Detection in Space Warfare
On This Page
First off, there are two broad classes of sensors: passive and active. Passive sensors just detect any emissions from the target, i.e., they passively look for the target. Passive sensors include telescopes and heat sensors. Active sensors emit various frequencies and detect their reflection off the target, i.e., they actively "shine a light" on the target. Active sensors include radar and lidar/ladar.
Active sensors are much better at detection, but have the annoying side effect of virtually placing a huge flashing neon sign on your ship that says: "LOOK AT ME! I'M HERE! SHOOT ME, SHOOT ME!!" . This not only lets all hostiles (detected and undetected) know where you are, but also gives their deadly radar-homing missiles some radar to home in on.
Passive sensors, on the other hand, are more blind but are undetectable. Much better if you are trying to hide. Passive sensors also generally can vaguely detect the presence of objects at a much greater range than active sensors. But active sensors can determine the precise location of an object with much greater precision.
Why? An active sensor emits "pings" of electromagnetic radiation in order to illuminate the target, the sensor "sees" the target if the energy returned by reflecting off the target is high enough to be detected. If the target has a small dimension compared to the angular and range resolution of the active sensor, the strength of the return signal is proportional to the inverse fourth-power of the distance to the target (i.e., signal fall-off is 1/r4). Why this fall off is 1/r4 instead of the 1/r2 you'd expect from the inverse square law is explained here and here. Basically only a fraction of the initial pulse energy is reflected back. So the target acts as if it was an active sensor emitting pings with a strength of 1/r2 of the original pulse. These pseudo-pings travel back to the original ship, suffering a further loss of 1/r2. This combines to make an effective loss of 1/r4.
But on the other hand an active sensor uses tightly focused pings while a passive sensor has to make do with whatever unfocused radiation flux the target emits.
In some SF novels, passive sensors are called "sensors" while active sensors are called "scanners."It would be a jolly science fictional idea to postulate a break-through that could detect passive sensors, keeping in mind that there doesn't seem to be any basis for this in reality. Wave your hands real hard, and vaguely mutter about "psionics", or something based on a Schrödinger's cat-like collapse of wave function (Captain, the wave function collapsed, it means somebody is peeking at us!) or specially trained experts who feel itchy sensations between their shoulders when somebody is looking at them. But to reiterate, this is strictly science fiction.
There is one cute real-world trick. If your active radar pulses mimic radio static, enemy radar detectors will filter the pulses out as random noise and fail to see them. This will make your active radar invisible. Until the enemy catches on to the trick and redesigns their detectors.
Ken Burnside notes you have to remember simple detection is NOT the same as a weapon target lock. A hostile ship can be detected a long time before your sensors have enough data for a targeting solution. Active sensors are better at obtaining a target lock. But, as previously mentioned, passive sensors have a greater range.
There Ain't No Stealth In Space
Wargames like GDW's STAR CRUISER describe interplanetary combat as being like hide and go seek with bazookas. Stealthy ships are tiny needles hidden in the huge haystack of deep space. The first ship that detects its opponent wins by vaporizing said opponent with a nuclear warhead. Turning on active sensors is tantamount to suicide. It is like one of the bazooka-packing seekers clicking on a flashlight: all your enemies instantly see and shoot you before you get a good look. You'd best have all your sensors and weapons far from your ship on expendable remote drones.
Well, that turns out not to be the case.
The "bazooka" part is accurate, but not the "hiding" part. If the spacecraft are torchships, their thrust power is several terawatts. This means the exhaust is so intense that it could be detected from Alpha Centauri. By a passive sensor.
The Space Shuttle's much weaker main engines could be detected past the orbit of Pluto. The Space Shuttle's manoeuvering thrusters could be seen as far as the asteroid belt. And even a puny ship using ion drive to thrust at a measly 1/1000 of a g could be spotted at one astronomical unit.
This is with current off-the-shelf technology. Presumably future technology would be better.
Read the essay in the Rocketpunk Manifesto entitled Stealth Reconsidered.
Now I know you do not want to accept the fact that stealth in space is all but impossible. This I know from experience (Every day I have new email from somebody who thinks they've figured out a way to do it. So far all of them have had fatal flaws.). The only thing that upsets budding SF writers more is Albert Einstein denying them their faster than light starships. But don't shoot me, I'm just the messenger. The good folk on the usenet newsgroup rec.arts.sf.science went through all the arguments but it all came to naught.
