Hacker News new | past | comments | ask | show | jobs | submit login
Gatwick airport trial for valet-parking robots (standard.co.uk)
56 points by based2 on Jan 26, 2019 | hide | past | web | favorite | 53 comments

> [...]and uses military-grade GPS to ferry it to a secure bay[...]

Something tells me Gatwick isn't actually in possession of the decryption key for the US Air Force's P (precise) signal and this is just some PR person taking it too far.

It's a myth that civilian GPS is less precise than military. A modern, dual-frequency or augmented civilian GPS receiver is just as precise as a military GPS.

P may have once stood for precise, but now days the main advantage of the encrypted, military P signal is that it prevents spoofing attacks.

It used to be true but in 2000 Selective Availability was disabled which made the civilian signal just as precise. This is basically why GPS took off around that time despite having been in operation for two decades at that point.

It's just one of many whys. Differential GPS was widely available and it gave cm-level positioning precision.

Equally if not more important why was your good old hardware minituarization and its cost reduction. In the late 90s a decent GPS reciever was still a size of thick book and cost prohibitively.

That would have been no problem in a car. But pre 2000s civilian GPS was useless for street level navigation.

No, it wasn't useless.

Again, coupled with differential corrections (available for free over radio) the resulting precision was in centimeter range. The reason it wasn't used was the physical size and cost (for hand-held navigation) and the lack of supporting hardware in cars (no user-facing on-board systems with screens and UI).

SA was _not_ the reason for the lack of consumer-level GPS adoption.

In my experience using DGPS around 2000, most of the DGPS sources were near the coasts and operated by the Coast Guard. There were a few scattered transmitters here and there, certainly not enough to provide the kind of coverage we now enjoy.

It's probably safe to conclude that they don't, but you don't need COMSEC crypto keys to achieve "military-grade" accuracy down to 10cm either, i.e. differential GPS[1].

[1] https://www.effective-solutions.co.uk/beacons.html

Yea I was on a project where a professor lent us a Trimble GPS (I think they ran $2k ~ $4k at the time). It had a backpack, synced up with a tower at the local airport and gave us information down to a couple of centimeters.

Are there any meaningful uses of that term? It's a fairly reliable indicator of cluelessness when followed by "encryption."

I wonder how they got GPS to work so well in a concrete car-park.

It's the long term car park, open outside not multi story and at about half a mile from the terminals. They are removing loads of lamp posts though...

Maybe they’re using GPS as a generic term, can access to military-level precision be bought for the non-US satellite positioning systems?

Probably not, since the main benefits of the military signals are improved resistance to jamming/spoofing, and having a signal that can continue to work if the civilian signal is turned off. If you sell those keys to random businesses they become less effective

This article desperately needs a video, here's a nearly 2 year old video from the same company:


Your video is the only non-animated one i can find, and really, its just of a dolly bumping into a car, then cut to 'car being on dolly' and disappearing out of view. All the pictures in the article seem taken from the animated promotional stuff for stanley-robotics.com which itself seems unrealistic as best. And even the 'already used in France' bit is misleading, as it was a trial at charles de gaulle with no information on the scale or success.

Their site looks like its there to attract investors rather than sell any kind of real product.. i might be wrong but i don't think this is going to amount to anything, shame.

Here is a non-rendered video where the lifting happens: https://www.youtube.com/watch?v=bocfNOjeBm8

I don't see how 2 tons are being lifted from such a low and thin surface. The whole thing looks more like a Kickstarter than a real product.

Edit: Nope this is 100% legit

Maybe something similar as pallet jack: https://en.m.wikipedia.org/wiki/Pallet_jack

Example of an electric one: https://youtu.be/w7MG3DOwAhQ

Quite close to this Stanley device. They claim capacity of 3,5 tons. Forks are not as long, but the mechanics seem such that extending them should not be a problem.

same replies as above. Our robots have been receiving in our opened service (Lyon airport) on a daily basis vehicles up to 3 tons (incl. heavy SUVs with hybrid powertrain). Our mechanical spec. is higher though

Can't find anywhere in Lyon airport to book this service;


Also they seem to use ectorparking.com at Lyon, are you affiliated with Ector?

Also no sign of it on Twitter, seems like people using it would likely talk about it on social media (maybe it's a private beta for airport employees?)

we understand your point, please see also our comment above. we are extremely serious and cautious people, and have not even started communicating ourselves regarding the Gatwick project. Our robots are very real, and impressive according to the professionals & users knowing them. Kind regards, Stanley Robotics

here Stanley Robotics, we are the makers of these robots. You are fully right, and btw this article is clearly not from us. We are careful in communication. Big news, this time by us, are coming in 2 months from now. The very same robots as the ones for London Gatwick, have already behind them more than 1 year of daily operational experience: at Lyon airport(France). Currently 150 parking spaces, managed by those robots, are managed every day there. This number is increasing rapidly week by week. Stay tuned, thanks :)

Elsewhere in this thread someone posted a few videos that make it much easier to understand what's happening. https://news.ycombinator.com/item?id=19011731

Have you had any communication with tire manufacturers about how you lift the car that way?

