Can you do Orbit Determination in Kerbal Space Program?

[@jamieloganaerospace6404]

As promised, my long awaited first voice-over video is here! I hope you've found it to be informative and entertaining. My question to you all is: how should I follow up on this? Should I use this to try to perform orbital rendezvous without using map view? Let me know your thoughts below. KTDS (KSP Trajectory Determination System): www.desmos.com/calculator/q6q0npozf3 Special thanks to Yukon0009 for letting me use his JWST replica for the opening shot. His stock KSP replicas are second to none: Video: kzbin.info/www/bejne/baOnnoRqr5tkoZY KerbalX: kerbalx.com/Yukon0009/James-Webb-Space-Telescope-+-Ariane-5

[@virutech32]

Orbital rendezvous without mapview sounds like a special corner of hell. unless it turns out to be really effective. idk it sounds pretty awesome. I can just imagine having a fuel depot in space & automating launch times, accent, & rendezvous of fuel deliveries. Actually i can see a lot of use in learning how to precalculating maneuvers & stuff. could make for a very interesting style of play.

[@Crimsonedge1]

@@virutech32 Its easy enough. Just ensure you're on the same inclination as what you're trying to intercept and then manoeuvre so your apoapsis is higher than the orbit of the target and so your periapsis is at the same height. The higher your apoapsis from the target orbit, the faster you'll make the rendezvous but that also means you'll have more velocity to zero out when you arrive at closest approach. Then just wait in orbit until you have a close approach of less than 10km. From then, wait until you're half way towards the rendezvous on final approach and then start using your OMS to bring the distance down until close approach is less than 1km. From there, switch to target mode on the navball and burn retrograde relative to the target to bring your velocity close to zero. Then just angle towards the target and use the OMS to bring you in for docking. Doesn't work as well if your target's eccentricity is greater than about 0.1 as you could be waiting in orbit for awhile waiting for a close approach.

[@aviin_videos]

0

[@mrhawkyy1]

There is a mod called Prinicipia you should try, it adds n-body physics to KSP and would make this challenge alot more realistic/challenging

[@ianhogben3472]

is the ground anchor stock

[@notfeedynotlazy]

"Hey guys, I'm an actual space engineer" Dude, -that's actually flexing- that explains a _lot_ of things!

[@scottmanley]

Fantastic job, I'd always wanted to do something like this, but hadn't considered doing this with stock in game robotic parts!

[@badethics7542]

Somehow, I knew I'd find you here.

[@Not.a.bird.Person]

Great to see the myth, the man, the legend Scott Manley here! I really hope you'll read this, though I realize it may get lost in the amount of messages you receive. I started playing KSP around 8-9 years ago and went to study in aerospace engineering because of your videos. I was just hired recently by a company to do mechanical designs on satelites as a junior engineer. You don't know me, but you personally have had one of the biggest impacts on my life. I really want to thank you specifically for this. If I hadn't watched your videos on KSP, I may have never discovered the game which ignited my passion for aerospace. From the bottom of my heart, thank you sir!

[@vovanikotin]

spAtnik!

[@40watt53]

Get yourself a man who can do both. Perfect Kerbal player and a god at Desmos.

[@jamieloganaerospace6404]

Every knee shall bow!

[@lorenzzorzt]

God is great

[@KarusKardin]

Despite being more confused than a ant in a supermarket this man managed to keep my attention the entire video

[@andrewcharles459]

Right? I probably looked like a four-year-old watching Teletubbies when I watched this. Totally engaged.

[@VercilJuan]

Wait a hot minute, did you just make... a stock ingame sextant? Jesus Christ that's brilliant! Why couldn't I have thought of that, all my Kerbin navigation problems would've been a breeze. Thanks for the idea mate!

[@VercilJuan]

Also, new sub! You've been putting out good content!

[@maxovercharge2961]

I love when actual space engineers play ksp.

[@mgaerospace2209]

Wow! This is seriously impressive! That JWST looked spot on

[@jamieloganaerospace6404]

All credit to Yukon0009 for that beauty. He’s the stock replica GOAT in my opinion.

[@mgaerospace2209]

@@jamieloganaerospace6404 Is it to scale?

[@yukon09]

@@mgaerospace2209 it's 1:1!

[@mgaerospace2209]

@@yukon09 Wow 🤩

[@maxwhite4732]

I thought it was an actual picture of the JWST edited in lol

[@jamdonut]

Jamie: Proceeds to prove and explain how it works Me: I'll take your word for it.

