≡ Menu

Dynamics of an Interstellar Probe

Yesterday’s look at radiation and its effects on humans in space asked whether any Fermi implications were to be found in the work described at the University of Rochester. One answer is that expansion into the cosmos does not need to be biological, for biological beings can build robotic explorers equipped with enough artificial intelligence to get the job done. A truly advanced civilization would be able to create large numbers of intelligent probes or, indeed, self-replicating probes that could spread throughout the galaxy on a timescale of perhaps ten million years.

Fermi speculation is always fun but, when we get into the motivations of extraterrestrial civilizations, it leads inevitably to unfalsifiable solutions, good for conversation over coffee but incapable of producing a scientific result. Thus the ‘zoo hypothesis,’ the notion that the Earth is intentionally left alone to pursue its own development by beings with an agenda of their own. It makes for terrific science fiction, but how much can we say about alien sociology? Working with two colleagues at the University of Edinburgh, Duncan Forgan thinks we’d be better off studying the physical constraints that might govern an expanding wave of galactic exploration.

Image: The spiral galaxy NGC 4414, as imaged by the Hubble Space Telescope. Can simulations help us understand how a galaxy of hundreds of billions of stars can be explored by a single civilization? Credit: NASA.

Forgan argues that even in our early exploration of the Solar System, we’ve learned the value of gravitational ‘slingshot’ maneuvers that let us produce velocity and trajectory changes without the use of fuel. Simple powered flight is an inefficient way to travel, especially in our era of chemical rocketry and, presumably, in the coming era of nuclear propulsion methods. At the galactic level, probes can be accelerated relative to the galactic rest frame by similar techniques. Forgan and team run simulations analyzing the ways a single probe can move through a population of stars.

Specifically, the researchers work with three scenarios: 1) The ‘powered’ approach, in which a single probe simply moves between stars by targeting the nearest unvisited neighbor; 2) The ‘slingshot’ approach, using gravitational slingshot methods to move along the same course as in scenario 1; and 3) The ‘maxspeed’ approach, where a probe selects the next destination star in such a way as to get the biggest benefit from a slingshot maneuver — all this depends on the destination’s star’s velocity relative to the current star being approached. The three simulation scenarios were carried out in a field of 1 million stars at a density of 1 star per cubic parsec.

What Forgan and team are looking for is the effect of these different strategies on relative travel times, as the paper notes:

We select a relatively low maximum velocity of 3 X 10-5c, where c is the speed of light in vacuo. This is comparable to the maximum velocities obtained by unmanned terrestrial probes such as the Voyager probes. Admittedly, the Voyager probes achieved these speeds thanks to slingshot trajectories, so the top speed of human technology under purely powered flight is unclear. To some extent, the maximum powered speed of the probe is less important – increasing or decreasing this maximum will simply affect the absolute values of the resulting travel times in a similar fashion. What is more important is the relative effect of changing the propulsion method and/or the trajectory.

The slingshot trajectory turns out to have a significantly shorter travel time than either the powered or maxspeed approaches, following the same course as the ‘powered’ probe but able to boost its velocity at every stage of the journey. In fact, Forgan’s simulations show the probe will have increased its velocity by close to a factor of 100 in the course of the simulations. The ‘maxspeed’ scenario optimized for velocity but did not consider the distance to the next star, creating lengthy journeys between widely spaced stars in the early stages of the exploration.

Forgan notes that with the current speed limitations of human probes, the galaxy could not be explored with a single probe in the lifetime of the universe, but the use of multiple probes changes the picture entirely:

Given our current ability to manufacture large numbers of similar sized craft for terrestrial uses, it is not unlikely that we can adopt a similar approach to building probes. A simple calculation shows that producing 1011 Voyager-esque probes would allow humankind to explore the Galaxy in 109 years. Given that around 5 X 107 automobiles are produced each year globally, it seems reasonable to expect a coordinated global effort could produce the requisite probes within a few thousand years. If the probes are made to be self-replicating, using materials en route to synthesize copies, the exponential nature of this process cuts down exploration time dramatically.

