it. So more people will drive there instead of the old supermarket which was closer to their home. This eliminates the newly created excess capacity, and make all other roads (which were not extended) even more congested than before. This is not even remotely a hypothetical scenario. People may or may not choose jobs further away, but the above is definitely happening right away.
@@juzoli The examples I gave aren't exactly hypothetical either. The job I have now is because a "new" (it opened a long time ago, but this story also happened a long time ago) freeway opened and suddenly it felt worth it to drive further for the better paying job. The existence of a convenient freeway connection is why I do my grocery shopping where I do instead of a closer store. I had a coworker who accepted the job because she knew that there was going to be a new freeway built in the next couple of years and she was willing to deal with the extra hassle now to get her foot in the door before there was even more competition for jobs from all the other people who would now be commuting from her neighborhood (she didn't last long, turns out the promise of future infrastructure doesn't induce as much demand as actual in place infrastructure does, but it does still induce some demand). My point being, I chose examples of things that had directly impacted my own decisions, so I'm not having to use hypotheticals. And, I try to do better now, I take the bus to work instead of driving, I try to make more decisions on where to go around where I can get on bike and transit (I'm still going to the grocery store that I have to drive to though, I'm trying to improve, but I'm not perfect, especially when that store literally costs between 20-30% less than the store that I could walk to).
@@smileyeagle1021 By “hypothetical “ I didn’t mean imaginary. It is just harder to find explicit proof that it is happening, so one could easily argue against it. Big stores opening next to big roads are much more obvious, and you can use it as an example to persuade people who are more skeptical.
This is just an equivalent of higher capacity roads, which is already proven to be NOT a solution. See California highways… Sure it might make the infrastructure cheaper, but will not reduce traffic jams. If you want to actually improve traffic, then you need: - Better city planning, so most frequently visited places (school, workplace, supermarket) are easily accessible without long driving - Better public transportation and trains
6:00 When you say that efficient traffic lights could streamline driving, wouldn't that just add more to induced demand? 6:55 HA! nevermind :D I still think that the short-term benefits of AI will just make things worse long-term and/or it will still inevitably be an excuse not to have proper infrastructure...
To do that, cars shouldn’t be substantially faster than the greener alternatives. Either make the bikes, buses, and trains go faster, or slow the cars down. Most suburbs won’t be converted back into farmland until the houses are let to rot and abandoned, and they won’t densify writ large unless there is massive population boom. At best, sprawling low density cities will have pockets of high density which satisfy a token appeal to “sustainability” (a.k.a. greenwashing) while missing the big picture which is that the rest of the city area will remain low density.
False dilemma fallacy. The cars already exist, intersections already exist, this technology already exists and is nearly free to implement. Why disregard a low-cost optimizer because you want a culture shift and new infrastructure?
If the goal is to maximize utilization of the available road infrastructure by cars, I have confidence that this would be effective at that by packing in more cars with the same degree of congestion. But due to Jevon’s Paradox and induced demand, it would reduce congestion in the long term. The only effective way to do that will be to provide high quality alternatives to driving.
Congestions make streets safer by reducing car speeds. removing congestion as what happen with covid made US streets much more dangerous because car speeds increased
Speed as a component of safety is important in collisions, but slow congestion has its own issues. When traffic moves slowly, people will pull into the intersection to get through before the next, making the spaces unsafe outside a vehicle. Slow traffic moving at capacity where drivers are encouraged to stop before clogging intersection is best.
I see this as attractive to cities that don’t want the expense of a traffic engineer but have enough signals to warrant them. This is the replace-a-job arena people are afraid of. Big cities have no fear-maybe used as suggestions. Also limited to signals that can be changed remotely and/or dynamically. Most signals I encounter are set-timer-once at the box with some with variable timers throughout a day, but rarely “smart”. The hard questions are about what the intention of the use & design capacity should be. What kind of controls should be used & whether an LOS is an appropriate guide for a path too.
I am sceptical that this will actually reduce traffic, improving intersection efficiency is analogous to adding one more lane and we all know how well that works. This technology would be much better used as a way to identify high-demand corridors which are under served by transit.
This might improve traffic flow a bit and allow more cars to get around, but I can't see it being enough. The only long-term solution is more walkability, cycle lanes and public transport.
It bears a serious risk. If the system is hacked and shut down, there will be chaos. Do this at a few key intersections and you shut down an entire city. Also, the AI solution might induce demand and cause more traffic, just like adding more lanes. A dense mass transport system with trains and trams is still a good solution for the long run.
