Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

200k Flemish drivers can turn traffic lights green

vrt.be

Post reply on HN
Print view View original

151–160 of 190 posts

Re: 200k Flemish drivers can turn traffic lights green

#151

Post by smeej »

With all the data companies certainly have about traffic flow, it boggles my mind that there isn't a company out there selling city-wide "traffic light optimization" where they come in, analyze the data, and set the traffic lights to optimize traffic flow. Heck, sell it as a subscription where the analysis continues running and factors in things like major events or road closures automatically.

Noise and air pollution would go down. Time waiting in traffic would go down. Crashes would go down.

Of all the things that have been optimized in the worod today by algorithmic analysis of data, it just surprises me that this hasn't been one of them.

Re: 200k Flemish drivers can turn traffic lights green

#152

Post by prmoustache »

gruez wrote: post #123
prmoustache wrote: post #122

Earlier quoted context omitted.

You are not held up by a cyclist riding at 10mph if your average speed is at 5mph because of the traffic lights. Besides there is a good chance you are the one holding up the cyclists creating a mini traffic jam with fellow drivers at every traffic lights.

>You are not held up by a cyclist riding at 10mph if your average speed is at 5mph because of the traffic lights. At a normal traffic light, maybe. But the whole premise of this thread is that the lights that only turn green when a car is approaching. >Besides there is a good chance you are the one holding up the cyclists creating a mini traffic jam with fellow drivers at every traffic lights. ???

the original statement was that drivers do it all the time, wether the traffic light might pass green or not.

Many cyclists slow down when they see a traffic light being red with a lane of stopped car in front. Because cyclists understand physics and drivers are generally bad at anticipating events.. Almost universally a driver try a quick pass to brake check the cyclists and stop in front of the cyclist while it would be more efficient for both of them to gently slow down instead of stopping so they still have speed when the light go green.

Re: 200k Flemish drivers can turn traffic lights green

#153

Post by Zanfa »

Ajedi32 wrote: post #97
iamtheworstdev wrote: post #94

Earlier quoted context omitted.

why would you need machine learning? just detect a minimum of one bright light source, assume it's a headlight, and switch the lights. What the worse a false positive could do? this doesn't need AI

Comparable systems based on inductive loops work during the day. A design based on detecting headlights wouldn't. Software/compute is cheap and as you said reliability is not a safety concern.

We’ve had camera-triggered intersections for years now in Estonia. They seem to work equally well in daylight as well as darkness. I don’t think you need any ML for it either, you basically only need to detect motion within a predefined region which can be done with traditional CV algos quite reliably.

Re: 200k Flemish drivers can turn traffic lights green

#154

Post by diggan »

whartung wrote: post #91

Late at night, at a friends house, and his car died, but I need a ride to my car. So, we took his motorcycle. We come up to a street light, and it's not sensing us. So, I hop off, run over, press the WALK button, and run back and hop back on. In moments, the light turns green, the WALK sign says WALK and we're on our way. Inevitably, a couple of blocks later, we're in the same boat. Being too clever by half, I rinse…

> So, I hop off, run over, press the WALK button [...] the light turns green, the WALK sign says WALK and we're on our way.

You mean the button for pedestrians to pass the crosswalk I guess? Why would it say "WALK" (be green for pedestrians?) and also be green for a vehicle to pass? The crosswalks around me are the reverse, if pedestrians can pass, no vehicles can pass, and vice-versa. Or maybe I misunderstood something?

Re: 200k Flemish drivers can turn traffic lights green

#155

Post by Aurornis »

rob74 wrote: post #69

> A simple example: at night, a main road mainly displays green, the side streets only get green for a very short time. If you arrive with the app from a side street, you can get green immediately, provided there is no traffic on the main road. In Germany, there's an established but much more low tech approach to this problem: simply turn off the traffic lights at night! All traffic lights already have yield/stop/pri…

> In Germany, there's an established but much more low tech approach to this problem: simply turn off the traffic lights at night!

Many American cities do this, too.

Well, they don't turn the lights off. They flash to distinguish this state from the power being off.

It's more common in smaller cities.

Re: 200k Flemish drivers can turn traffic lights green

#156

Post by Aurornis »

phkahler wrote: post #71
rob74 wrote: post #69

> A simple example: at night, a main road mainly displays green, the side streets only get green for a very short time. If you arrive with the app from a side street, you can get green immediately, provided there is no traffic on the main road. In Germany, there's an established but much more low tech approach to this problem: simply turn off the traffic lights at night! All traffic lights already have yield/stop/pri…

I don't understand why they need apps for this either. The infrastructure should be in place to detect cars approaching from the side street. Once they have the ability to detect cars, it's simple to assume everyone would be requesting a green light.

