• Content count

  • Joined

  • Last visited

Community Reputation

57 Excellent

About Tullius

  • Rank
    Rocketry Enthusiast
  1. Exactly that, if you have two sides, resp. the ratio of their lengths, you can get the angle with arctan, in the same way as you can get the ratio of these sides by taking tan of the angle. arctan is the inverse function of tan, i.e. arctan = tan^-1. What does inverse function mean? If I have a function f mapping x to y, then f^-1 maps y to x (provided that no more than one x is mapped to y). What does this mean for tan? tan maps angles in (-pi/2,pi/2) resp. (-90°,90°) (depending on your notation of angles) to values in the real numbers, and arctan maps real numbers to angles in (-pi/2,pi/2) resp. (-90°,90°). For example tan(pi/4) = tan(45°) = 1, and arctan(1) = pi/4 = 45°. Or take a look at this graph tan maps values on the horizontal axis (usually called the x-axis) to values on the vertical axis (usually called the y-axis), arctan does the inverse: it maps values on the vertical axis to values on the horizontal axis.
  2. It would mean starting from zero, or accepting Falcon Heavy and New Glenn as alternatives. And while the former isn't really interesting for the senate to fund, since they would see the results even later, the latter two are less capable rockets than SLS, which, while one would quickly see some results, would have difficulties even putting a space station module into lunar orbit. And if you start anew with private contractors, every development works needs to be paid for 2 or 3 times, since you want competition. Also, either NASA lets 2 or 3 companies build the rocket entirely on their own, which would mean that NASA has little influence on the design and you have to find as many companies willing to do it, or NASA does the development itself, but buys different parts from different companies, which would reduce the efficiency gains. In short, cancelling SLS now would not just mean a huge blow to NASA's manned space exploration, but also leave NASA with something that might not necessarily be better. It might be necessary to redesign some parts of SLS, like maybe find a private contractor to develop new, cheaper engines to be build in NASA's facilities, but just cancelling SLS and putting the whole work in the hands of private companies won't necessarily improve anything. On the other hand, if NASA finds out in 3-5 years that SLS is indeed a dead end, there is a good chance that SpaceX and Blue Origin might have developed by that point an acceptable replacement, which NASA could buy directly without any of the lengthy procurement procedures, since there are only 1 or 2 viable possibilities.
  3. sin and tan are the trigonometric functions sine and cosine (https://en.wikipedia.org/wiki/Trigonometric_functions). arctan is the inverse function of tan (https://en.wikipedia.org/wiki/Inverse_trigonometric_functions).
  4. I think that SpaceX just wants for the time being save money on the development of a specialised 2nd stage for Falcon Heavy. The Falcon 9 2nd stage is good enough for the first flights of Falcon Heavy and provides it with a meaningful payload capacity, even if it means that recovering the core stage will cost a lot more fuel and thereby cost a lot in terms of payload fraction. And lets not forget that SpaceX wants to recover 2nd stages, so they are planning for quite some development in terms of Falcon 2nd stages, unlike the first stages, where most of the development already seems to have happened. After all, Falcon wouldn't be the first rocket that got a new second stage at some stage in its life cycle.
  5. I think the "if something where to happen during ingress" is relating much less to problems that astronauts might have to get into the capsule, but to a mishap with the rocket happening during that procedure, which would force the astronauts (who cannot yet use the LES, since they are not yet secured in their seats with the hatch closed) and the support crew to get away from the rocket as fast as possible, i.e. to use the zipline. This is relevant, because in case of the Shuttle, the external tank would already be fully fueled before the astronauts and their support crew would access the tower. And it is quite likely that it will be the same for SLS. In short, the zipline is the rescue plan for those people that will access the tower with a fully fueled rocket sitting next to them on the pad.
  6. The AI doesn't grasp anything, it just notices that certain combinations of actions lead to certain scores. If you look at the console output in the stream, you see after each try the line "fitness: ...", which is the score the AI received in the previous try (the higher, the better). And at the moment the highest achieved score is 6,9... I am still unsure what the "adj fit" in the table means; maybe it is something like the average score achieved, and thereby showing the progress. But I am really not sure; I just see it being higher, if I check back after a few hours. So the AI might notice that certain directions of thrust lead to good scores, but it won't notice the behaviour of the Pe independently of the rest. But most of its tries ending up in quick turns might suggest that it noticed that thrusting horizontally leads to good scores, since it increases the periapsis. And "will take some time", it something like the motto of neural networks. It took AlphaGo a couple of months to increase from barely the level of good European Go player to beating one of the best human players. And that despite running on a supercomputer. In the experiment above, we are not even at 6000 tries. But in the end, we all needed quite a few tries before achieving our first orbit in KSP. And we knew what an orbit is, and maybe also read up on how to achieve orbit in KSP. And now imagine someone, who starts literally from 0, knowing absolutely nothing about the task ahead.
  7. The score that the AI gets as a result after each try is also based on periapsis and apoapsis height. Or put differently, the scoring system knows what an orbit is and gives the AI a score based on how good its achieved "orbit" was. But sure, the scoring is much more difficult in the case of KSP than those of Go or Super Mario. For Go, it could be binary (win = good, loss = bad), or for Super Mario, one would probably base the score on the distance travelled before it died (since after all, winning means travelling the whole distance of a level. Or, to compare again with Super Mario, while the AI doesn't know that, if it falls into the holes in the ground, it will die, it will notice after a couple of attempts that jumping over them allows it to travel much further than falling into them. And since travelling further is good, it will start jumping over the holes.
