I assume they are doing something much more clever/hardened, but you can trivially achieve much greater spatial accuracy with a Vive Tracking Puck for like $100.
Your positioning system needs to acquire a fix at least 100m out in variable atmospheric conditions on a rocket undergoing heavy acceleration and dumping all kinds of heat, smoke and vibrations into itself, the atmosphere, and everything around it.
In addition having a fix on your tracking device is only half the game, not you have to figure out where the rest of the rocket is in relation to your tracking device. Which again, vibrations, temperature and manufacturing all have an effect.
So while yet, a vive tracking puck isn't entirely unlike the workable solution it is also entirely unsuitable as a solution and should not be used as a baseline to measure off of.
Mixing up control errors with absolute errors is a very common form of miscommunication in robotics.
I work with relatively big robots and often my colleagues would say something like this "During the test we had 0.5m cross track error, so we did X, Y, Z ...".
And I always ask them for clarification. Were they looking at the robot and seeing that it is half a meter off where it should be, or were they looking at a screen and seeing that the robot thinks it is half a meter off from where it wants to be? Because those are two very different situations. And both can be described with the same words. (And sometimes it can be both, or just one of them.)
Control error is defined as the difference between desired value and measured value. So this is pretty good?
Even if they use some crude method to obtain position (e.g. gps), they can still easily refine that using e.g. triangulation using cameras around the landing platform.
Not sure what you are talking about. If you are asking if 0.5cm is good controller error for an orbital class launcher on landing? Yes, it is extremely good. Without doubt.
If you are asking about my tangential story where there is confusion between total error vs controller error then no, it is not good. Confusion is never good. Especially if the system is not within the total error budget. Because to improve it you need to know if you are dealing with measurement error or controller error.
> Even if they use some crude method to obtain position (e.g. gps), they can still easily refine that using e.g. triangulation using cameras around the landing platform.
Sure. I doubt that their total error is within 0.5cm, but both of their landings were extremely succesfull.
This always stuck out in an otherwise excellent bit, because you should definitely _not_ be taking the absolute value of your control error.
i.e. the booster doesn't know it's actual position to within 0.5cm but it knows it's position relative to a buoy or the catch arms to that precision.
But survey grade gnss is a web of rabbit holes, if you want to get into it.
And there are ways to get sub mm accuracy both relative and absolute, but idk of one that would be quick enough for the required reaction time of dynamic landing via 'catching'.
But multi-centimeter (4-5) that's really easily doable is probably good enough for other systems to take over from.
[1]: https://www.youtube.com/watch?v=pAPt5vbr-YU (don't recall timestamp, sorry)
No reason you couldn't use RTK GPS for <10cm accuracy for most of the flight, then in the last few meters of landing switch over to to high-precision, short-range tracking - like optically tracking a marker on the grabbing arm.
For other specific cases - like bridge monitoring - there are reports of 2–3 mm precision [1]. Of course, bridge monitoring has quite distinctive requirements; a 5Hz vibration component and a 0.0001 Hz thermal expansion component. So there's a lot of potential to average over lots of readings to reduce noise.
[1] https://www.sciencedirect.com/science/article/abs/pii/S02632...
Or in fact, you need even better than that, since you don't want your whole error budget used up by the GNSS system.
Anybody else thinking this quite the time to be alive?
You can see the arm comes in, then there's some side-to-side bounce (not sure how much is the rocket bouncing off vs. the arm fine-tuning its position). Just after contact seems to be made, and before the shock absorbing (or yaw-correcting) pistons drop much, there's a large flash from the engine. Is that a characteristic of engine shutoff, or was there a last-second "hover" push just before shutoff and drop? I wonder how much force the arms felt.
Another perspective showing both arms, and (as mentioned in the article) how the left one adjusted more significantly at first: https://youtu.be/JlcrNakUGVs?t=3
https://nextbigfuture.s3.amazonaws.com/uploads/2023/12/Scree...
I understand engineering is complicated but this honestly seems like the easiest part of the problem to solve.
It's more likely that SpaceX determined they didn't need super tight tolerances and called it a day.
The meatball Fresnel lens is canted slightly side-to-side, and only places the hook in the right spot at a given angle of attack. Which is a design compromise necessitated by having to allow multiple types of aircraft with multiple hook-to-eye distances to land on the same aircraft carrier while using a visual input in one location (the cockpit) to properly place a device in another location (the hook point) with high precision.
Source: I've done it.
So just as it is not "very easy" to trap on board the boat with "just" a light signal, I would assume landing a building-sized booster has a similar if not bigger list of potential "gotchas."