Whatever the case — a fleet of probes created by an extraterrestrial civilization or an exponentially growing fleet of probes each engaged in self-replication — adding gravitational slingshot dynamics into the mix can shorten travel times and strengthen the Fermi paradox. Slingshot effects reduce a single probe’s travel time by two orders of magnitude. Forgan’s team is now repeating the analysis of this paper using self-replicating probes as the basis of exploring whether slingshot dynamics fit into the overall cost-benefit picture of that kind of expansion.

The paper is Forgan, Papadogiannakis and Kitching, “The Effect of Probe Dynamics on Galactic Exploration Timescales,” accepted for publication in JBIS (preprint).


Comments on this entry are closed.

  • Ed Reed January 3, 2013, 11:13

    Seems like you need to separate, or at least treat differently, the two described approaches: slingshot for ever-increasing velocity vs self replication. I’d think a self replication strategy would require matching speeds with available resources, which may occur by sucking in interstellar materials in route (with drag associated with changing the velocity and direction of material thus corralled by the probe), or by matching velocity with a target star system and NOT using slingshot to further the trip, but rather to decelerate into the system for a period of self replication.

    AI-driven probes using sling-shot techniques to “scout” prospects makes sense, followed by AI-driven probes using sling-shot techniques to approach and orbit selected target systems to foster subsequent in-depth exploration, self replication and AI-colonization also makes sense, over a span of millenia and eons. So, mixed mode strategy.

    I wonder what the criteria will be for such self-replicating AI? This is truly the question of creation. Whether you use a model of parent-child or god-creation, how you decide whether to release an autonomous being to roam all of creation and whether you’ve instilled enough of your own high-minded values to discourage it from turning into the dreaded Saberhagen Berserker is really an amazing question to contemplate.

  • Chris Rose January 3, 2013, 13:11

    So, here’s a tyro question – is it possible to use the objects in a single stellar system (say, our Solar System) to accelerate a spacecraft using a series of (increasingly fast) slingshot maneuvers? If we accept Sol, Jupiter and Saturn as slingshot ‘helpers’, what is the maximum speed that could be attained? What would cause the upper bound on the speed?

  • Tarmen January 3, 2013, 16:05

    Cheap probes are an obvious method of procreating and extending one’s kind. So it seems likely that many of these cheap mites have likely been spitting to and fro through system after system since who knows how long. Like plankton in the Pacific. Maybe a wave of curious fly by probes floods the galaxy every tens of millions of years or so, and they maybe fire off a data packet backward toward their origin, eventually sputtering out one way or another. Ambitious civilization blurs gradually into anonymous space junk. Perhaps there is some old space junk trapped in the Oort Cloud of every star. Ours likely has been targeted repeatedly by snoopers of this kind.

  • Andrew W January 4, 2013, 2:26

    All seems unlikely to me, at these speeds, travel times, even between the closest stars, is tens of thousands of years, and speeding up the probes significantly would mean that the slingshot effect, even with very close passes to stars , would have a negligible effect on speed and direction.

  • Anthony Mugan January 4, 2013, 11:36

    Whilst this point is clearly anthropocentric and therefore who knows if any ETCs might think the same way, I do wonder about the economics of this sort of model. The payback period on such a huge investment is so long that I wonder if humanity would ever embark on mass produced, automated, instellar probe programmes that operate at a small fraction of c…? The risk factors involved in creating self replicating probes have been discussed elsewhere and are not ones I’d personally want to take.

    That said, alternative scenarios such as some breakthrough physics or perhaps more tangibly sheer neccessity if some civilisation is facing a choice of leaving their homeworld or facing eventual extinction could well alter the equation quite dramatically.

    One thing that also puzzles me a bit about these models is that they never factor in the possibility of allowing an intelligent probe to decide to ‘hang around’ and do some more detailed sampling of any interesting flora, fauna or other subjects of interest (e.g. natural resources, to carry on the economics theme) it (if that is the right word for a self replicating intelligent entity) might come accross…

  • JoeP January 4, 2013, 12:13

    I wonder if the biological vs. technological machines dichotomy exists for very long at all to intelligent life elsewhere. It may be there is a relatively quick transition from planetary beings to descendants made of other materials that are have no issues with interstellar travel. Resistance to radiation, perception of vast time between voyages, are all limitations imposed on us due to our structure and limited life span.

