Hacker News Reader

Comment by phkahler | Hacker News Reader

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.

View on HN

We have those in Poland. Sucks if you are on a bike. You need to wait for a car to turn it green for you.

In the Netherlands we have ones that detect warm bodies.

Sucks if you are a zombie...

Here in Zombieland we have cameras that can detect any bipedal movement.

Sucks if you drive a car

There's a road closure near me here in Rotterdam which has encouraged some people to get around it by cycling on a short patch of the tram track. It's quite funny to hear the "tram approaching" bell ringing only to look and see a small boy on his bike.

If they're using inductive vehicle detector loops in the pavement, those can be tuned to detect motorcycles and bicycles (probably not carbon fiber bikes though). Where I've seen and used loops tuned for bicycles, they have a different cut pattern and a painted bike to accent positions where the detector is likely to detect you... Sadly, they don't tend to have far back detectors (advance loops) tuned for bikes so an idle intersection won't be green when you get there like if you were a car. You can sometimes get a detection on a car tuned detector by riding right on top of the wire, but it's iffy.

I'd imagine camera detectors can be tuned for bicyclists as well, but I have less experience with those.

Here's a nice survey of bicycle detection options [1], if you're willing to work with whoever does traffic control near you, this might help get a conversation started. This is focused on US applications, but the ideas should be broadly applicable.

[1] https://altago.com/wp-content/uploads/ALTA_Bike_Detection_Wh...

My city has inductive bike sensors at specific intersections but many intersections (even ones with bike lanes) lack them. Cycling infrastructure is so half-assed in the US.

Cost per benefit? My friend is a commercial property guy and sometimes has to bring properties up to date for handicapped access.

He jokes that with how few handicapped people use the features, he could probably give people $1k each as an apology for not being able to go into the building and be far ahead.

I started thinking that for the most part most handicapped people would even take that deal. I think I would.

I don't think it costs that much more to use a D or Q pattern for the inductive loop instead of a circle for new loops. But, if the loops are already there, it is a lot of expense to put in new ones.

Adding a loop to bike lanes might be reasonable, or maybe it forces a controller upgrade and that's not reasonable... but if you make the lowest speed through lane and at least one of each protected turn lanes bike accessible (while otherwise doing loop work), that's good. I think the loops are pretty shallow, so if the road has significant rebuild/repair, it's time to address the loop design.

The key word for ADA is reasonable accomodation. It does add expense, and not every feature is used often, but it's 35 years since the ADA passed and the US is a lot more accessible now than it was in 1990. The window of reasonable changes over time too, of course.

Definitely didn't work for my carbon bike. It's annoying because sometimes those are put in place on very minor - major road crossings and those are dangerous to cross on red light (and the minor roads are exactly the ones you want to be on the bike). I had 2 of those on my usual loop when I lived in Poland. Sometimes I just waited for a car but especially on weekends I needed to risk it and run the red light as otherwise I would be stuck there for a few hours.

> Definitely didn't work for my carbon bike

Didn't work for my 200kg metal motorcycle either most of the time.

Since when did cyclists take notice of traffic lights?

If you think cyclists break more traffic rule than drivers your view or reality is distorted. Try looking at the data and then around. You just notice them breaking the rules because they are different than you. It's a bias worth correcting.

I notice them when I cycle. Most common infraction, which I can't understand due to the sheer danger, is running red lights. Wouldn't occur to me in a car, much less in a smaller vehicle.

Different than me? When I used to cycle, I skipped through traffic lights, when it seemed safe to, and I've seen many other cyclists do the same. It's not a bias, it's an observation. But I mainly said it as a joke, because it's a well known stereotype, for good reason. At least where I am.

I have one in my city in Canada. Basically, on the bike you have to mount your bike from the bike path onto the sidewalk and press the pedestrian walk button. That turns the light green. Extremely annoying.

So they assume perfect vehicle detection? Sensors should be used to speed up a regular cycle. They should not be the only thing changing a light.

Presumably the app would be able to signal the presence of an approaching car from a much longer distance than the inductive loops you typically see these systems, ensuring cars don't even need to slow down when approaching the intersection. (Saving gas, brake/tire wear, and time.)

It's also probably cheaper than the extra physical infrastructure needed for inductive loops, given that all the necessary equipment (phones, GPS, cell infrastructure) is already in place for most people.

>It's also probably cheaper than the extra physical infrastructure needed for inductive loops, given that all the necessary equipment (phones, GPS, cell infrastructure) is already in place for most people.

Assuming the procurement/contracting for said app is done competently, which seems questionable for most governments. Moreover even if it's cheaper, surely some sort of camera/sensor setup would be safer than drivers fiddling with an app?

I'm guessing they wouldn't need to fiddle with an app; the software would just automatically signal their approach based on GPS/navigation info.

I agree theoretically a camera + machine-learning based system could work just as well; I've always wondered why that wasn't standard on traffic lights these days given that the hardware would probably cost a few hundred dollars (if procurement for said hardware were done competently, which I suppose as you said, it probably wouldn't be).

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.

Are we are the stage where it is normal to be giving governments always on background access to our gps location voluntarily?

I mean, if your concern is government, your already shit out of luck between your cellphone pinging cell towers, license plate readers, CCTV cameras, Ring cameras, Wi-Fi pinging, etc.

(I would agree that I wouldn't necessarily want fine-grained gps location data being sent to the lowest bidding for-profit traffic light monitoring company.)

I can think of a few ways to do this in a privacy-preserving way. The traffic light doesn't need to know your location or who you are, just whether there is a car approaching. But yeah good chance whoever designed this didn't put that much thought into it.

> Are we are the stage where it is normal to be giving governments always on background access to our gps location voluntarily?

Yes. See Palantir for examples.

The drivers don’t “fiddle with an app”. The apps use geo location to notify the traffic control system when they are near an intersection.

https://www.brusselstimes.com/512118/flemish-traffic-lights-...

> As soon as a driver approaches an intersection with smart traffic lights, a signal is sent via their app and the cloud to the computer controlling those traffic lights.

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

AFAIK background location is opt in, but only requires the developer to check a box in their xcode project. It's not like you have fill out a form and have apple approve it or something.

https://developer.apple.com/documentation/corelocation/handl...

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.

Probably because maps apps can do useful things like tell the light whether you're intending to turn right or not.

Because users don’t want to install apps that have access to gps in the background.

> a signal is sent via their app and the cloud to the computer controlling those traffic lights

No network, red light. /s

Flemish person here: no one here understands it either. Probably CV-driven development for someone...

Or the meeting started with the open question: "How can we get our citizens to give us 24-hour location data on their whereabouts? How could we possibly spin it as something beneficial to them?"

So, uhhhhhhh, gps spoofing time?

> 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.

We've had traffic lights like that in the UK for ages. Where I am in the States these days, they're very common. There's one very effective one that manages my turning from the residential street onto to the main road. I was getting irritated at the weekend because I came across some lights that didn't have sensors. 5 cars stuck waiting at the lights for literally no one.

Of course, the other thing they do in the UK is use small roundabouts, that can handle a fair amount of traffic nice and smoothly.

I’m guessing that with location access they can probably do it without adding any infrastructure, just an API.

> I don't understand why they need apps for this either.

Data.

You would be amazed how bad drivers are at actually hitting the correct stopping position for the inductive loops in the ground that sense cars. When I'm on my bike stopped on a small side street I often try and gesture for drivers to pull up so their car is over the vehicle sensor in the pavement and the light changes faster. Drivers have gotten angry and yelled at me for giving them a little "inch up" hand gesture.

Reply on news.ycombinator.com