Yup! This is my conclusion in the article - the landing box for the Super Heavy booster is 5x13x18 meters on each side, with 5-15 degrees of angular tolerance in each of the vehicle axes. So the margins are big enough that you don't need millimeter level precision for the rocket position.
The actual question is literal: Can SpaceX land a rocket with sub 1 cm (1/2 cm) accuracy? GNSS RTK can get you down to a couple of centimeters, but getting more granular resolution than this isn't reliably possible with current professional grade technologies.
I'm personally unsure if the military has greater resolution than what's possible with RTK or w.r.t. military use GPS, but I would not be surprised if they did. If that's the case, NASA would most likely have access to it, I would assume. But the article specifically calls this out saying that it's not accurate enough to surpass the resolution of using RTK.
What's really cool about these questions is that the same problem space is applicable to self-driving cars and SLAM, if you're into that sort of thing. Lane detection, etc.
Edit: I think I misunderstood the comment. Yes, you can use the absolute methods for rough guidance and then use relative positioning for the final approach. The article has a line about why the author doesn’t think that’s likely though.
That is what they told us in missile maintenance school.
And gyros have gotten a lot better. Especially if you're throwing money at the issue like you know those folks are.
I'd expect that the rocket has a ton of sensors, and a ton of passive and semi-active tracking devices all over the body.
E.g. I'd put a bunch of NFC-type responders in a number of key positions, responding at different frequencies. Then a typical sweeping-frequency radar pulse would activate them all, and the response time and the Doppler shift would tell about positions and speeds of many points on the rocket. I'd do a similar thing with reflectors and IR/optical tracking.
All these points should follow some reasonable trajectory for some point the top of the rocket, near the chopsticks, would move towards some desired catch location point. Probably this motion is where "with precision of whatever cm" relates to.
/raptor 3 pumps look like you could hold them in your hands but iirc they deliver over 100k horsepower each.
I actually think there is some old Starbase tour interview where a SpaceX guy implied it was Musk's idea, though I could be misremembering. Catching the booster kind of makes sense, since they needed the tower arms anyway for stacking and unstacking.
I keep finding myself watching the catch every few days, and it does not tire to impress.
If you don't know how precise GPS receivers can get with dead reckoning techniques, this demo of someone "drawing" onto a map of their driveway using a GPS receiver is very impressive: https://youtu.be/3tQjIHFcJVg?t=245
It looks like they're getting measurements that are only a few inches away of the module's real position, although of course the conditions seem favorable with an unobstructed sky and consistent alignment.
The module they use is a ZED-F9P by u-blox. I've used ~$50 u-blox GPS modules in DIY electronic projects before since they're often the brand you'll get when buying GPS modules, but this particular type with dead reckoning is much more expensive. Sparkfun has it for $275 for example: https://www.sparkfun.com/products/16481.
[0] https://en.wikipedia.org/wiki/Real-time_kinematic_positionin...
While a sideways position error of even ten meters is not fatal, it is critical for the rocket to be quite close to zero altitude when deceleration brings the velocity to zero. (Any residual error must be dealt with by the shock absorbers, and their capability is modest.)
Good article. It is nice how it goes through all the points systematically.
https://youtube.com/watch?v=QHikx6kVvAo
It talks about how real time control system algorithms work with algorithms like like PID and MPC. I assume the SpaceX solution is likely one of the most advanced control systems in the world.
It becomes more technical, if one is specifically interested in the methods used by SpaceX. But there are some overviews that describe the general idea. Here is one presentation by Behçet Açıkmese: https://nescacademy.nasa.gov/video/eda2b96bddf945629be2c9d2e...
Note that before joining SpaceX to lead their autonomous landing software development, Lars Blackmore worked at JPL, where together with Behçet Açıkmese they developed the autonomous precision landing algorithms based on real time optimization methods. So, even though SpaceX has undoubtedly developed additional nuances to match their needs and capabilities, they were building on this prior work at JPL.
Maybe the 1/2 cm accuracy refers to the final position of the booster's catch points on the arms after they've closed, after the booster's engines are off, and after the booster settled, and maybe they mean lateral accuracy. I would forgive them for that because that's the accuracy that actually matters here.
If the catch points were off then that might spell disaster, so the catch points' landing accuracy including the help of the catch arms is what matters.
The booster rotation angle error and the catch point placement error were much too small to detect with the naked eye on the published videos. Every other measure of accuracy was clearly within tolerances -- and also hard to discern with the naked eye.