Hey, I love this type of video. I'd like to see more videos regarding researches especially in the technology and sustainability sector, just like this one! Subscribed :D
It may be new google technology but here in Europe we have selflearning and adapthing traffic lights all over for many years now. But it's only a small benefit.
Americans assume that pubic transit is supposed to be dirty, slow, and unreliable. Just like American's assume that healthcare is unaffordable and mostly unattainable, and that an MRI at the hospital really costs $10,000. We've never experienced anything else so we don't know the possibilities.
I'd choose traffic analysis cameras and sensors than this. why? because AI can only make assumptions based on given data and is not as dynamic as Traffic Analysis System ( plus traffic analysis systems already have AI now so why bother on this?)
So if everyone entered their destination into their phone linked to their car, local Ai would know where everyone else was travelling and could organise everyone’s route, minimising the disruptions. Self driving cars would be the way to go with this tech
Everyone in here are screaming for more public transportation, but ignore that fact it would cost billions to get better transit. The video says the a.i. is much cheaper. We dont have unlimited budgets.
Already exists. Agencies are too cheap to deploy the sensors (you don't need a lot but they're still to cheap) and integration services with signals. Tech has been around for 20 years.
How long will it take, until they realize the only way to solve traffic are trains.
hopefully someday... trains really are the best
Yes. Trains. Metro. Trans. Regular and affordable. Plus cycling infrastructure to get even more people out of their cars by making cycling safer. Making driving easier and more pleasant will just create more drivers.
Never
@@cosmicgreenThat's what people in the Netherlands thought back in the '70s. Now it's an urbanist's dream come true.
Or have housing closer to where people work and shop.
I'd love to see this idea expanded upon by applying it to public transit and biking as well:
- Analyze specific intersections and streets along bus routes to identify good candidates for new bus lanes and signal priority.
- Analyze specific routes that are underserved that could significantly reduce traffic if a bus were added.
- Analyze specific streets where adding a bike lane would improve traffic flow.
- Analyze areas where drivers consistently speed to identify spots where traffic calming measures would be useful.
Cities are always struggling when it comes to funding public and active transit, so being armed with that kind of data would give them the insight they need to prioritize the right projects.
5:46 - a lot of the US was already designed around pedestrians and streetcars in the 1920s and 1930s and yet we found the willpower to destroy them and rebuild them to cater exclusively to the automobile. "It's already built for cars" is poor rationale for why we can't make major changes for sustainable cities.
That Google AI should just have one answer:
"build denser neighbourhoods with high frequency transit"
Us urbanists in seattle are very wary of this solution. They are implementing it on 15th avenue, which is a dangerous stroad, but it is congested a lot so it brings down the speed of cars. I worry that this plan will therefore increase average speeds and make it more dangerous for cyclists and pedestrians. And what happens to the beg buttons for crosswalks when a pedestrian wants to cross? Well, that would create congestion so we shouldn’t let them cross… or what about the cutting of the traffic lights- how would that affect the timing of bike lights? This could really be bad for safe streets for people outside of cars.
Nah. Not dangerous stroad. Not “dangerous stroad.” /
I think, this becomes really interesting, if it is used in combination with a road diet. Many US cities tend to widen roads or just overbuild them in the beginning without even having traffic issues. Imagine the AI saying "this 6 lane stroad would be more efficient with 2 drive lanes, 1 center turning lane, bus lanes and bike lanes on each side" :)
Yeah, I think there's some cool potential uses for this
Another potential negative: as cars and especially pedestrians/cyclists become familiar with the specific timings of the lights they pass through frequently, they sometimes kind of… cheat the system by starting to move before they have a green. So widescale light retimings have some potential safety implications as well. Obviously that would go away over time and can probably be somewhat mitigated with proper communication about the retimings.
I’m glad you acknowledged that your title is basically clickbait by admitting that induced demand would probably ensure traffic to some degree at peak times at least.
the AI: "use trains". U.S. city planners: "AUGH NAURR NOT MY CARRRS"
It'd be fun if the AI would answer.... "More Public Transit mate"
Yeah it would!
The only critique that I'd give on your discussion of induced demand, it is a bit more nuanced than "oh, I might as well start driving now that there is less traffic." It is "Oh, I might as well start driving there now that there is less traffic." The first sentence almost makes it sound like a shift in mode share, people who previously walked, biked, or took transit deciding to drive for a trip instead of what they were doing before, when in reality these are trips that previously never would have happened that are now happening and it is specifically trips by people choosing to drive.
