Starship sensors

[Tom Mazanec]

We have all seen/heard/read it...an FTL starship reaches a system. The captain asks for a report from the science officer. In two seconds he says "Ten planets...six rockballs, a gas giant, an iceball and two ice giants."
Is this completely bogus? Is there any sensor technology on the horizon which would give such results in a moment? In a day? A month?
Say the starship is the size of a skyscraper.

[danscope]

Spectral analysis yields much which the eye would not see, but it surely takes time. If one is on a starship , surely one has time to spare.
On TV, they are in a hurry.

" Good health , good friends , good food , and time to enjoy them . "

Best regards,
Dan

[ravens_cry]

Well, potentially, an FTL starship has FTL sensors and FTL probes. Sometimes it is simply reading data from a database from already charted, at least basically, systems.

[cjameshuff]

We have all seen/heard/read it...an FTL starship reaches a system. The captain asks for a report from the science officer. In two seconds he says "Ten planets...six rockballs, a gas giant, an iceball and two ice giants."
Is this completely bogus? Is there any sensor technology on the horizon which would give such results in a moment? In a day? A month?
Say the starship is the size of a skyscraper.

It's really just a matter of scaling up...number and size of instruments. A skyscraper-sized starship could carry hundreds of telescopes and other instruments of a wide variety of sizes and capabilities. Planets aren't difficult to tell apart from stars, once you have an instrument pointed at them that can detect them.

[vasiln]

The really cheeky answer is, Of course! Just send a ship to make two FTL trips such that it travels backwards through time with respect to you. Then it can scout the planets in detail and add them to the database before you access it.

The less cheeky answer is, kind of. The difficulty isn't in the paucity of information, the difficulty is in the amount of time to analyze that information. That's probably not an insurmountable barrier.

The determination of the number of planets strikes me as the most difficult of these analyses-- doesn't that seem strange? But if one planet is obscured by another, or by the sun, it's very hard to see it. You might occasionally see, "Ten planets... Wait, no, eleven," as the ship moved around the solar system, granting better views. That's assuming a VERY mysterious system-- we're already charting distant planets, and by the time we have FTL, who knows what systems we have mapped, but not fully explored?

Most of the information we get about planets comes from very few factors: their gravitational influence on other bodies, and the light coming off of them. It's not impossible to imagine getting all of this information in a matter of seconds, but it does suggest ridiculous sensor precision, and ridiculously fast computers to run simulations and pop out useful data. There are hard physical laws that limit precision and computing speed, but I can't say where those laws top off. Besides, you've already ignored one hard law in your scenario, by allowing FTL, why not another?

With gravity and light, you could make some strong guesses about mass, size, temperature. Light data is spectrum data, which gives you good info about composition. You can look at the shadow cast in the solar wind, at least for objects in between you and the sun, but given sufficient precision about objects elsewhere as well, and make a good guess about magnetic fields. You can guess at the age of the system by looking at the sun (but chances are that if you traveled to it, you already know a LOT about the star).

Precision and computing aren't the only factors-- the strength of modeling systems are important as well. Some of the stuff you can guess at means running a lot of simulations, and those simulations aren't any better than the models that drive them. We can expect that a civilization with FTL capability has visited a lot of stars and verified a lot of observations, and that makes their models much stronger than our own current models. They can have theories like our old Bode's law, but verified and refined with lots of samples. In fact, there's probably a lot more that they can say than these rough classifications of planets.

With some of the more detailed things they can say, they'll probably just be indicating a likelihood rather than a certainty. They might have huge databases that specify mean details, with distributions and standard deviations, based on things that can be observed quickly. So they might say things like, "Compatible with models of anaerobic life," or "Significant planetary uranium reserves expected."