[@virutech32]

Man that was awesome. I love it when proper science is done inside games. It's just nice to see. Great way to get familiar with a topic too given how many hours people sink into games like this. Keep up the awesome work👍🏼

[@jamieloganaerospace6404]

Thanks!

[@notfeedynotlazy]

@@jamieloganaerospace6404 Games that teach science by DOING science are the best.

[@-_-4624]

This guy inspires me! Cuz rn im a 18 yo that is trying to get into uni for aerospace engineering and eventually want to work with a space agency! I found this passion through games like kerbal space program and other games like it! So it’s also great to see that a JWST engineer also thinks the same!

[@davee844]

Do it buddy. It was the best choice in my entire life to study space engineering. Of course it's not easy but if you're passionate about it you can do it.

[@NoSTs123]

Thanks for your interesting video, it was a real treat to watch. BTW I recommend the mod Distant Object Enhancement which allows the game to draw vessels which are incredible far away. Distant Object Enhancement paired with a telescope mod which enable really high magnification I was able to spot a mun arch and my fairly big lander next to it from the surface of kerbin albiet only for split seconds at a time due to missing tracking technology. Camera tools might be better suited for this. I will be checking out your desmos

[@alexwang982]

Use InfernalRobotics if you want tracking, and set it up on a equatorial mount. (possibly with 0.01 gravity cheat...)

[@christianhinton2030]

This is fantastic, instant subscribe. As an aerospace engineer I love seeing stuff from classes I took being directly applied to KSP.

[@redneck472]

I only understood about 20% of what you were saying but as a casual KSP player, it was very interesting to watch someone perform actual science in the game.

[@nintenplsdont128]

this is a really nice perspective on a part of ksp which isn't as commonly explored :) very neat explanation and video!

[@fatitankeris6327]

In reality you indeed don't get the info as in KSP. Truly amazing how that's done.

[@TalpaDK]

This is some Scott Manley grade entertainment, definitely warrants a subscription

[@GigaKerbal]

in the action groups menu under the second section, you can set the motor movement to the rotation shortcut, and simply turn SAS on and select target, allowing you to quickly and easily pinpoint the target, keep it locked, and use some mathematics to determine the orbit from orbits further from 100km away.

[@dennyrulos7370]

You could use the Physics Range Extender mod. The craft will stay simulated and you can target is for longer

[@phoenix0166]

Fellas, this is the moment we’ve all been waiting for! A VO *AND* an awesome tutorial!

[@ShadowZone]

That was really educational. Thanks for showing us how it's done!

[@cartler]

This is why I love Kerbal! It's really easy to play, but has so much potential that real engineers can enjoy playing it

[@everettmassic4050]

I started watching your channel when i saw your HUGE Apollo mission. now, well... IM HERE TO STAY!!!!!!

[@joeymurphy2464]

Wow, very cool video! I really enjoyed the perspective of "This is a thing we do in real life, it's called Lambert's Problem, let's use KSP as a way to visually show what's going on". More of this style would be super cool to see! Maybe discussing how pointing works and a small intro to quaternions and why they matter?

[@thetechnoninja3859]

Your job is actually what i am trying to achieve currently. I am going to college for aerospace engineering and i would love to work with satellites and orbits. Hope you enjoy your job, to me that looks like a ton of fun and something i would really enjoy doing.

[@ivolol]

This is pretty darn cool. A tour of the maths involved would be very neat (I assume mostly rejigging and combining trig / spherical / elliptical / kepler orbital properties to solve for stuff). I find "space math" some of the coolest "real application" of semi-advanced math around. On a separate topic, describing what the main pertubations are for JWST' orbit is, and what the main measurements and their mechanics / technology used are to update its orbital solution, maybe what software you use, would also be a very cool companion video.

[@samcg1492]

Wow, sincerely this is so impressive, I love when people use KSP as a simulator to show things up, I loved the creativity of your tracking station instrument

[@buddysadventuregame1255]

2:44 that has to be the smoothest landing in ksp ever

[@synthicalmix]

Amazing job, best video yet (apart from the chinook, which was my favourite)! But omg, this video was so ingenious, it really taught me about OD! Thank you for an amazing video!

[@Ipwnpwnage1337]

Awesome video! When I saw the title I thought you were going to do filtering in kOS or something, but for KSP’s dynamics that would be overkill! Using this to do some maneuver design would be really neat, and like this video I think it would be very informative to demonstrate the concepts behind real nav.