  8. The whole idea is that the AI starts from nothing, and doesn't even know what it is doing. It can play around with certain commands, gets a few bits of information (such as the navball, the altitude, etc.), but it doesn't know what those commands and pieces mean or what is an orbit or gravity turn, as it should learn that all by itself. The only way the AI gets information about how well it is doing is after each attempt, when it gets a score. So the AI starts with a completely random behaviour and gets a feedback in the form of a score. If the score is good, it will be more likely to attempt something similar in the future, if the score is bad, it will be less likely. For KSP, this is not the best idea, since there are well established concepts of going to orbit, but it is a school project about applying neural nets to playing KSP. So it is no problem, if it is completely pointless. There have been others before that let a similar AI (based on neural nets) learn how to play Super Mario, or Google that developed AlphaGo to beat even the best human Go players. Those two examples need a bit more "creativity" by the AIs, since there are no well-established solutions for their respective problems, unlike KSP, where people have already written kOS scripts to fly rockets to orbit. And, besides the above, why should the AI fly east? You can achieve orbit in any direction, it is just a bit easier, if you go east. But the AI probably doesn't know that yet, and depending on the scoring, it might never learn it (if it is just about getting into orbit).
  9. Noticing that you are from Switzerland, I just checked on the website of the library of the ETH Zürich. While it is on loan from the ETH (and the EPFL), it is still available from one university in the NEBIS (https://www.nebis.ch/en/frontpage/). So if you are at a swiss university with access to the NEBIS catalogue, you might request it from there.
  10. Maybe some overpressure valve locking up or a fuel leak. But I think it is probably more about having an escape plan "just in case", than any specific danger. After all, needing 2 minutes to get away from the rocket is better than 10 or more minutes (just think of the distance they have to cover from the top of the tower until they are safe).
  11. No, 15 kg. It is a bit difficult to find informations, but Germany's similar ATF Dingo protected its crew, when it drove over an anti-tank mine with 6 kg of explosives. The explosion created a crater with a diameter of 2 metres and hurled the Dingo 2 metres sideways, but only 2 crew members got lightly injured.
  12. First, I am a mathematician that looked up on wikipedia what Navier-Stokes and the question are about and not a physicist, but I think that is not a bad thing for answering the above question. The Navier-Stokes equation, like the heat equation or most differential equations, can be solved numerically, i.e. approximately, and that to a rather high precision (given sufficiently large computing power). And that is also the most we can ask for, as explicit solutions generally can only be found for special cases. So everything is good and your question is pointless? Well, it is not so easy. If you want to approximate a solution numerically, it needs to exist and be sufficiently nice. And usually it is also very nice to have a unique solution. This is were the mathematics come into play. For the heat equation, under minimal restrictions on your domain (the shape of your object) and the initial values (the initial distribution of the heat in your object), you can mathematically prove the existence and uniqueness of a bounded solution (boundedness is good, since it prevents the temperature from rising to infinity). So while mathematics is in general incapable of providing us with an explicit solution, it ensures us that the equation actually makes sense for physics and more importantly that our numerical approach to find a solution is sane (approximating a solution, which does not exist or is not unique, doesn't make sense). For Navier-Stokes, we don't know under which conditions a solution exists: So trying to find an approximate solution might, under some circumstances, make no sense, since there is no exact solution to approximate. And to make matters worse, we don't know when this happens. So everything you do with Navier-Stokes on the beaviour of fluids is based on the wild guess that, in the case you are looking at, a sufficiently nice solution exists. Since the Navier-Stokes approach seems to work nicely in practice, people have come up with the hypothesis that there exist minimal assumptions for the existence of a sufficiently nice solution. If you prove that hypothesis, you would earn yourself a nice prize (the problem is one of the Millenium Problems) and would make the physicists and engineers happy, because you just told them that the approach they were taking all these years in approximating the behaviour fluids actually makes sense.
  13. Sure, it is ridiculously expensive, but if you bring everything to Mars with you, you know the stuff that brings you back. If you for example rely on ISRU for the fuel for the trip home, you have to be pretty damn sure that it works, while bringing the fuel involves much less risks. It is a station in lunar orbit and thereby can be used as a simulation for the transfer. We neither put humans into space outside of LEO for extended periods of time, nor have we tested any of the technology for the transfer flight. Lunar orbit allows for a return within a few days if something happens, while a problem half way to Mars can't be resolved by returning earlier. That is why NASA is taking this baby steps. They are slowly developing as much technology for a flight to Mars as their budget permits. But if no administration in 2030s decides to push for Mars at all costs, all would have been for nothing. But if an administration decides to do it, they already have some of the technology. On the other hand, were do you get a launch for 10 million dollars? Even SpaceX charges 60 million dollars for Falcon 9, maybe 20-30 million dollars if they get reusability working as they like. And that is without the cost of payload, which usually costs more than the rocket itself, especially if it needs to be newly developed for Mars. As are most technologies NASA plans to use. The only problem is that they were never used for that purpose and in that configuration. So you need extensive testing of all relevant components before you send a crew to Mars with no easy abort plan. Also Mars Direct involves a Saturn V sized rocket. And what is SLS? A Saturn V sized rocket.
  14. They were a Lunar XPrize team until a few days ago, when they finally gave up, since they would only be able to launch their mission in 2018, and the deadline for the XPrize is end of 2017. Source: http://www.spiegel.de/wissenschaft/weltall/mond-mission-part-time-scientists-gewinnen-vodafone-als-partner-a-1139371.html (in German)
  15. You also have that risk, if the ship is standing still. And New Glenn might be able to cope better with it, since it will be programmed that the landing spot will be constantly updated during flight, unlike Falcon 9, where the landing spot is probably somewhat hardcoded before liftoff.