I assume this was done with a FPV piloted drone. I just got into FPV flying and this was pretty impressive! My stomach lurched a few times recalling times I've tried to squeeze into tight spaces only to drift where I shouldn't be.
There's lots of cuts and morphs, so it's not continuous. Also lots of speed adjustments. Many of the shots were probably done very slowly and sped up after.
No reason that it needs to be FPV - if they have a 3D model of the factory they could have created a path for it to fly. I don't think Ardupilot or Betaflight is up to that(?), but a small team of engineers working for a couple of months would be able to make it work.
This is almost certainly FPV as the cost of building an autonomous solution would be incredible. It's more likely parts are CG than that the camera vehicle flew autonomously, and I think neither are likely, this looks like "top 5% FPV pilot" work to me.
Inside-out indoor 3D positioning is an _exceptionally hard_ problem. Like, really hard. It parallels inside-out road positioning enough that there is crossover (and IIRC there is a lot of employee motion between Tesla and DJI and other drone autonomy makers), but I don't think this would be an easy "few months" project and certainly not cost effective.
Picture drawing lines in 3d space in something a little more sophisticated than SketchUp, and then having the drone follow those paths. What's tricky about that? Indoors you have no weather, and could use a beacon system to pinpoint location.
And this is the typical dismissive rebuttal of an HN post.
They built a factory using a high degree of automation in a 3D space - i.e. they have a significant staff of automation engineers.
Intel has been doing tight formation drone light shows outdoors (in weather) since at least 2014, if not earlier. Positioning drones precisely enough (i.e. within N centimeters) is therefore objectively possible if you have the right feedback system (and as a sibling comment points out, IR sensors could work for that). Offload all the processing to a computer rather than the drone, and it isn't an insurpassable problem.
A small team of engineers for a few months is <$1m, which for a marketing campaign this splashy is nothing. Granted, per one of the sibling comments they did use an FPV pilot for at least some of the shots, but there's nothing stopping them from having automated other parts of it.
yeah or it could take a week of planning a week of shooting and a week of editing with a crew of people that make drone videos for a living… they’re not pulling automation engineers off of their projects to work on something completely unrelated for a one-time promo shoot, especially not for an entire month
An array of active IR cameras in fixed, known positions, plus IR reflector markers on the UAV works well out of the box over ranges of a few meters. Eg. Vicon. Same tech as motion capture for movies.
I don’t have experience using it in volumes as large as the video, but in moderate sized lab space it works very well, and at high frame rates.
The trick is to invert the problem and have the active sensors off-board and the robot is the (reflected energy) beacon.
How do you propose you use outside-in to fly through an operating robot? There are too many sources of occlusion for off the shelf outside-in solutions to work at all in this video.
It sure was a nice video, and I found it insightful to see what goes on in a car factory. But I have honestly quickly gotten spoiled by seeing some really good FPV videos, recently, it seems that every self-respecting company wants/gets one, these days. And with that in my memory, I found the cuts surprisingly off-putting, as silly as that is, I realize that. On the flip side, how close they were allowed to go into the molding and other machines, I found very impressive, imagining that a steering mistake could cause disruption and cleanup, although it might also speak to how much buy-in you can get from the people on the floor for making a cool video.
I'm just getting into FPV RC cars! I was pretty blown away that there are basically two choices: low res analog or DGI's digital system. Side by side here: https://youtu.be/1POaGs99y-M?t=79
I always assumed there were some neat tricks to get HD from multiple analog channels, after all these years, but none are remotely affordable.
I'm tempted to slap an old 4g android phone onto the car (or two), and turn the cars max speed down, and use that for highish latency, but nearly infinite control range. With a little solar panel stopped on, infinite battery life, with some long, periodic, naps!
Hobbyist: RubyHD, OpenHD (make your own system using raspberrypi zero/jetson nano etc..) and get a decent connection by using off the shelf wifi hardware in monitor mode. Can cost up to $100 - $150. (2 raspberrypis, 1 camera, wifi dongles) - https://www.youtube.com/watch?v=oFGsZ_d3KBw
Commercial: HDZero (proprietary low latency HD system that is fairly popular with FPV quadcopter racers these days). Can cost up to $300. ($50 for camera, $50 for vtx, $150-250 for receiver with hdmi out).: https://www.youtube.com/watch?v=CHa08sdKoGs . They recently released a 1Watt vtx that costs about $100, and can give you a fairly good range.
If analog is potato, then those systems look more like yams. Their effective resolution, in the optical sense rather than the image size sense, seems to be really low, compared to DJI: https://youtu.be/llRaTs3bVuE?t=12
Definitely cheaper though!
disclaimer: I just bought a DJI headset. I may be biased and trying to justify my purchase to myself. ;)
Yeah on paper HDZero and DJI both do 720p, but i think that's just the number of pixels. HDZero does seem to lose more quality than DJI due to the way it encodes data. There's also 1080p30 mode on HDZero - but there don't seem to be any cameras that make use of it. For now I am disappointed with my HDZero system: I still can't see ghost branches and electric cables in time.
The video you linked doesn't represent HDZero fairly though, because they're 2 different cameras - with different settings. Here's a more accurate representation: https://youtu.be/SmfbMTe80Uc?t=126 . (So far, of all the HDZero cameras I have tested, I liked Digisight V3 the best)
RubyHD seems pretty good - except for latency. On a better hardware I think RubyHD would be as good as DJI.
Also do check out the recent root exploit for DJI goggles. It will soon bring you "Canvas mode" (OSD like we had on Analog systems). Happy Flying!
I am still waiting for Orqa to come out with their HD system.
I definitely had some uncanny valley vibes watching it, but I think it might be a combination of the camera's perspective (distorting distances away from center) and the footage being sped up/slowed down.
More like footage being stabilized in software like Reelsteady / Gyroflow . Typically fpv videos are a little more "jerky" and these software remove the lens distortion and then stabilize the footage.