[@samspeed6271]

This is very, very cool. Definitely going to try this out and experiment a bit with this in KSP

[@D.C.7.3]

Incredible video!! It’s people like you that inspire me to become an aerospace engineer. Please keep up the amazing work.

[@stvden]

This is very impressive, hope to see more videos like this, also it would be cool to see the math behind it in video

[@JoelleTheAbsurdist]

Distant Object Enhancement might help out for more... distant objects... I'd love to see a demonstration of say, predicting the orbit of a helio-captured object from a Kerbin(or other) base station.

[@andrewcharles459]

What I like most about this video - apart from plain admiration of the skills I will never possess - is how it demonstrates just how versatile the game itself is. Application of real world techniques which were probably not envisaged by the designers is so satisfying. Sadly, in a world of such impressionists, I draw stick men. But I still like to go to the museum. ;-)

[@donchaput8278]

A small piece of my mind just exploded. Well done!

[@SofieBrink]

This is really really cool, love this kind of stuff

[@thorodinson7573]

this is by far the most instructive ksp video i've ever seen

[@tophertycoon]

Amazing! love watching your videos since the start!

[@anthonyzawacky2184]

I use the planet I am orbiting as a bore sight, I line up the target perfectly with the planet im orbiting, then put the maneuver where the planet touches my orbital path. I'm no engineer, but this video is awesome!!

[@jitteryjet7525]

Excellent video, thank for posting. I am still trying to get my head around Lambert's Equation.

[@bubski6981]

This is an awesome video! Mad desmos skills.

[@andyherle]

wow why has this not gotten more attention!! amazing

[@leonmareno4599]

Great one. Thanks for sharing your knowledge.

[@xander2853]

Wow your builds look amazing

[@justananonymousperson7011]

This is really cool!

[@1320crusier]

Very cool. More of this please.

[@Danbearpig]

Very cool, glad I stumbled across this!

[@Ykswad]

You have one of the coolest job on this planet, I wish my country had a space program

[@napoleonbonaparte4382]

Even though I have been quite busy recently to not be able to do anything with ksp I still love finding things like this that could help me in useless ways. Ps. I really love your work and I Really want to see your channel succeed!

[@maxpaulsen6207]

That's impressive! Now do a rendezvous manoeuvre with one active spacecraft using only angles, relative velocities and range :)

[@ollieross6137]

Awesome. thanks for doing the work you do :)

[@jupiterbjy]

If KSP gets a mod where we have to determine every body's orbit like this at the start of the game, it would be awesome

[@anonymous-rb2sr]

This is really intresting! This was always the part of "space related activities" I was wondering about, precisely how you managed to turn "???" into a precise computation that has orbits down to the meter My first thought watching your video was "well ok, but if you are given the distance by the game that's obliviously very easy, you don't even need to triangulate or do complex maths, just aim a gigantic ruler in the right direction and you know the exact position of the spacething in a reference frame"... but then I realized that it's probably something that can actually be done irl, using for example the time to send and return signals with a known travel speed like light, so that answers my question about how it's done irl with such high precision, they basically DO cheat (by they I guess you since it's your job lol), using something that works like lidar to get very precise distance measurements of everything, maybe using several stations and trigonometry more for doublechecking the results rather than the actual main course of information gathering but that raises another question, let's say you only have access to telescopes, and have NO idea how far from said telescope what you are staring at is, let's also say you don't really have any far away station (nothing on other planets or in orbit, and if possible nothing outside of one continent) How then would you go about making a very PRECISE model of the orbit of an object in the solar system? basically put yourself back in the 18th century, with what they had available but modern maths and knowledge, how accurately could you track an asteroid, and how would you go about doing it? (as an added precision let's say you already have a near perfect knowledge of all the orbital motion of the known planetoids)

[@piperboy98]

This was solved, see Gauss' or Laplace's methods of orbit determination. You need three observations though. In effect you have three lines you know the orbit passes though (from the observation locations along the corresponding observation directions), and that the points where the orbit crosses those lines must be coplanar (since a Keplerian orbit is planar). Coplanarity means you need only choose distances along two of the lines and you know the third, and also from those three points you know the full ellipse. In principle then you have two time deltas derivable from your two parameter space of possible orbits so you can solve from there. It's gets rather messy in the details though, and ends with an 8th order polynomial for the distance to the second observation (and that's after some small time delta approximations). Modern methods are usually just iterative solvers instead where you predict the observations from an initial guess orbit and simply figure out how to tweak the orbit parameters to reduce the error against the actual observations and do this repeatedly in a computer until they coincide as well as possible (typical multidimensional gradient descent type stuff).