But this level of speed and precision of sensors and computers, and strength of models, suggests things that aren't usually a part of those television shows. If starships can do that, what do you imagine marketing firms have discovered about human behavior? Do you think which program you're going to have the holodeck run is really a mystery to anybody that cares to dedicate resources to making a prediction? In other words-- it (and not just FTL) suggest technology that radically changes human experience-- that changes what it even means to be a human.

[CJSF]

Why would the spectral analysis take a long time? Once the light hits the sensor array, characterization of the spectra should be done in seconds, I would think. A reading of the total EM spectra for the objects should quickly reveal atmospheric or other elemental composition, temperature and other interactions with the system's star. With a starship the size of a large building, even imaging the planets to get a quick measure of size and distance shouldn't take long at all, in my opinion.

CJSF

[IsaacKuo]

We have all seen/heard/read it...an FTL starship reaches a system. The captain asks for a report from the science officer. In two seconds he says "Ten planets...six rockballs, a gas giant, an iceball and two ice giants."
Is this completely bogus? Is there any sensor technology on the horizon which would give such results in a moment? In a day? A month?
Say the starship is the size of a skyscraper.

No, it's not bogus. Even human eyes could find the planets within minutes. Telescopic sensors could find and roughly characterize the planets within in seconds.

Here's the critical thing--the science officer merely reports the planets which have been detected. He obviously doesn't report on any planets where weren't detected. If there was by some coincidence a planet which was hidden behind the star? Oh well, not part of the report. If there were some far away planets which were too dimly lit by the star to be immediately obvious? Oh well, not part of the report.

Such a starship entering our system would likely have the science officer report the existence of 7 to 9 planets, depending on whether or not the sensors found Neptune and/or Ceres. The visible eye planets? A piece of cake. Pluto and other small bodies? Probably not detected right away.

[PraedSt]

We have all seen/heard/read it...an FTL starship reaches a system. The captain asks for a report from the science officer. In two seconds he says "Ten planets...six rockballs, a gas giant, an iceball and two ice giants."
Is this completely bogus? Is there any sensor technology on the horizon which would give such results in a moment? In a day? A month?
Say the starship is the size of a skyscraper.

I think you simply need large sensors, and fast actuators and computers. If you're not on the ecliptic plane, a couple of snapshots a few seconds apart should be enough to count most the significant objects that are orbiting the star. Another instrument with high magnification would give you size, and other instruments, like a spectrometer, would give you information about composition.

[IsaacKuo]

In case it isn't obvious--the naked eye planets aren't just some obscure dots which would only be obvious after observing them moving. They are bright compared to all but the brightest stars. It doesn't take very sophisticated sensors to spot these right away.

[cjameshuff]

The determination of the number of planets strikes me as the most difficult of these analyses-- doesn't that seem strange? But if one planet is obscured by another, or by the sun, it's very hard to see it. You might occasionally see, "Ten planets... Wait, no, eleven," as the ship moved around the solar system, granting better views. That's assuming a VERY mysterious system-- we're already charting distant planets, and by the time we have FTL, who knows what systems we have mapped, but not fully explored?

The latter almost never happens, while the former is nearly impossible, and this is for viewing locations within the plane of the solar system, while starships could come in from any direction. Odds are overwhelmingly in favor of all planets being in plain view.

Most of the information we get about planets comes from very few factors: their gravitational influence on other bodies, and the light coming off of them. It's not impossible to imagine getting all of this information in a matter of seconds, but it does suggest ridiculous sensor precision, and ridiculously fast computers to run simulations and pop out useful data. There are hard physical laws that limit precision and computing speed, but I can't say where those laws top off. Besides, you've already ignored one hard law in your scenario, by allowing FTL, why not another?

It'd take a trivial amount of processing to determine if an object is a star or planet from its spectral signature, ridiculously fast computers are unnecessary. It's simple enough it could even be embedded in the sensors themselves. Classification as terrestrial or gas giant worlds is not much more complicated. It also does not suggest ridiculous sensor precision, you should be able to do the job with rough readings from a handful of spectral bands.