    A 10,000 year journey to us seems intolerable. To another intelligent being, it may only be a simple wait time. I can see us making AIs that don’t care about the passage of time quite easily. Forever patient.

  • Rob Henry January 4, 2013, 17:41

    Chris Rose asks, “What would cause the upper bound on the speed?”. And to me the answer is in terms of opportunity and turning angle.

    Our solar system doesn’t have dozens of Jupiter sized planets, just one, and another Saturn sized one. To continually take advantage of the slingshot effect, you need a chain of close passes where each of which BOTH turns the probes trajectory towards the direction of its own orbit, AND sends it on a new trajectory. Our solar system has little potential here, our galaxy lots.

    No matter how fast our probe is moving, a (star) can give it the same velocity boost that it can a slower one from the same direction, as long its speed wrt that star does not exceed the turning angle that that (star) can impart to its trajectory. If it is greater than this, its speed can still be boosted, but not as much. For our sun, surface escape velocity is 300km/s, so, taking it as a typical star, this speed is about where the effect drops off dramatically. In the above article, they calculate cumulative implied boosts up to 900km/s.

  • Astronist January 4, 2013, 18:17

    I don’t understand this post at all, because gravitational slingshot manoeuvres are essentially irrelevant to interstellar flight, unless you have something like a binary neutron star in your backyard.

    In your quote, Forgan says he considers a maximum velocity of 3 x 10^-5 c, which is 9 km/s, which is actually considerably less than the hyperbolic velocity at infinity of the Voyager probes of around 16 or 17 km/s. This, I would suggest, is far too slow for any kind of realistic interstellar probe. But then he seems to be talking about boosting that speed by a factor of 100 during the course of the simulations: how is this achieved? It could be achieved, I suppose, by flybys of a large number of double or multiple stars, or perhaps of stars with gas giant planets, but this seems to be hypothesising a probe that (a) endures for the millions of years necessary for tens to hundreds of such flybys, and (b) is not in fact a probe at all, since it never stops at one target for sustained close-up observations. There seems to be no explanation of what purpose such a vehicle would have, or why one should expect it to continue functioning for such long period of time. Altogether a problem posting, I’m afraid!


  • Rob Henry January 5, 2013, 15:05

    Astronist, you’re spot on about that initial probe velocity being superlatively slow, but I fear that you have made an awful mistake on your second point, by using the incorrect reference frame.

    A single star with no planets can’t boost a probes absolute hyperbolic excess wrt itself, but it can wrt the average star if it moves relative to them . It does.

    Thinking of mistakes, above I should have written as the condition for OPPORTUNITY after the AND “targets a new massive object” not just “sends it on a new trajectory”.

  • Astronist January 5, 2013, 18:16

    Rob Henry, thanks, I do believe you are correct. In fact, more than correct: boosting the probe’s velocity would usually be done at single stars rather than using stellar companions, as those companions would not generally be in the correct geometry to take a probe where it wanted to go next.

    However, we are still left with the apparently bizarre situation where these three scholars are either vastly overstating the ability of gravitational slingshots to accelerate a probe on a single interstellar crossing, or are speculating about the possibility that someone might launch a probe which is not a probe, because it never stops in any planetary system to do any exploration, but merely flies from one star to another at ever-increasing speeds (from an extremely low starting point) until it wears out!

  • Brett Bellmore January 5, 2013, 19:33

    The only reason the slingshots can accomplish such a high amplification of the original velocity, is that the original velocity was chosen to be extremely low. If you used a propulsion system that was already capable of 100 times that speed, (Even a crude Orion drive, IOW.) slingshots would hardly help at all.

  • Rob Henry January 5, 2013, 21:30

    Astronist, I should address your other points further.

    I agree that this scheme makes no sense in a formally coordinated effort at interstellar exploration, but the prerequisites for such travel always seem to be a Sol wide industrial base anyway. Now if in this period each university spends just a tiny amount on an annual new interstellar probe as one of their traditional projects, and coordinate their efforts with each other, this would explain the number and slow initial velocities of probes that end up being built and launched there.

    Furthermore, we currently have little experience building in space, but when we do we should know exactly what would have zero deterioration in the cold changeless vacuum of space, and can withstand a few thousand close passages to stars, and has the ability to dish out a very small delta V course correction at each.

    Sure, you worry that the abilities of each is pathetic, but here they must derive kudos, by their relative superiority over predecessors and rival universities. The appeal is that many students would leap at any chance to be a larger part of the cosmos, and have the fruits of their labours live on way after they were dead. The actual data itself would be almost irrelevant to this drive, since they would always realise that the very institutions themselves might be long gone by the time of the first download.

  • Eniac January 6, 2013, 2:30

    …, or are speculating about the possibility that someone might launch a probe which is not a probe, because it never stops in any planetary system to do any exploration, but merely flies from one star to another at ever-increasing speeds (from an extremely low starting point) until it wears out!

    I think that is exactly what they are speculating. Also, because of the travel times and number of encounters needed to get at the higher end speeds (900 km/s?) it would take tens or hundreds of millions of years to top out the speed. Way out there, but maybe worth a little speculation.

    You are right, though, that it is hard to imagine a “probe” that lasts long enough to even report on its first few encounters, much less those millions of years later. Or, that it would be worthwhile waiting for a few days of observation in between 100,000 year long dead periods

  • Eniac January 6, 2013, 15:02

    Another way to look at this issue is to view the collection of stars and craft to be in thermodynamic equilibrium. Since they interact via gravitational encounters, when you wait long enough they should tend each to have the same average kinetic energy. Then the average velocity of the craft would be the average velocity of stars (~10-100 km/s?) multiplied by the square root of the mass ratio between stars and craft. This comes out to (of course) astronomical numbers, or, taking relativity into account, large gamma factors.

    The rub, I think, is that the “long enough” mentioned above may be trillions of years. Plus, the craft will exit the galaxy (i.e. “evaporate”) long before they can even start to get close to c. Unless you use Lorentz forces, that is, to keep them inside. Lorentz force may also be a better way to interact with the stars than gravity.

    I wonder sometimes if that is the origin of galactic cosmic rays. Rather than being generated by some mysterious source, ions may just be circling the galaxy, contained by the galactic magnetic field and stochastically accelerated by the constant jostling of stars.

  • Rob Henry January 6, 2013, 19:53

    Hey Eniac, these probes are not zombies. Your thermal equilibrium calculations are an interesting counterpoint to the slow speeds mentioned here, but I see the “leaving the galaxy velocity” through the eyes of a clever probe. Our distance from the galactic centre is about 8000 parsecs, and star separation here is about 1 parsec. So each deflection need be 1.3*10^-4 radians.

    Okay, things get a bit complicated thereafter, but I calculate that if the average encounter is like an approach to our sun of a million km from its centre, then the answer to the maximum retention velocity is 46,000 km/s. So the theoretical limit for the extremely patient of the galactic outskirts would be about 0.1c.

  • Brett Bellmore January 6, 2013, 22:41

    That requires interactions capable of producing such speeds, which gravitational interactions with stars aren’t. Charged particles are essentially points, you can reach arbitrarily strong interactions by sufficiently close approaches. (Until you reach the sort of energies where you end up with different particles coming out.) The strength of interaction via gravity with a star is limited by the fact that you actually *hit* the star if you get too close to it’s center of mass. So the slingshot effect really loses most of it’s “umph” once you approach the escape velocity of the star.

  • Rob Henry January 7, 2013, 18:36

    Brett Bellmore, your above explanation sounded wonderfully simple and logical. I believe that, other than your point about having to take into account what we might call the capture cross-section of a star per interaction your assertions are all based on incorrect assumptions.

    Let me cut to the basics. Those molecular interactions really are arbitrarily strong, but as far as eventual thermalisation is concerned, that is a red herring. I was going to give you a long explanation as to why, when I realised that all I need say is that you just can’t beat thermodynamics.

    Note though that you can use your argument to say that it would take longer to thermalise though, so that cross-section becomes important. As Eniac pointed out, (in an infinitely large galaxy) this would give a probe mass object a high gamma, so these hyperbolic interactions have linear trajectories, thus calculating that cross-sectional capture area becomes easier. Now if we define an interaction as coming within 0.1 parsec of a star, its capture probability of 1.5*10^-15. So, it seems, that we can only allow about a quadrillion such interactions along the path to thermalisation.

  • Eniac January 7, 2013, 23:04

    Don’t forget that it takes only a very moderate and easily achievable charge/mass ratio to circularize the motion of a probe with respect to the galaxy, even at highly relativistic velocities. Similarly, Lorentz forces near a star can greatly exceed gravitational interactions, so I think that this mode of transportation, while still only for the VERY patient, becomes a lot more interesting with charged craft.

  • Chris Rose January 8, 2013, 13:22

    Eniac, do you have a reference somewhere that would be helpful in understanding and calculating Lorentz interactions? I’m trying not to get too excited about this – but it sounds awfully cool.

  • Eniac January 9, 2013, 0:21

    Chris Rose: I do not really have a reference handy. I did back-of-the-envelope (literally) calculations a while ago, and if I remember correctly a curvature radius of a few light years at relativistic velocities is not too much to ask in the galactic magnetic field. The trick is to generate a sufficiently large charge/mass ratio, best accomplished by charging up a web of ultra-thin carbon fibers (nanotubes) until they nearly tear themselves apart from electrostatic repulsion.

    This could also be used to accelerate to perhaps a few thousand km/s around Jupiter, using its strong magnetic dipole field. The solar magnetic field is less suitable, as it is apparently quite chaotic with all that solar wind tangling it all up and blowing it to pieces.

    I do think that this has been written about, few and far between, so a thorough Google search should bring up something useful. Start with “Lorentz turning”. Maybe Paul knows a few relevant sources, also.

  • Eniac January 9, 2013, 0:41

    Ah, turns out that some of my back-of-the-envelope calculations were actually type-into-blog-comment instead. You can find them here: https://www.centauri-dreams.org/?p=22174. For what it is worth…

  • Thomas Hackney January 20, 2013, 7:15

    The utility of A.I probes decreases the further out they go, since the probe’s report time increases accordingly. The important thing these scouting craft will provide is information about space before manned craft are sent. The importance of sending and establishing human colonies off the earth is elementary and obvious. They are not only stepping stones they are insurance policies for longitudinal human survival and quality of life on Earth. These colonies would not only be able do science that is impossible to do on Earth but they would ease population pressure and resource depletion for the home planet. Both space and basic resources are infinite out there. Should a catastrophic or extinction-level event occur on Earth (due to a pole shift, nuclear world war, comet strike, “super”-volcano eruption, or what have you), the colonies would be unaffected, and could help pick up the pieces on Earth. Such colonization would be hard and expensive, but this is what separates men from the boys, biogalactically speaking. Those civilizations that make it deserve to make it; those that fail will, well, fail. So are we men or boys? (just wondering.)

  • Eniac January 20, 2013, 21:15

    While it is true that the survivability of the human race can be improved only by “true” colonization, i.e. independent colonies with actual humans included, for the purpose of the Fermi question machines work just as well, since we can ask the modified question “Where are they? Or their machines?”.

    I doubt that even in the absence of true colonization there is much of a chance the human race will be rendered extinct anytime soon. None of the ones you mention (pole shift, nuclear world war, comet strike, “super”-volcano eruption) strikes me as being radical enough to eradicate humanity to the extent that not a single survivable group would remain anywhere on Earth. Our civilization and technology may be more easily destroyed, but those can and will be rebuilt eventually by a surviving humanity.