Not that that's gonna stop you from trying.
It is not like the absence of stealth in space takes all the fun out of things. Sometimes things are more interesting this way. For example, John Reiher shows how to incorporate this in to the tabletop role playing game Diaspora (incidentally, Diaspora has been awarded the Atomic Rocket Seal of Approval).
If you want to really argue on this topic, I'd advise you to cut out the middle man and go directly to rec.arts.sf.science and lay your case out before the experts. You might also want to review the section on Respecting Science.
It is a truth universally acknowledged that any thread that begins by pointing out why stealth in space is impossible will rapidly turn into a thread focusing on schemes whereby stealth in space might be achieved.
This is true. Take my word for it, I know from bitter experience.
Why Not? Let me count the ways
First off, the answer is NO, you cannot solve the problem by using a thermocouple to convert the heat into electricity.
Ken Burnside said:
Most of the arguments on thermo and space detection run through a predictable course of responses:
- "Space is dark. You're nuts!"
- "OK, there's no horizon, but the signatures can't be that bright?"
- "OK, the drive is that bright, but what if it's off?"
- "But it's not possible to scan the entire sky quickly!"
- "OK, so the reactors are that bright, what if you direct them somewhere else..."
- "What if I build a sunshade?"
- "OK, so if I can't avoid being detected by thermal output, I'll make decoys..."
- "Arrgh. You guys suck all the fun out of life! It's a GAME, dammit!"
For reference purposes, here follows some brief summaries of the more common arguments and their rebuttals.
But Scanning The Entire Sky Takes Too Long
If you are hoping to lose your tiny heat signature in the vastness of the sky, I've got some bad news for you. Current astronomical instruments can do a complete sky survey in about four hours, or less.
Ken Burnside said:
A full spherical sky search is 41,000 square degrees. A wide angle lens will cover about 100 square degrees (a typical SLR personal camera is about 1 square degree); you'll want overlap, so call it 480 exposures for a full sky search, with each exposure taking about 350 megapixels.
Estimated exposure time is about 30 seconds per 100 square degrees of sky looking for a magnitude 12 object (which is roughly what the drive I spec'd out earlier would be). So, 480 / 2 is 240 minutes, or about 4 HOURS for a complete sky survey. This will require signal processing of about 150 gigapizels per two hours, and take a terabyte of storage per sweep.
That sounds like a lot, but...
Assuming 1280x1024 resolution, playing an MMO at 60 frames per second...78,643,200 = 78 megapixels per second. Multiply by 14400 seconds for 4 hours, and you're in the realm of 1 terapixel per sky sweep Now, digital image comparison is in some ways harder, some ways easier than a 3-D gaming environment. We'll say it's about 8x as difficult - that means playing World of Warcraft on a gaming system for four hours is about comparable to 75 gigapixels of full sky search. So not quite current hardware, but probably a computer generation (2 years) away. Making it radiation hardened to work in space, and built to government procurement specs, maybe 8-10 years away.
I can buy terabyte hard drive arrays now.
I can reduce scan time by adding more sensors, but my choke point becomes data processing. On the other hand, it's not unreasonable to assume that the data processing equipment will get significantly better at about the same rate that gaming PCs get significantly better.
Now, this system has limits - it'll have trouble picking up a target within about 2 degrees of the sun without an occlusion filter, and even with one, it'll take extra time for those exposures.
It won't positively identify a target - it'll just give brightness and temperature and the fact that it's something radiating like a star that moves relative to the background.
On the other hand, at the thrusts given above, it'll take somewhere around 2 days of thrust to generate the delta v to move from Earth to Mars, and the ship will be in transit for about 1-4 months depending on planetary positions.
Surely Sheer Distance Will Hide Engine Burns
Rd = ( 17.8E6 * sqrt( Ms*As*Isp*(1-Nd)*(1-Ns) ) ) * (sqrt(0.04 * π))
- Rd = maximum detection range (kilometers)
- Ms = bogey spacecraft mass (tons)
- As = bogey spacecraft acceleration (G)
- Isp = bogey drive specific impulse (seconds)
- Nd = bogey drive efficiency (0.0 to 1.0)
- Ns = bogey "stealth efficiency", i.e. fraction of waste energy which can be magically shielded from enemy detectors. (0.0 to 1.0)
- π = 3.141593...