Wouldn't that put unusual stress on the tire plies? Maybe once or twice might be OK, but if someone did this to tires 500 times a year (e.g. commuting to work, once at arrival at garage at work, once at departure from work) would the tires still be OK?

I found a VR video (which needs dragging around quite a bit on my computer to find the desired angle): https://www.youtube.com/watch?v=zIaxPa2aUDI that showed the car being put into its slot. Looks pretty neat.

thanks. but really old now already, and it was in an easier - indoor - environment. far better coming very soon, please stay tuned ;)

In the distant future, I might be able to just get out of my car and walk away to let the robots do their thing. But in the near future, any time saving is negated by standing there watching the robo-magic :)

Huh ... so that answers my big question: What about Rear-wheel-drive and AWD cars? AWDs are difficult to dolly. Tow truck drivers (that don't own flatbeds) have to either disconnect the rear differential or put dollies under the wheels.

Here’s one from Lyon: https://youtu.be/V4bgiMP_WfM

I wonder if the movement triggers car alarms to go off?

That's clever. But all it does is rearrange the cars in the parking lot. It doesn't pick them up at the terminal or deliver them back there. Good first step, though.

thanks. what you say is a nice vision, level 5 of such AGVs ;) and as you say, the prior levels are already pretty interesting (our customers, the parking lots owners, saying) as first steps

London's other major airport has a very cool autonomous rapid transit system.



One benefit I can see from this:

Car owners probably drop off their car and then get valeted into a spot.

This means the car is started and ran for a few minutes before being shut down again, draining some precious battery power that isn't recharged in time.

Repeat when the owner requests the car's return.

This may make the difference between a car that starts when its owner returns vs not.

This seems like a non-problem. Your car battery is going to be full when you arrive at the airport. Two extra starts isn't going to cause a problem.

Apparently some modern cars with a lot of electronics do actually have a problem keeping their battery alive for just two weeks in an airport garage!

These robots and electric car chargers could solve the daily game of musical chairs in corporate parking lots. A single human attendant could handle the "last inch" problem.

I assume I have to leave my car off, but in neutral with the parking brake off? How else could they drag around an AWD/RWD car by just pushing it around from the front?

The vehicle isn’t pushed; all four wheels are lifted off the ground.

There is a risk if the vehicle is dropped off on level ground that the driver forgets to put it in gear+parking brake.

Now if it's parked onto a slope or exposed to heavy winds...

That's hardly obvious from the pictures in the article.

In addition, cars have very precise jacking locations which must be used to lift them without damage. Those bots probably don't know where to jack every particular vehicle. Do they just jam a dolly under the wheels?

I'd give this a hard pass. It is probably just a way to separate some money from some dumb VCs.

One picture in the article shows that the sled has lifted the car by the wheels. The video posted above shows this more clearly, all four wheels: https://www.youtube.com/watch?v=e3hmbYLQedc

I guess that much tighter parking is the goal, to make this pay, but it's not clear how far they can go here -- it looks like every car can be accessed without moving others, needing clear aisles.

Also, the column spacing in many buildings is designed around parking-bay sizes, which will limit how much you can save in any existing building. And it would be one hell of a 50-year bet to design a new building only for this system.

The video posted above shows this more clearly, all four wheels

The video shows nothing. It only shows "before" and "after", not "during". It's totally fake. The video doesn't answer the question of how the wheels magically get on top of the sled.

Is the body somehow jacked up to lift the car off the ground? If so, how is this done safely? Most unibody cars can only be jacked up by lifting at very specific points. You can't just slide a large piece of metal underneath to lift the car. Well, you can, but you'll probably damage the car.

It seems pretty clear. The fingers on the sled go underneath the wheels of the car.

Watch 1:12 in the video. You can see the white-tipped prongs around the wheel

Here's a side view.

   _O_    _O_

Come on, guys. It's emphatically NOT CLEAR.

There's a jump between 1:11 and 1:12 in the video. Just before, the sled bumps into the tires in the front of the car. You can actually see the car physically jerk back as the sled hits. Just after, all four wheels of the car have magically appeared on top of the sled.


If this were as simple as everyone thinks, there would not be a jump in the video at the exact point that the magic happens!

Thank you. Those videos illustrate the concept much better than any others.

You keep saying "on top of the sled".

The sled moves under the car. When the sled is under the car the fingers extend under the wheels. They do not, cannot, extend under the part of the wheel that is in contact with the ground. But they can, and do, extend to the parts of the wheel that are close to the ground. Again, look at 1:12. You can see the prongs go around the wheel, not directly underneath it.

You can see the prongs go around the wheel, not directly underneath it.

The videos linked to by improbable22 make it easier to see this.

My (very limited) understanding of radial tires (or at least steel belted ones) is that it would be bad to stress the tires in that manner for long periods of time. But lifting the car by the tires this way for a few minutes should be OK?

It would be interesting to learn what tire manufacturers feel about this. (Of course they probably would want to disclaim liability because there's no reason for them to want to bless this).

> He said that passenger feedback was “overwhelmingly good” during a five-month trial at Charles de Gaulle last year, where the system operated so seamlessly that some customers were not even aware a robot had been involved.

Only 4 robots? Why so few?

And only for 4 travelers? Why that arbitrary limitation? Why can’t they handle 5 travelers?

Where does it mention 4?

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact