Would there be a way to put it in orbit around Alpha Centauri? Or would it just be a fly-by?
Given the distance, we couldn’t send signals in time to put it in orbit. Would it be programmed extensively with flight algorithms that could brake its flight and put it in orbit?
And how could it brake its flight? Would there be enough fuel left for a braking manoeuvre?
It’s possible to use multiple massive bodies to change orbits with very little thrust - the trick is that it can take a lot of time Interplanetary Transport Network - Wikipedia - so it might be possible to use the Lagrange points in the binary star system to get into an orbit around one of the stars in Alpha Centauri
Would that somehow be pre-programmed so the probe can do it?
Everything is pre-programmed, since it is too far to control your probe remotely. That does not stop the program from being very sophisticated, and indeed in this case it probably must be. You can imagine the probe doing things like choosing on its own the precise moment to perform certain manoeuvres.
Considering any information takes 4 years to get to us and our instructions in reaction take 4 years to get back - it better be running on autopilot.
Whether it goes into orbit or does a flyby - well, pick your propulsion method. Let’s say we can get to 5%C so - ignoring minor relativistic effects - a one-way trip would take 80 years. Will it include enough fuel to slow down? Or maybe you use a solar sail, hundreds of square kilometers of extremely thin mylar relying on reflected light to slow down. Or Bussard RAMjet, where the interstellar hydrogen is sucked into a fusion drive so no fuel necessary. There’s a whole thread possible on discussing propulsion methods; but light is 186,000 mi/sec so even getting to 10,000 miles/sec will be an accomplishment, let alone slowing down at the other end.
The fastest human vehicle to date, Voyager 1, would take 72 thousand years to get to Alpha Centauri, and would not be able to stop there at all (I think, unless we could somehow point it with enough precision so it gets captured?)
That’s like saying that crossing the Atlantic is impossible in the 1400s, and pointing to some floatsem left over from a galley in the Mediterranean that made its way to the Americas after a few decades of drifting randomly through the ocean. Voyager would take 72,000 years to reach another star because it wasn’t built to go to another star; it was built to explore our outer solar system and just so happened to have enough energy left over after passing by Titan to escape the sun’s gravity well. Comparing that to a probe that’s actually designed to leave the solar system as its main objective is, frankly, ludicrous.
The OP’s question about controlling the probe once it gets to a remote stellar system is one I’ve been wondering about too. Let’s say tomorrow a magic space drive is invented. It can accelerate up to 0.25c and decelerate at the other end. We still haven’t the tech to build reliable life support systems that can last for decades, so for now we will concentrate on building and programming our interstellar probes to explore remote star systems for us, light-years away. Is modern AI up to the task of doing it for us? Could a probe be programmed to identify planets, navigate to appropriate orbits, and do a complete survey of an unknown system without human intervention?
You can’t control it in any practical manner. It would take 4 years to receive an image and 4 years to send a command based on that image. That’s 8 years. I
Oh I was’nt trying to imply that it was the faster we could get there, just posting the data about voyager because I thought it was interesting.
That’s why I said "the fater human vehicle to date.
Nonetheles I understand how could it be taken that way.
Building an interstellar probe is difficult. Building an interstellar probe that can stop at the destination is much more difficult than that.
Let’s say we had a miracle fuel, so efficient that it only takes 100 tons of it to accelerate a 0.1-ton probe to 0.2c. Decelerating from 0.2c takes as much energy as accelerating to 0.2c. Which means if you want this probe to stop at the destination, it needs to be carrying 100 tons of fuel while flying at 0.2c. So you have to accelerate 100 tons of fuel + 0.1 ton of probe to 0.2 c. Which would take 100,000 tons of fuel.
Gravity assist can help a little bit, but the faster the probe, the less effective.
All this means the first interstellar probes will almost certainly be flybys. Presumably, by that time we’d have telescopes that can actually image exoplanets, so before the probe gets there, we’d know which planets we want the probe to observe, and where they are. The probe would be programmed to all the observations, store all data, then spend the next few years transmitting that data back to earth. Which is how we operate interplanetary probes. (New Horizons made a fast flyby of Pluto and spent the next few months transmitting that data.)
Yes, that’s what I said in the OP.
So, is our AI currently sufficiently advanced to design pre-programmed algorithms that would allow choice of targets, braking manoeuvres, and so on?
Not from an artificial intelligence approach - no.
However, there would not really need to be a requirement for an AI approach - we are perfectly capable of building a complex decision tree approach that could handle a fly-by/fly-through and target items of interest.
And the entire process wouldn’t be hands off. Assuming an 80 year trip to the target, that is at least 60 or so years of observations getting closer and closer to the target - the closer you get, the more the probe can observe (logically, the probe would be accompanied by multiple external telescope probes to allow wide baseline interferometry). You couldn’t control the probe directly, but you would have a very good idea of how to direct it at year 70 or so (with enough time to send the final instructions and get a response).
NASA already is doing and has done this sort of thing with probes round Jupiter/Saturn/Uranus - albeit with hours of delay, not years, but the concepts are the same.
Given self-driving cars can more or less successfully navigate crowded city streets, I would think that identifying a planet orbiting a star and steering relatively close to it for a flyby would not be too difficult. The algorithm would start taking pictures of the star system, look for lights that move relative to the background stars, pick one light and determine where it will be when the probes arrives, and make the necessary course correction.
Doing a gravity assist braking maneuver would be quite a bit harder
We’ve already flown spacecraft with autonomous navigation and control, like the Deep Space 1 probe. Mostly you are asking the spacecraft to determine its position by measuring the position of stars & asteroids, make appropriate course corrections, and point the science instruments at a target that meets certain criteria. It’s much easier than, say, navigating a car through traffic.
And for Mars landers also. While the basic technique is preprogrammed, it is definitely adaptive to take current conditions into account. For control the lag to Mars is about as bad as the lag to Alpha Centauri. You find out if it worked much faster though.
It may help to realize that “what it finds of interest” is going to a planet, and how it responds to something of interest will be “fly close to it”.
How would it “pick one light”? I guess it would depend on what the original programmers thought was the most interesting target.
Suppose they’rE looking for the most “Earth-like”. How would a probe coming into the Solar system be able to distinguish between Earth and Venus? And what if the orbital situation is that some planets would be out of reach because they’re on the opposite side of Alpha Centauri when the probe arrives? Sound like it could be quite an elaborate décisif -making tree.
Good point - hadn’t thought of that.
Most likely, we won’t even launch a probe unless we know there’s an interesting planet there. Otherwise, how would we decide which star to send the probe to? There is already a lot of research into developing a telescope that can directly see exoplanets, and it’s just about possible with current technology + several billion dollars in funding. OK, they say it’d cost a few billion, so it’ll probably be tens of billions of dollars, but still, we know how we would do it. Which is more than can be said about interstellar probes.