[@bfournier1884]

This video was a great introduction, congrats ! Maybe now's the time to add the estimation of the error that your method yields, not from the comparison with the true value but statistical methods on top of several measurements, even involving the Principia mod to add some variance ! For example, a follow-up video could use a Gauss OD to demonstrate estimation from angles only and perfect least squares inversion in stock KSP, then the need for optimal estimation and error control once n-body physics gets involved.

[@FL2070]

This channel is an absolute golden find.

[@sudantarescosmonautics9422]

I feel myself dumb after your video, doesn't matter I built the very first hanging base under the munar arch without the anchor (1.9), and from stock parts only. I did it all through the trial of guessings and errors. But man, you opened a brand new horizon in the game. I wish I'd be as half as smart as you. Amazing video! 👏👍

[@iloveemeralds4622]

Time to use this in ksp 2 to figure the orbits of my friends stuff so I can attack them

[@railworksamerica]

I have no damn clue what you’re talking about but by God is this interesting

[@alexanderryan-jones600]

Far out! That's incredible!

[@EvanG529]

You have my dream job

[@teamruddy611]

3:43 - you can actually use the debug menu to set eccentricity to 0, even though there's a law of physics (or some other science i can't remember the name of) against it - but it is very glitchy.

[@Mochu_s_Junkyard]

would love to keep hearing voice-over videos whatever the mission

[@aarong.4691]

The algorithm must have known that just yesterday I was wondering how this was done and recommend this to me today. Was reading about the big radar imaging facility they have in Germany as they take some awesome photos of spacecraft and things and was wondering if they where doing anything for Artemis I, and saw that they determine orbits and was wondering how they managed to do that. This answers that very well! My only question is how you find the spacecraft, I'm guessing you have the initial estimate from physics and then are able to find it using that and then refine to the real world number using these observations. Very cool stuff and great video!

[@keksentdecker]

Insane! This is so cool

[@jamieloganaerospace6404]

Thanks!

[@Twitchi]

As a first year astro student this is fascinating :D

[@renzo00]

this is way too cool

[@revolutionstudios5052]

I like your funny words, magic man.

[@charimuvilla8693]

Always great to see professionals explaining something using ksp, definitely sticking around and subscribing!! I think you will find the k-OS (kerbal operating system) mod very interesting. You can basically program stuff and make things as easy or complicated as you want. I've used it for anything from achieving a cm-perfect orbit to landing a booster on a barge. I also have a question: In real life how do you get distance information from James Webb and how much more accurate does it make things compared to pure angle measurements?

[@bfournier1884]

For spacecrafts further from Earth, passive ranging (for example by bouncing a laser on them) can be quite difficult. Active ranging is used (think, GPS, but reverse, where the ground sends the time and the satellite sends it back), but techniques based on angles (and relative speeds, Doppler ranging is a nice addition !) are very precise already. With POD techniques (complicated math on top of precise sensors aboard satellites to reconstitute the exact trajectory a posteriori), you can go below one meter in a posteriori location error box for earth observation satellites. I wouldn't be surprised the Webb is in that order of magnitude, although maybe for a space telescope the exact position is less crucial for science than an Earth Observation satellite needs to be able to take accurate pictures

[@charimuvilla8693]

@@bfournier1884 Very interesting thanks!

[@mmb3006]

tnx for making this vid!

[@Ramash440]

I've always wondered how real spacecraft know their orbit. In KSP you always have accurate knowledge on the stuff but I can't imagine any amount of INS being able to predict that by itself. I figured that it was up to tracking stations to determine that, I mean it's in the name, but I never quite knew how.

[@Gopnik_B57_2]

absolute legend!

[@OfficialMageHD]

Found your videos the other day, and now (not so) slowly binge watching them all. Is it possible to do a fully automated launch, flight, and moon landing with the kal controllers? I messed around with them a bit quite awhile ago, and didn't really know what I was doing, but it seemed like they have a lot of potential when combined with action groups. I'm curious how far those can take a mission between the amount of control the controllers have, and proper planning for launch times and sequences so everything lines up.

[@jamieloganaerospace6404]

That is a very good question. I’ve thought about how I might go about doing something like a fully automated Mun mission. The major challenge that presents is probably getting the timing of it to work autonomously. Like, when to start up or shut down a landing burn for example. Another option could be finding a way to use the environment to trigger a KAL sequence, perhaps by having a solar panel eclipsed from the sun. But then managing attitude control autonomously would also be another can of worms. I’ll keep thinking on it.