A good example would be someone who currently drives 5 miles for their commute, there is a job that pays slightly better 15 miles away, but it just isn't worth the inconvenience of driving that far... now that there has been that highway expansion though, those extra 10 miles don't seem like such a big deal, so they go ahead and apply for that job. There could be some amount of mode share shift, but it will almost always be part of moving to a longer trip. I used to walk to the grocery store a few blocks away, but now that the highway has been expanded, I can drive to the one across town that I prefer.
It's worth also adding, you can get the same induced demand can work on high quality transit as well. My in laws live in Ogden (about 40 miles north of Salt Lake City), they previously would never consider going to Temple Square to see the Christmas lights, but now that Frontrunner has service between Ogden and Salt Lake, with a station convenient to them in Ogden and timed connections to the Trax light rail in downtown to get them literally to the front entrance of Temple Square, it is now an annual tradition. Why pay to go to a Utah Jazz game and deal with the hassle of getting to and from whatever the hell that arena is called this year and deal with parking while there, when we can watch it on ESPN for free (well, included in what they already pay for ESPN), but now that they can take Frontrunner and there is a Trax station literally at the front door of the arena (that they have wisely just named Arena to avoid having to change maps every time the naming rights change), you know what, maybe it is worth occasionally splurging and buying tickets to see them in person.
It's like a lot of urbanists are now saying, we need to build infrastructure for the cities we want, not for the cities we have, because our built environment determines our behavior more than our behavior influences how an environment is built.
More straightforward examples:
Build a bigger road, and you will soon see a shopping center next to it. So more people will drive there instead of the old supermarket which was closer to their home. This eliminates the newly created excess capacity, and make all other roads (which were not extended) even more congested than before.
This is not even remotely a hypothetical scenario. People may or may not choose jobs further away, but the above is definitely happening right away.
@@juzoli The examples I gave aren't exactly hypothetical either. The job I have now is because a "new" (it opened a long time ago, but this story also happened a long time ago) freeway opened and suddenly it felt worth it to drive further for the better paying job. The existence of a convenient freeway connection is why I do my grocery shopping where I do instead of a closer store. I had a coworker who accepted the job because she knew that there was going to be a new freeway built in the next couple of years and she was willing to deal with the extra hassle now to get her foot in the door before there was even more competition for jobs from all the other people who would now be commuting from her neighborhood (she didn't last long, turns out the promise of future infrastructure doesn't induce as much demand as actual in place infrastructure does, but it does still induce some demand).
My point being, I chose examples of things that had directly impacted my own decisions, so I'm not having to use hypotheticals.
And, I try to do better now, I take the bus to work instead of driving, I try to make more decisions on where to go around where I can get on bike and transit (I'm still going to the grocery store that I have to drive to though, I'm trying to improve, but I'm not perfect, especially when that store literally costs between 20-30% less than the store that I could walk to).
@@smileyeagle1021 By “hypothetical “ I didn’t mean imaginary. It is just harder to find explicit proof that it is happening, so one could easily argue against it.
Big stores opening next to big roads are much more obvious, and you can use it as an example to persuade people who are more skeptical.
People will rather summon a machine god to run your cities' traffic than build walkable neighborhoods.
This system doesn't control anything. It's a data analyzer that makes optimization recommendations to civil engineers. (for free)
Subways dont care about traffic. Build them and expand them as much as possible and you will be good
This is just an equivalent of higher capacity roads, which is already proven to be NOT a solution. See California highways…
Sure it might make the infrastructure cheaper, but will not reduce traffic jams.
If you want to actually improve traffic, then you need:
- Better city planning, so most frequently visited places (school, workplace, supermarket) are easily accessible without long driving
- Better public transportation and trains
6:00 When you say that efficient traffic lights could streamline driving, wouldn't that just add more to induced demand?
6:55 HA! nevermind :D
I still think that the short-term benefits of AI will just make things worse long-term and/or it will still inevitably be an excuse not to have proper infrastructure...
Maybe you could maintain the same intersection efficiency and reduce number of lines.
You don't need AI to solve traffic, you just need to get more people out of cars and onto bikes, buses and trains.
This, the U.S. needs a world class public rail transportation system.
That would cost more.
It will be so worth it though. @@MrPAULONEAL
To do that, cars shouldn’t be substantially faster than the greener alternatives. Either make the bikes, buses, and trains go faster, or slow the cars down. Most suburbs won’t be converted back into farmland until the houses are let to rot and abandoned, and they won’t densify writ large unless there is massive population boom. At best, sprawling low density cities will have pockets of high density which satisfy a token appeal to “sustainability” (a.k.a. greenwashing) while missing the big picture which is that the rest of the city area will remain low density.
False dilemma fallacy. The cars already exist, intersections already exist, this technology already exists and is nearly free to implement. Why disregard a low-cost optimizer because you want a culture shift and new infrastructure?
If the goal is to maximize utilization of the available road infrastructure by cars, I have confidence that this would be effective at that by packing in more cars with the same degree of congestion. But due to Jevon’s Paradox and induced demand, it would reduce congestion in the long term. The only effective way to do that will be to provide high quality alternatives to driving.
ah yes, lets add more capacity to our roadways, that always fixes traffic
BUILD TRANSIT AND BIKE INFRASTRUCTURE!!
When moving 1000 a hour is seen as a success you know it's not officiant
Congestions make streets safer by reducing car speeds. removing congestion as what happen with covid made US streets much more dangerous because car speeds increased
Speed as a component of safety is important in collisions, but slow congestion has its own issues. When traffic moves slowly, people will pull into the intersection to get through before the next, making the spaces unsafe outside a vehicle. Slow traffic moving at capacity where drivers are encouraged to stop before clogging intersection is best.
Wrong! Congestion will exponentially increase pollution and waste much more energy. The added pollution will lead to many more deaths.
I see this as attractive to cities that don’t want the expense of a traffic engineer but have enough signals to warrant them. This is the replace-a-job arena people are afraid of. Big cities have no fear-maybe used as suggestions. Also limited to signals that can be changed remotely and/or dynamically. Most signals I encounter are set-timer-once at the box with some with variable timers throughout a day, but rarely “smart”.
The hard questions are about what the intention of the use & design capacity should be. What kind of controls should be used & whether an LOS is an appropriate guide for a path too.
I am sceptical that this will actually reduce traffic, improving intersection efficiency is analogous to adding one more lane and we all know how well that works. This technology would be much better used as a way to identify high-demand corridors which are under served by transit.
This might improve traffic flow a bit and allow more cars to get around, but I can't see it being enough. The only long-term solution is more walkability, cycle lanes and public transport.
There is a thing called rails and something called bikes.
It bears a serious risk. If the system is hacked and shut down, there will be chaos. Do this at a few key intersections and you shut down an entire city. Also, the AI solution might induce demand and cause more traffic, just like adding more lanes.
A dense mass transport system with trains and trams is still a good solution for the long run.
Hey, I love this type of video. I'd like to see more videos regarding researches especially in the technology and sustainability sector, just like this one! Subscribed :D
Put money in public transit that has dedicated right of way, slow other traffic to 20mph, remove on street parking and traffic will solve itself.
It may be new google technology but here in Europe we have selflearning and adapthing traffic lights all over for many years now. But it's only a small benefit.
I am disappointed in Google. We know what solves traffic and this is not it. This just encourages more cars - which do not belong in our cities.
We should just take all this traffic, and push it somewhere else!
LMFAO
Americans assume that pubic transit is supposed to be dirty, slow, and unreliable. Just like American's assume that healthcare is unaffordable and mostly unattainable, and that an MRI at the hospital really costs $10,000. We've never experienced anything else so we don't know the possibilities.
You dont need AI. You need roundabouts
I'd choose traffic analysis cameras and sensors than this. why? because AI can only make assumptions based on given data and is not as dynamic as Traffic Analysis System ( plus traffic analysis systems already have AI now so why bother on this?)
no way there's a shot of my school in this video
So if everyone entered their destination into their phone linked to their car, local Ai would know where everyone else was travelling and could organise everyone’s route, minimising the disruptions. Self driving cars would be the way to go with this tech
Self-driving cars won't solve transport problems because pedestrians.
Everyone in here are screaming for more public transportation, but ignore that fact it would cost billions to get better transit. The video says the a.i. is much cheaper.
We dont have unlimited budgets.
Already exists. Agencies are too cheap to deploy the sensors (you don't need a lot but they're still to cheap) and integration services with signals. Tech has been around for 20 years.
This is the opposite of what it says! Don’t believe it