As amazing as .5cm accuracy sounds, if SpaceX meant catch point placement error, then it's quite as impressive because that only implies everything was within tolerance _and_ only the booster rotation angle error need have been impressively near-zero measure. That's... still amazing, honestly. If you can get the booster rotation angle error near zero then you can get the other errors way down too.
His choice of stainless steel is panning out well here - I doubt if aluminum or composite body structure would hold up as well to the "grab" forces from even minor misalignment. A composite structure would likely be entirely compromised by a big scrape.
It would be interesting to see a test where the landing speeds were deliberately too high - how much deceleration can the arms handle safely?
I think the chopstick mechanism is probably the best possible catch mechanism for such a tall object. The booster will be suspended from the top, which means the booster isn't subject to tipover as it would be if landing legs were involved. We've already seen this many times in the Falcon 9 booster series.
I can't see chopsticks ever working from a droneship, though - too much induced rotation for chopsticks to compensate.
As an alternative to chopsticks, a catch 'sleeve' might be possible, though it would magnify alignment errors considerably.
There are so many other methods that the lander can use to know where the tower is.
> Could you use other real-time distance measurements like laser rangefinding or visual processing? I don’t think so – the surface of the vehicle is too irregular to get a reliable fix point, especially while it is moving, and these are vulnerable to smoke/fog/gas/ambient lighting. Technologies like Ultra Wideband are vulnerable to multipath reflections and attenuation by the booster’s steel walls, and aren’t more accurate than RTK anyway.
That is not exactly an exhaustive list of methods to locate an object.
I have no idea whether 0.5 cm precision is feasible or even needed, but this part felt a bit off.
> At the most precise, an RTK positioning system could lower position accuracy all the way down to 2.5 cm (+1cm per km of distance). If SpaceX put a receiver on the launch tower or the ocean buoys, then the landing position could be incredibly accurate. But even the most advance positioning tech won’t guarantee it down to 0.5 cm. And RTK does rely on being able to acquire and maintain a link between the booster and ground for this precision.
I don't understand this last sentence. Afaik RTK correction only requires receiving correction frames on the booster's side, which can be distributed via l-band just like GPS. I suspect the latency constraints are also quite low as the conditions aren't going to change quickly near the tower with the kind of good weather they choose for launch.
It's still wildly un-nerving to me that there's no publicly stated option other than the chopsticks for landing(edit: some future passenger craft). Imagine if you've got enough fuel to avoid slamming into the ground, and a nice big ocean, or a lake sufficiently deep... couldn't a water landing happen and let future passengers survive?
But judging from the bouncing the rocket did when in the chopsticks the error for positioning into the initial catch position is much larger in all directions. The chopsticks coming closed around the rocket do the heavy lifting for final alignment to that 5cm I imagine.
Impossible apparently.
From the Article:
"Why can't SpaceX do a catch with a Falcon 9?
-It does not have separate landing propellant tanks, so propellant slosh will disturb its trajectory. The Super Heavy booster has dedicated central header tanks for landing propellant, so there should be minimal propellant slosh to disturb the vehicle attitude.
-It lands with a single engine which cannot throttle low enough to hover the vehicle, and as such must perform a “hoverslam” maneuver to bring the vehicle to a stop right on the ground. While the Super Heavy booster must perform most of a hoverslam maneuver to slow down just before coming in to the tower, it can hover for the final fine positioning.
-Because it lands with a single engine, roll control is minimal close to touchdown when the airspeed is low and the grid fins can impart minimal torque, and is limited to its weaker cold-gas thrusters. The Super Heavy booster can control roll with its 3 engines all the way to the ground.
-Falcon 9 has no engine-out capability for landing. SpaceX has not confirmed it for the Super Heavy booster, but I believe one engine out is likely possible (more on this later).
-It is smaller with a lower moment of inertia. Rockets get more stable and easier to control the larger they are, much like it’s easier to balance a broom on your finger than a pencil.
-It is smaller, and so thanks to the cubed-square law has a higher area:mass ratio. This means that it will be more affected by wind gusts that might blow it off course."
There are cameras and other electronics sending telemetry (via at least Starlink) on the things.
The "acquire and maintain a link" seems to be a very solved problem.
This talk was given at NVIDIA GTC 2015 and inspired me to go into manufacturing
https://www.youtube.com/watch?v=vYA0f6R5KAI
title: "GPUs To Mars: Full Scale Simulation of SpaceX's Mars Rocket Engine"
But of course this ignores that before the contract, SpaceX had already been working on this for many years. Raptor development began in 2014ish.
I would estimate SpaceX pays more then 50%. By reports they are currently investing like 1.5 billion $ a year. And the government contract can't finance half of that.