[PraedSt]

Of course the physical report is often followed by biology report. "Life-form readings indicate the third planet from the star is inhabited. Population 3 billion, technology level X". Now that's pushing it.

[cjameshuff]

In case it isn't obvious--the naked eye planets aren't just some obscure dots which would only be obvious after observing them moving. They are bright compared to all but the brightest stars. It doesn't take very sophisticated sensors to spot these right away.

Another system might have brighter nearby stars, though.

Still, with measurements of the relative intensities of visible/near infrared and far infrared, they should be quite obvious with a single observation (not detecting them by parallax or motion), even with instruments that can't resolve them as objects. A starship the size of a skyscraper could easily have a separate high resolution telescope investigating each planet in the system almost immediately after arrival.

[PraedSt]

In case it isn't obvious--the naked eye planets aren't just some obscure dots which would only be obvious after observing them moving. They are bright compared to all but the brightest stars. It doesn't take very sophisticated sensors to spot these right away.

Relying on movement certainly isn't necessary; but it is easy. Brightness is a good discriminator, but on occasion the initial viewing angle might be wrong. Having said that, I suppose you could get round this with the use of a good database.

[vasiln]

It'd take a trivial amount of processing to determine if an object is a star or planet from its spectral signature, ridiculously fast computers are unnecessary. It's simple enough it could even be embedded in the sensors themselves. Classification as terrestrial or gas giant worlds is not much more complicated. It also does not suggest ridiculous sensor precision, you should be able to do the job with rough readings from a handful of spectral bands.

You're right that occlusion is improbable (not impossible) but the probability changes with the scale that you look at. What are the odds of Jupiter occluding one of its moons from any particular position in its ecliptic? Probability increases, and Jupiter's moons are probably much more interesting to the Enterprise than Pluto is, than Mercury is.

Identification of "gas giant" status is trivial. Look at a picture, look at the distance, you're likely to be able to tell that it's a gas giant, at least after you've seen a few thousand of them. I wasn't limiting analysis to "rocky planet" vs "gas giant." That's, it's true, pretty simple.

Any kind of analysis by gravity takes a lot of simulation power. Maybe if you have quantum computers, it wouldn't take a lot of computation. Maybe. Writing the algorithm would be hard.

Noticing position (on the projection) is simple; trying to grab velocity isn't easy, and it increases in difficulty as you try to figure it out in increasingly small periods of time, and you need it before you can do any meaningful analysis. Describing the ellipse of its orbit with data grabbed over fractions of seconds from a ship much smaller than earth is hard. It requires extraordinary precision. To begin any kind of meaningful description, you need distance from sun or mass of object. Determining distance via parallax is hard without the kind of parallax possible from a planetary body. Determining mass from gravity is harder. (Determining the distance of stars requires more parallax than we can get on earth-- we need to use our whole orbit). I haven't done the math on how hard these hard things are; I just know that they're hard. If anybody wants to do the math, it'd be lovely

I'm not a spectrographer; reading WP, you'd think that it can be done at the speed of light, but in reality, analysis currently requires times larger than seconds by many orders of magnitude. Days, I understand. That's with the effectively limitless resources we have on Earth. I can't comment on the feasibility of speeding this process up.

Many values you're looking at are probabilistic over any time period. Looking for short periods of times necessarily means less certainty of mean values. That's unavoidable, but probably a relatively minor factor compared to the precisions required. EDIT: On reading, this sounds like I'm being a little silly, because of course planetary position is not probabilistic over any meaningful time or volume. But temperature is, and more importantly, the temperature of your mirrors are, and hence the shape of your mirrors. I don't think you have to add too many orders of magnitude of precision before this starts becoming a real concern, but again, I haven't done the math.

Like I said, doing complicated, passive analysis isn't hard; what's hard is doing it fast. If you try and do it a million times faster than we currently do it, you need instruments a million times more precise. EDIT: Which on Earth would be a million times larger. If you try and do it a million times faster than we currently do it, with a tiny fraction of the mass of sensors that we have on Earth....