[@MichaelS-pr9qn]

This is awesome, Thank You!

[@Slaintemaith]

You mean the Sally Ride Space Telescope?! Awesome!

[@moocowpong1]

How much harder would this be if you couldn’t measure the distance to the orbiter directly?

[@jamieloganaerospace6404]

It would be a lot harder, but still doable. The OD would have to be performed using angle measurements only, which would require that you either: 1) Perform two pairs of simultaneous Az/El measurements at two different tracking stations, and then combine them to triangulate the target’s position at the observation time or 2) Perform Az/El observations from a single station at several points in time, and then feed them into a more sophisticated OD algorithm (E.g.; Batch Least Squares or Sequential Kalman Filter)

[@russellthorburn9297]

I was going to ask the same question. 🙂

[@Von_DBL]

Well I guess when it comes to precision, the distance increase could make a 0.27% margin quite non optimal but considering the tools, it sounds great ! Now can we imagine a video of a Hohmann transfer using your stock sextant? This would be a great feature and a really awesome way to get out of Kerbin’s Orbit bodies.

[@Hasmo_]

Bro is one of the few qualified to play ksp, would love to work for something as big as JWST some day!

[@Yoctopory]

Itˋs like the characters of the game discovered the scientific method to find out more about the universe they live in. Only to eventually discover that they are living in a computer simulation…

[@Darthsinduk]

Dude you are so smart holy cow dude

[@GrilledCheeseSandwichReal]

This video was awesome! Would you ever consider doing this again with principia? (N-body physics mod)

[@SofieBrink]

I just thought of a way that could potentially make this easier, what if you put a reaction wheel on the actual moving part of the antenna and turned off the motors on the robotic parts to let it free spin. And you used SAS to point straight at the target. Then you just simply take the readings from the motors and do the determination with those values

[@jamieloganaerospace6404]

I tried this very same thing initially, but it turns out the SAS target hold option doesn’t remain pointed at the target with the accuracy needed for this to work. It tends to jump around.

[@SofieBrink]

@@jamieloganaerospace6404 ah alrighty, good to know!

[@MagneticEel]

This taught me better than my teacher did!

[@youtubehandlesux]

Can you do some videos about KSP's Principia mod? It adds n-body simulation to the game and I think that would be interesting for a satellite navigation engineer, it needs a beefy CPU though

[@nikkiofthevalley]

As in, the whole system is simulated, or that every planet's attraction is computed on the craft?

[@youtubehandlesux]

@@nikkiofthevalley Whole system, they have to change stock Jool moons' orbits because they aren't stable

[@taskfroce80th95]

I like your funny words magic man

[@Quad373]

What are your graphics mod? I just wanna know which one you have which removes the skybox so its much more dark

[@notravstar]

The satallite knows where it is. It knows this because it knows where it isn't

[@joshuasims5421]

That was so incredibly cool! We really take for granted that our KSP spacecraft know their own orbits so precisely. I'd love to hear more about your work, is most of this work computerized or do you still have to do a lot of tracking manually?

[@13ratsinatoyota]

this is amazing!

[@nomaxnula1]

Great video!! You're awesome

[@eduardopupucon]

now do it with principia installed! it would be awesome

[@AgentCryo]

This is so cool.

[@rignall05]

I tried something similar but instead used a reaction wheel connected to unmotorised servos on a static base. This made it much easier to track (i used the point to target on SAS) however the servo readings werent accurate.

[@jamieloganaerospace6404]

Yep, I initially tried the same thing and got the same results.

[@Mrlimabean01]

badass

[@robonthecob5092]

Hi Jamie, what degree did you do at university? Great video

[@ktw5819]

You should play RSS/RO, I would like to see your take on a manned mars mission.

[@HieronymousLex]

Wow, that’s awesome

[@NavidIsANoob]

Instantly subbed.

[@1ndragunawan]

What's next? Scientific paper using KSP as the simulator? Use Principia mod please.

[@fernandovalner]

dude, sadly, I understood close to nothing in your video. But I am saving it, and I'll come back to it in a few more years of playing ksp (hopefully ksp2) to see te progress I am making. If you are willing to do more videos teaching space science, you have my views, subscription, and likes.

[@hoteltrivago8186]

What KSP mods do you use when making this?