Also, of course in all of NASA history, this is the first time that anybody ever expected to invest themselves. The idea that a company would spend so much to build a moon lander for NASA is not something that was even an option a few years ago.
So why should tax payers be angry even if NASA paid 100%. Unless of course people are just angry that the space program exists at all.
My view is that Tesla, SpaceX, etc. are just cults of personality. Sure, they've produced a non-zero amount of technological progress, but for the most part they're leading us to dead ends.
Like he might have good point hidden somewhere, the issue is that we know that he will reflexively "debunk" anything SpaceX does, and predict failure every time. Meaning that his takes on SpaceX are beyond useless.
The SpaceX budget is several orders of magnitude smaller than the Apollo budget. Also, the Apollo era rockets were entirely unable to launch a large constellation like Starlink at reasonable cost like Falcon and Starship.
> Sure, they've produced a non-zero amount of technological progress, but for the most part they're leading us to dead ends.
Not sure whether you are serious. Assuming you are, what would be the better alternative then?
I know several space startups that's currently limited by launch prices
He has no credibility at all, like literally 0. Nobody that follows this topics, even if they don't like Musk or SpaceX consider him reputable.
He is literally a professional hater. He just continually moves the goal-posts to always present himself as 'the reasonable skeptic'.
> Currently, around 80% of SpaceX's rockets are dedicated to launching Starlink satellites.
And the other 20% are still more launches then the whole of the Western world would have had before SpaceX.
SpaceX is simply so fucking successful that you have to totally reconsider how you look at the numbers.
> and things like starship don't really have a reason for existing
And if Starship is cheaper, then Falcon 9, and that's literally the whole reason it exists, then making Starlink and other launches cheaper still makes sense. Stopping innovating because of market share is an idiotic thing to do. Its what people argue who don't understand the difference between absolutes and %.
The market for space launch and space economy will be much bigger in 10-30 years. Now we can argue how much bigger, but its defiantly gone be bigger.
And also, Starship literally needs to exists because the government want's to land on the moon. And using Starship and sharing lots of technology with a launch vehicle that is needed for Starlink anyway just makes a whole lot of sense.
> Also, in spite of massive improvements in technology since the 1960s, they're vastly underperforming the track record of the Apollo program.
What an absurd criticism is this? In the 1960 they basically had no regulation, 100% political support and literally unlimited money.
SpaceX had barley any money at all until like 2020 for this project and even then they could only invest a fraction of that into Starship.
If you compare Apollo, the NASA budget was around 20-40 billion $ per year compare to SpaceX who only crossed the 1 billion $ per year in like 2022.
So this is exactly what I mean when I say ThunderF00t has 0 credibiltiy. Nobody should talk arguments like this seriously. Its just arguments from bad faith. I suspect he knows that he is fully of shit, but he has makes a lot of money from continually telling everybody what they want to hear.
> My view is that Tesla, SpaceX, etc. are just cults of personality. Sure, they've produced a non-zero amount of technological progress, but for the most part they're leading us to dead ends.
ThunderF00t is closer to a cult of personality. He has achieved absolutely nothing, he can't even edit videos as well as most 13 year old on tictac. He is a man in his 40 and his humor is even more childish then that of Elon Musk.
And your claim is simply bananas. Calling landing rockets and Starlink 'non-zero' is the understatement of the century. Starship has literally revolutionized space internet and global connectivity, and literally everybody, including all their competitors and every government in the world understands and reconizes that.
And literally some of the largest cooperations on earth, plus major nations are trying to replicate it. Amazon is investing 10+ billion $. Europe, China, Russia all will try to replicate even part of it, but they all know they can't match it.
Lower LEO internet will basically not go away for the rest of human history, unless civilization crashes or somebody comes up with something even better, but we have no idea what that would even be. But somehow you call it a 'dead end'. What are you even talking about.
I mean seriously, please break out of your bubble and look at the world with objective eyes.
Eg if you space sensors 20m’s apart.
Love the analysis.
The tricky parts (that we don't really know as non-SpaceX employees) are:
- how accurate is the clock onboard the satellites? Given that it's likely an OFDM signal the timing is probably pretty good, but given that they're launching zillions of them they probably don't all have atomic clocks onboard
- how accurately is SpaceX tracking their orbits? Kind of a similar answer here... they're doing beamforming to the ground terminals, so it has to be pretty good but we don't really know how good.
- how many SVs are actually visible at a time? We need a minimum of four but the more the better. If there's lots visible we can somewhat work around the first two issues statistically but if there's a limited number than the orbit and clocks need to be super accurate.