[neilzero]

Approaching our inner solar system and half way between Uranus and Neptune when they are about as close to each other as they get, the four inner planets are almost as difficult to see as Pluto because they are only about one degree from the sun and have a diameter of about one millisecond of arc at this distance = about 20 AU. Worse some are back lighted as thin crescents. nor are the inner planets more easily separated from back ground stars of 25 th magnitude, when your space craft was twice as far away. The time consuming techniques by which we find exoplanets would be necessary if the milky way was behind the sun = on a few seconds examination the 4 inner planets would look much like a million stars in the same direction. The crecent shape is perhaps the best indicator, if the starship has that good resolution at 20 or 40 AU. Most of my numbers are guess work, so please correct. How long does it take to do a spectum of a million stars about 25 th magnitude, with all the present resources of Earth? Neil

[cjameshuff]

You're right that occlusion is improbable (not impossible) but the probability changes with the scale that you look at. What are the odds of Jupiter occluding one of its moons from any particular position in its ecliptic? Probability increases, and Jupiter's moons are probably much more interesting to the Enterprise than Pluto is, than Mercury is.

The scale we were looking at was planets. However, if you happen to be in the plane of Jupiter's inner moons, you will still most likely see Io (the moon most likely to be occluded) in any given observation. The probability again goes up greatly when you consider that the observing spacecraft is most likely not in that plane. Io will spend about 2.6% of its time in the part of its orbit where it could be occluded from a given point in its orbital plane, and overall has a chance of ~1/144 of being behind Jupiter from any given point of view.

Any kind of analysis by gravity takes a lot of simulation power. Maybe if you have quantum computers, it wouldn't take a lot of computation. Maybe. Writing the algorithm would be hard.

It doesn't take that much processing power at the numbers of bodies we're talking about, is not all that hard to write simulations for, wouldn't necessarily benefit hugely from quantum computers, and is pretty much irrelevant to the situation as it requires high precision data on the orbits of all major bodies and further observations to determine the accuracy of your model. Any objects that might be detectable though this method would also be easily detected directly...it's not a technique that would be of much use in this situation.

Noticing position (on the projection) is simple; trying to grab velocity isn't easy, and it increases in difficulty as you try to figure it out in increasingly small periods of time, and you need it before you can do any meaningful analysis. Describing the ellipse of its orbit with data grabbed over fractions of seconds from a ship much smaller than earth is hard. It requires extraordinary precision. To begin any kind of meaningful description, you need distance from sun or mass of object. Determining distance via parallax is hard without the kind of parallax possible from a planetary body. Determining mass from gravity is harder. (Determining the distance of stars requires more parallax than we can get on earth-- we need to use our whole orbit). I haven't done the math on how hard these hard things are; I just know that they're hard. If anybody wants to do the math, it'd be lovely

Velocity is fairly irrelevant to most sorts of analysis...it lets you determine the orbit. Which is great for going there, but not of much use otherwise. Most things of interest could be determined through spectroscopes.
It's also not that hard. The ship's velocity relative to the star is presumably at least roughly known and generally quite a bit higher than that of the planets (given that FTL ships rarely take years to travel through a system in sci-fi), a few observations will help narrow down the possible locations for the planets. The range and radial velocity to the planet will soon be knowable via radar, allowing higher precision measurements. I'm not sure why you think parallax from the ship will be difficult.

I'm not a spectrographer; reading WP, you'd think that it can be done at the speed of light, but in reality, analysis currently requires times larger than seconds by many orders of magnitude. Days, I understand. That's with the effectively limitless resources we have on Earth. I can't comment on the feasibility of speeding this process up.

A commercial digital camera does spectrography at a sufficient level of detail to separate planets from stars...3, sometimes 4 broad bands. It doesn't take days to get a picture. Detailed analysis of the planet's atmospheric and surface composition may take time, separating planets from stars can be done essentially instantaneously. This is not a hard problem...stars are hot emitters, planets are cold reflectors. Earth (or for that matter, Mercury) doesn't produce remotely as much infrared as a body hot enough to emit as much visible light as Earth reflects.

Like I said, doing complicated, passive analysis isn't hard; what's hard is doing it fast. If you try and do it a million times faster than we currently do it, you need instruments a million times more precise.

This isn't complicated analysis we're talking about.

[vasiln]

I'm sorry, I think we're talking past each other, because we're talking about different levels of analysis. Yeah, I agree, differentiation between a gas giant and a rocky planet and a star is pretty much trivial. I believe I've said so.

I'm reading a little bit past the literal question, to a question that is maybe only implied, which is, "How much of that data that they can get two seconds in to jumping into an unknown system could they reasonably get?" I think that's an interesting question. Distance and velocity are some of the first things that you'd want to grab. The importance of distance is of course paramount, because you can't talk about things like size or position until you know distance. Velocity is probably the next thing you'd grab.

I actually just did a little bit of math about the parallax part, because the math and data is accessible. You''re just dealing with similar triangles.

Sounds like our current parallax telescope is the Hipparcos. One of the most distant stars it's given us data about is V762 Cas, which is about 15000 LY away. I use "about" very loosely-- we're talking plus or minus twelve thousand LY. Still, let's give it the benefit of the doubt. It uses both sides of earth's orbit to figure out parallax-- its stance is 3E8 km. The Sears tower is a pretty big skyscraper, at around 500m tall. So we'll take two Hipparcos and put them 500m away, rather than gathering data from a single telescope at opposite points in Earth's orbit.

Let's say that we're at the mean orbital distance of Pluto, and we're looking across the system to a point on the far side of that distance-- in other words, we're going to try to estimate the distance of something 1.2E10km away. This is no longer being generous, but we figure that's what is asked of the star ships involved-- I've never heard them say "probably" yet

All that we need is 25 times the precision of Hipparcos (although we're still at some miserable tolerances).

Hipparcos is about a little over a metric ton. So, with 2000 kg in two even chunks separated by about 500m, we only need precision about 25 times as good as we're using. (Although I sure hope nobody asks us to do any rapid rotation. I don't think our telescopes are built with that much centripetal force in mind.... It's going to take us at least an hour to rotate through a full circle without putting more than 1g on each telescope. Maybe we got anti-gravity.)

EDIT: Did some math on velocity, because I was curious about that. Jupiter moves 13km a second. I just used mean orbital distance of Pluto as a viewing distance; the mean distance of Jupiter from any particular point in Pluto's orbit is higher than that. At 6E9km, 13km takes up 4E-4 arcseconds, just about Hubble's resolution. Hubble would give you the velocity of Jupiter with pretty lousy tolerance (it'd say between zero and 26km/s) in one second. Hubble's about ten tons big.

[Jens]

Of course the physical report is often followed by biology report. "Life-form readings indicate the third planet from the star is inhabited. Population 3 billion, technology level X". Now that's pushing it.

Yes, it's pushing it, but not anywhere as far as the original proposition (an FTL drive)!

[cjameshuff]

The ship's moving, likely at several hundred km/s given typically depicted travel times. You can get a baseline of hundreds of km for most of the system with two observations a second apart from a single instrument. Spectroscopy gives you radial velocity from the star, which together with parallax observations of the sun will let you pin down motion relative to it and let you determine distances to other bodies. In the worst case of a planet aligned with the ship's trajectory, you're limited to cruder or more difficult measurements, active sensors like radar, or waiting until the ship makes a maneuver that changes its trajectory, but this is an exceptional case.

(And you'd have to spin the ship up to nearly 2 RPM to get accelerations over an Earth gravity 250 m from the rotation axis.)

[vasiln]

You can use doppler spectroscopy to get relative velocity with the star, which is definitely the single most important value you want. But that's because you mostly know the makeup of stars to begin with. Don't you need to know a base spectrogram to do it for something like a rocky planet? It seems to me that you can't just solve for the doppler of starlight reflecting off a planet of a particular velocity until you know either position or velocity, to solve for the other. (In fact, don't you need to know radial velocity to do the kinds of useful planetary spectrometry discussed earlier in the first place?)

It also seems to me that using your own motion to generate parallax requires already knowing your velocity relative to the object you're trying to find the position of.

Obviously, it's a silly discussion, because as soon as you stop being totally passive all of the problems disappear Not like 5 seconds is so much more painful than 2 seconds anyways. But I enjoy thinking about it and hearing what people have to say. Even if I'm wrong

Pardon my error on rotation rate-- I squared something I shouldn't have. You're right, ~30 seconds for a full circle at 1g.

[neilzero]

The ship can start sending terawatt unique radar pulse trains toward the star as soon as they slow below c: Reflections = radar echos by bodies will start returning in an hour or less, compressed, and dopler shifted by the still high approach speed. The local beings could detect these powerful pulse trains, which would be obviously sentient, so the approach would not be covert. Radar results also require some patience, unless there is such a thing as subspace radar. Neil

[neilzero]

The ship can start sending terawatt unique radar pulse trains toward the star as soon as they slow below c: Reflections = radar echos by bodies will start returning in an hour or less, compressed by the still high approach speed. Radar results also require some patience, unless there is such a thing as subspace radar. With subspace radio, we can send probes thoughout the solar system and get instant data from the probes. Neil

[Tom Mazanec]

I specified FTL because there are enough sci-fi FTL stories to form the mass of a small asteroid...it is sort of an accepted trope in SF. For an STL starship, one would expect telescopic examination of the target star for years before the mission. Someone once said that you could have ONE exception to known science in sci-fi, FTL is the most popular. Also, if there are an unlimited number of exceptions to physical law, then they could just use magic sensors.

[cjameshuff]

You can use doppler spectroscopy to get relative velocity with the star, which is definitely the single most important value you want. But that's because you mostly know the makeup of stars to begin with. Don't you need to know a base spectrogram to do it for something like a rocky planet?

I didn't mention using spectroscopy for planets because of the greater difficulty in reliably picking out good spectral lines. Doing it for the star is good enough, you then have parallax data for the planets.

It seems to me that you can't just solve for the doppler of starlight reflecting off a planet of a particular velocity until you know either position or velocity, to solve for the other.

Why? It's a simple doppler shift with a very limited range, that due to the ship's motion relative to the system as a whole +- a few tens of km/s. It's independent of position, and solving for the doppler shift is solving for the radial velocity.

(In fact, don't you need to know radial velocity to do the kinds of useful planetary spectrometry discussed earlier in the first place?)

No. Why would you? Doppler shift affects the spectrum as a whole, it doesn't move lines around relative to each other. Doppler shifted oxygen still looks like oxygen.

It also seems to me that using your own motion to generate parallax requires already knowing your velocity relative to the object you're trying to find the position of.

Yes. Which you have, given the parallax and radial velocity data from the sun. Not exactly, but good enough to determine where the planet is in the system, given how much faster the starship is going than the planets. The planets moving the fastest and thus having the highest errors are also those changing direction the fastest, giving you a way to reduce those errors with multiple observations.

Obviously, it's a silly discussion, because as soon as you stop being totally passive all of the problems disappear Not like 5 seconds is so much more painful than 2 seconds anyways. But I enjoy thinking about it and hearing what people have to say. Even if I'm wrong

It'll take a bit more than 5 seconds for a radar pulse to propagate 60 AU across a system and back. More like 22 hours, using the distance you gave in your Pluto example. You'll probably want this for navigational information, but it's not necessary for just getting a picture of the system.

[JustAFriend]

Yup we already have powerful radars for decades that have 'pinged' moons and planets in our own solar system.

It's just a matter of how much power you want to put into them... of course you are announcing
your own presence in that system when you beam out those pulses so your FTL Captain is a huge target...

[vasiln]

I believe that the frequency shift from a moving mirror is a function of both the mirror's velocity and the angle of incidence*. Assume knowledge of your distance to star, angle between planet and star, and reflected spectrum of star. If you know position, you can solve for velocity; or if you know velocity, you can solve for position. (If, instead, you know position, velocity, and star's spectrum, you can solve for reflected spectrum, so that you can begin analyzing the planetary spectrum. That's one reason why I believe you need to know position and velocity to do spectrometry.)

Let's say you assume speed relative to star of -50 to +50 km/s (about the mean orbital velocity of Mercury). You build that error into your parallax computation, and you travel 100km/s for one second. Your certainty about the distance of the object is that it's between 1 and 3 units. This uncertainty never shakes out of the system-- if you then use your imperfect knowledge of the position of Mercury to determine the velocity of Mercury (say, via doppler, as above), you find that the only thing you know is that it's traveling between -50 and 50 km/s. In other words: when you figure numbers based on the assumption that the body is something in-between Mercury and Pluto, the only thing you discover is that it's something in-between Mercury and Pluto. If you decide not to worry about velocities so mercurial (sorry, couldn't help myself), no problem-- you just discover that Venus is something in-between Venus and Pluto. (And so is Mercury now, as far as you can tell.) I don't see how this would change if you travel 200km/s-- the only thing you discover about their velocities are your starting assumptions. I'll see if there's some easy math I can do to verify.

I didn't mean to suggest that 5 seconds is sufficient for any pings-- but 4 seconds of passive observation are worth twice as much as 2 seconds of passive observation. Pings are sweetest; if you can't get pings, get bigger mirrors and more telescope time.

I'm just going to post this now, but I'm hoping to find more information about realistic astronomical spectrometry. I don't know enough about things like molecular vs. elemental spectrometry, spectra outside of the visual range, diffraction, etc to talk very well about what kind of knowns you need to start out with. If I find something interesting, I'll at the very least edit this post. My feeling is that analyzing the spectrum of a body composed of a large number of different materials, for a potentially infinite numbers of compounds (O2 is interesting, Fe2O3 much less so), through a poorly understood if sparse cloud of dust, without even having any "start here" line suggested by knowledge of velocity, is, at the very least, hard-- but I've got to find a lot to read in order to confirm or reject my suspicions.

*Page 4 of http://cdsweb.cern.ch/record/791343/files/0409014.pdf gives a formula for this doppler shift-- for a mirror of uniform velocity. While it may be possible that this changes with acceleration, my suspicion is that solving it becomes harder, not easier.

EDIT: Spent a few hours looking through spectrometry stuff. Can't say my suspicions are either refuted or confirmed. I did find more complicating factors, as well as examples of spectrometry failing-- of people attempting to do spectroscopic analysis and not coming up with any values that made sense. I'm probably done looking into this for now.

[Jens]

Yup we already have powerful radars for decades that have 'pinged' moons and planets in our own solar system.

Yes, but remember that OP specified "2 seconds." You can't ping them that quickly--the signal doesn't return that fast (unless you have a very compact star system!).

[cjameshuff]

I believe that the frequency shift from a moving mirror is a function of both the mirror's velocity and the angle of incidence*. Assume knowledge of your distance to star, angle between planet and star, and reflected spectrum of star. If you know position, you can solve for velocity; or if you know velocity, you can solve for position. (If, instead, you know position, velocity, and star's spectrum, you can solve for reflected spectrum, so that you can begin analyzing the planetary spectrum. That's one reason why I believe you need to know position and velocity to do spectrometry.)

You've basically got two subsequent doppler shifts, one from the scattering object's point of view of the sun (which will generally be quite low for anything but severely eccentric orbits), one from the observer's point of view of the planet. The frequency shift of emission/absorption lines from the planet itself depend purely on radial velocity of the planet from the observer. For really detailed analysis, you can use a few key emission/absorption signatures to estimate the radial velocity of the planet from both the spacecraft and the sun and take into account the exact spectrum the planet is reflecting to resolve fine spectral details near spectral features of the sun, but you won't have to do this to recognize a few compounds of interest in classifying the body.

In fact, you should be able to fairly easily solve for the planet's radial motion with respect to the sun from the observed reflected solar spectral lines compared to its own spectral lines. But again, this is going to be near zero for anything in a reasonably circular orbit.

Let's say you assume speed relative to star of -50 to +50 km/s (about the mean orbital velocity of Mercury). You build that error into your parallax computation, and you travel 100km/s for one second. Your certainty about the distance of the object is that it's between 1 and 3 units. This uncertainty never shakes out of the system-- if you then use your imperfect knowledge of the position of Mercury to determine the velocity of Mercury (say, via doppler, as above), you find that the only thing you know is that it's traveling between -50 and 50 km/s. In other words: when you figure numbers based on the assumption that the body is something in-between Mercury and Pluto, the only thing you discover is that it's something in-between Mercury and Pluto. If you decide not to worry about velocities so mercurial (sorry, couldn't help myself), no problem-- you just discover that Venus is something in-between Venus and Pluto. (And so is Mercury now, as far as you can tell.) I don't see how this would change if you travel 200km/s-- the only thing you discover about their velocities are your starting assumptions. I'll see if there's some easy math I can do to verify.

The faster the ship is moving, the less the planets move between the parallax measurements of a given baseline, and the closer the measured position is to the actual position. And the planets do not move in straight lines...those that move the fastest, and hence have the most error, also change direction the fastest and yield data to refine their trajectories the fastest. The more observations you make, the fewer possibilities there are that are consistent with observation.

My feeling is that analyzing the spectrum of a body composed of a large number of different materials, for a potentially infinite numbers of compounds (O2 is interesting, Fe2O3 much less so), through a poorly understood if sparse cloud of dust, without even having any "start here" line suggested by knowledge of velocity, is, at the very least, hard-- but I've got to find a lot to read in order to confirm or reject my suspicions.

You've got the same contaminating dust between you and the sun and other bodies in the system, and it's sparse, spread over a wide range of relative velocities (blurring its spectral lines), and independent of the doppler shift of the sun and each planet. As for that, it's a simple doppler shift. All spectral lines are shifted the same way. CO2 will never be shifted relative to OH-. Your pattern-matching algorithm has a single scale factor variable to compensate for a doppler effect of some dozens of km/s while scanning for known patterns of spectral lines. You don't need to know what the planet's made of first or how it's moving, you just have to recognize a few key spectra. Matching those is a far more complicated problem than accounting for such a shift. It's like reading a sentence regardless of its horizontal position on the screen, or recognizing an image that's been scaled down slightly. It's the easy part of the problem. Accounting for initially-unknown doppler shift is certainly not going to make a starship's computers churn for any notable amount of time.

Will you know everything there is to know about composition of the planets and have positioning data accurate enough to fling probes on ballistic trajectories without correction maneuvers within seconds of entering the system? Most likely not, not with known types of technology. It's entirely plausible that you could get a good overview of the system, however, with positions accurate enough to draw a decent diagram, basic planetary classifications, and any interesting highlights like the presence of water or spectral signatures expected from life or industry.

[MikeMcc]

This seems to be assuming that you are only able to use light to detect the planets and other bodies and don't have an FTL equivalent of radar that you can ping and get an almost instant response.

[Nick Theodorakis]

What will take the longest is when they start arguing about which small, distant, icy bodies are really planets or just minor planets. No amount of computing power will speed that up.

Nick