> The infrastructure should be in place to detect cars approaching from the side street. Once they have the ability to detect cars, it's simple to assume everyone would be requesting a green light.

Historically, traffic control systems detected cars with sensors embedded in the road. The cars had to come to the light to be detected.

It's the most reliable method, but it requires the car to nearly come to a stop first to get to the right spot.

It's easy to imagine more complicated computer vision or radar systems detecting cars. These are orders of magnitude more complex, though. When cities install traffic infrastructure they want to leave it alone and service it as infrequently as possible. It also needs to last for decades before replacement. Modern systems that are cheap, robust, and can detect traffic from a distance throughout different conditions like rain, snow, dark, and fog reliably are a recent invention. It's going to be decades before these are widely retrofitted to long-standing infrastructure.

Re: 200k Flemish drivers can turn traffic lights green

#157

Post by exiguus »

In amsterdam there are traffic light systems that prioritize their green phases and rhythms according to the number of traffic participants. For example, streetcars have priority 1, then buses, bicycles and finally cars. This works via weight or pressure sensors. No camera or app.

Re: 200k Flemish drivers can turn traffic lights green

#158

Post by frosted-flakes »

TYPE_FASTER wrote: post #86
OtherShrezzing wrote: post #13

I have no idea why the app is required here. Cars & bikes are massive, hot, & mobile objects. They're perfect for a modest-cost IR sensor atop the traffic light to detect them and adjust the lights accordingly. This type of system is commonplace in cities in the UK and works effectively. Having drivers reach for their phone when they're approaching traffic lights - common pedestrian crossing points - is categorically…

Do bicyclists and pedestrians really require the use of sensors? There are a lot of paths around me that cross busy roads with traffic signals. We just press a button. It speaks after you press it for visually impaired users and shows a red light after you press it for audibly impaired users.

Do pedestrians need automatic detection? Not necessarily, although a beg button adds unnecessary friction.

Do cyclists need automatic detection? Yes, absolutely. If you think otherwise, would you want to park and exit your car every time you want to cross an intersection? It's no different for a cyclist.

Re: 200k Flemish drivers can turn traffic lights green

#159

Post by robocat »

fn-mote wrote: post #130
robocat wrote: post #120

Earlier quoted context omitted.

And the reason only some maps apps work, is that Apple don't allow a background App to keep running that tracks GPS and reports location over data??? I'm guessing

Surely the reason is that neither Apple Maps nor Google Maps is willing to add the code to interface with the special Flemish Traffic Signal API. The market is too small, and there’s no profit to be made. Unlike the “you could take an Uber” ad in Google Maps.

Yes, but why do they want to be integrated with a maps App?

If it can work as a background App then why not? And if you're in charge of the Google Maps App, you know you'll get blamed for draining the battery and transmitting location data continuously.

It smells like someone in Belgium is lying or is looking for a scapegoat excuse.

The reasoning just doesn't sound right, technically.

Is there an international standard API for this? Perhaps one used mostly for emergency services? Neither Google nor Apple would want to code crap in for every single country's weird features.

Re: 200k Flemish drivers can turn traffic lights green

#160

Post by shkkmo »

gruez wrote: post #143
shkkmo wrote: post #140

Earlier quoted context omitted.

> just that the cartoonishly evil caricature of the driver's reasoning that OP dreamed up isn't accurate "Cartoonishly evil"?!? I'm pretty sure it is accurate for some subset of the driving population, which was the claim. > Unfortunately, judging by the downvotes it seems like people are interpreting my opposition to OP's claim that "Some people apparently just can't stand to have anyone driving before them" as me a…

>I'm pretty sure it is accurate for some subset of the driving population, which was the claim. It might be true, but that's not the same as "accurate". Let's imagine the same thing but for cyclists: "I saw a cyclist run a red light and I had to brake to avoid T-boning him. Apparently some cyclists hate motorists and want inconvenience them as much as possible." I'm sure there's probably "some" radical cyclists that…

> It might be true, but that's not the same as "accurate". Let's imagine the same thing but for cyclists:

I don't understand the disctinction you are trying to draw here. The statement itself is both accurate and true? At best the implication might be slightly hyperbolic but definitely not a "cartoonishly evil caricature."

> There's probably some far more innocuous explanation out there, like that the cyclist thought the way was clear (because there were no cars in sight).

Except your explanation is also not innocuous. Not wanting to be behind a bicycle is non an excuse for dangerous passing and especially when you are approaching a red light. The fact that you think this is an acceptable reasoning for that behavior is why you're being called out and downvoted. None of your excuses for how this might cause a light cycle to trigger sooner make this acceptable behavior.

Post reply on HN