This assumes about one hour for a full sky scan. Current chemical rockets have Nd of roughly 0.95. Ion drives get about 0.50, and steady-state plasma thrusters 0.65 or so - both can in principle be pushed to 0.90 with some difficulty, but not much beyond that. For realistic rockets, Ns = 0.0. There really isn't any way to hide your waste energy from your opponents, short of science fiction.
A Russian Oscar submarine has a mass of about 15,000 metric tons. Say it was accelerating at a tiny one-tenth of a g (As = 0.1). A chemical rocket has an Isp of around 450 seconds, an ion drive has 21,000 seconds, and a steady-state plasma has about 30,000 seconds.
This means the maximum detection range of the chemical Oscar is about 1.2 billion kilometers (7.7 AU), and both the ion Oscar and the steady-state plasma Oscar is 25 billion km (167.4 AU). For purposes of comparison the distance between the Sun and Pluto is about 40 AU.
What If I Run Silent And Cold?
"Well FINE!!", you say, "I'll turn off the engines and run silent like a submarine in a World War II movie. I'll be invisible." Unfortunately that won't work either. The life support for your crew emits enough heat to be detected at an exceedingly long range. The 285 Kelvin habitat module will stand out like a search-light against the three Kelvin background of outer space.
The maximum range a ship running silent with engines shut down can be detected with current technology is:
Rd = 13.4 * sqrt(A) * T2
- Rd = detection range (km)
- A = spacecraft projected area (m2 )
- T = surface temperature (Kelvin, room temperature is about 285-290 K)
If the ship is a convex shape, its projected area will be roughly one quarter of its surface area.
A Russian Oscar submarine is a cylinder 154 meters long and has a beam of 18 meters, which would be a good ballpark estimate of the size of an interplanetary warship. If it was nose on to you the surface area would be 250 square meters. If it was broadside the surface area would be approximately 2770. So on average the projected area would be 1510 square meters ([250 + 2770] / 2).
If the Oscar's crew was shivering at the freezing point, the maximum detection range of the frigid submarine would be 13.4 * sqrt(1510) * 2732 = 38,800,000 kilometers, about one hundred times the distance between the Earth and the Moon, or about 129 light-seconds. If the crew had a more comfortable room temperature, the Oscar could be seen from even farther away.
To keep the lifesystem in the spacecraft at levels where the crew can live, you probably want it above 273 K (where water freezes), and preferably at 285-290 K (room temperature).
Well I'll just beam my heat the other way!
Glancing at the above equation it is evident that the lower the spacecraft's temperature, the harder it is to detect. "Aha!" you say, "why not refrigerate the ship and radiate the heat from the side facing away from the enemy?"
Ken Burnside explains why not. To actively refrigerate, you need power. So you have to fire up the nuclear reactor. Suddenly you have a hot spot on your ship that is about 800 K, minimum, so you now have even more waste heat to dump.
This means a larger radiator surface to dump all the heat, which means more mass. Much more mass. It will be either a whopping two to three times the mass of your reactor or it will be so flimsy it will snap the moment you engage the thrusters. It is a bigger target, and now you have to start worrying about a hostile ship noticing that you occluded a star.
Dr. John Schilling had some more bad news for would be stealthers trying to radiate the heat from the side facing away from the enemy.
Besides, redirecting the emissions merely relocates the problem. The energy's got to go somewhere, and for a fairly modest investment in picket ships or sensor drones, the enemy can pretty much block you from safely radiating to any significant portion of the sky.
And if you try to focus the emissions into some very narrow cone you know to be safe, you run into the problem that the radiator area for a given power is inversely proportional to the fraction of the sky illuminated. With proportionate increase in both the heat leakage through the back surfaces, and the signature to active or semi-active (reflected sunlight) sensors.
Plus, there's the problem of how you know what a safe direction to radiate is in the first place. You seem to be simultaneously arguing for stealthy spaceships and complete knowledge of the position of enemy sensor platforms. If stealth works, you can't expect to know where the enemy has all of his sensors, so you can't know what is a safe direction to radiate. Which means you can't expect to achieve practical stealth using that mechanism in the first place.
Sixty degrees has been suggested here as a reasonably "narrow" cone to hide one's emissions in. As a sixty-degree cone is roughly one-tenth of a full sphere, a couple dozen pickets or drones are enough to cover the full sky so that there is no safe direction to radiate even if you know where they all are. The possiblility of hidden sensor platforms, and especially hidden, moving sensor platforms, is just icing on the cake.
Note, in particular, that a moving sensor platform doesn't have to be within your emission cone at any specific time to detect you, it just has to pass through that cone at some time during the course of the pre-battle maneuvering. Which rather substantially increases the probability of detection even for very narrow emission cones.
(Somebody suggested using a continuous blinding barrage of nearby nuclear detonations in order to hide thrusting.)
The timescale of the radiant emission from a nuclear detonation in vacuum is measured in milliseconds. The recovery time of a good CCD array is measured in microseconds. You'll need to detonate nuclear explosives at a hundred hertz, minimum, to cover an accelerating ship. That's going to get expensive.
It also rather clearly indicates where the enemy should start looking...
The problem with directional radiation is that you have to know both where the enemy sensor platforms are, and you have to have a way of slowing down to match orbits that isn't the equivalent of swinging end for end and lighting up the torch. Furthermore, directing your waste heat (and making some part of your ship colder, a related phenomena) requires more power for the heat pump - and every W of power generated generates 4 W of waste heat. It gets into the Red Queen's Race very quickly.
Imagine your radiators as being sheets of paper sticking edge out from the hull of your ship. You radiate from the flat sides. If you know exactly where the enemy sensors are, you can try and put your radiators edge on to them, and will "hide". You want your radiators to be 180 degrees apart so they're not radiating into each other.
Most configurations that radiate only to a part of the sky will be vastly inefficient because they radiate into each other. Which means they get larger and more massive, which reduces engine performance...and they still require that you know where the sensor is.
The next logical step is to make a sunshade that blocks your radiation from the sensor. This also requires knowing where the sensor is, and generates problems if the sensor blocker is attached to your ship, since it will slowly heat up to match the equilibrium temperature of your outer hull....and may block your sensors in that direction as well.
Well I'll Just Make A Burn Then Coast
If you are actually trying to apply thrust, the upper equation comes into play, and they can see you all over the solar system. What's worse, they can measure the spectrum of your drive to estimate the thrust and use a telescope to observe your acceleration. Simple division will reveal the mass of your ship.
"Well fine!", you say, "I'll just burn once and drift silently"
But now you will be months in getting to your target. The extra time increases the chance that the enemy will spot you. It will be harder to keep your directional radiator aimed away from any enemy observers. And if you are spotted, so much of your ship mass will be radiators instead of weapons, so that the enemy ships will out-gun you by an obscene margin.
Not to mention the fact that once your initial burn is spotted, the enemy will be able to calculate your future position anytime in the future. They can set a computer controlled telescope to track your current calculated position, and will quickly spot any future course correction burns.
(Somebody suggested a ship shutting down and stealthly coasting into enemy range from a billion kilometers away)
That's nice if you can plan your tactical operations six months in advance. Not very likely, at least against a maneuvering foe. Sometime between when you boost and when you arrive, he'll redeploy and you'll have to correct your course accordingly. Which will give you away.
And you can't beat that effect by coasting in really, really fast so as to cross a billion kilometers in a week. Boosting to such a speed in the first place will require so much energy that you'll be detected even from a billion kilometers away. You can back off to twenty billion kilometers, of course, but then you're dealing with that six-month planning cycle again...
Distance cancels out of the math on that one. The detection range scales as the square root of the target spacecraft's drive power, and the drive power required to cross a distance in a given time scales as the square of that distance. No matter how far away you start, you find that there is an irreducable mimimum of time that must be spent on boost-and-coast to avoid detection. Which is generally measured in months. Fine for strategic planning, but not for tactical operations.
Only if you can predict the strategic positions well enough to plan the tactical deployment of your forces during the attack months in advance. Otherwise your space fleet will have to chose between correcting its own course and blowing its cover, opening fire from the wrong position, or aborting the attack entirely.
Accelerating to a proper vector while beyond detection range runs into the fundamental problem of how you figure out what the proper vector is. Even granted that you know the present location of the enemy fleet, you're going to be coasting for a very long time, and you've no way of knowing where they will be months in advance. So you'll probably have to adjust your course somewhere along the line, which means lighting up your engines, which means giving yourself away.
So much for being ambushed by a space pirate appearing out of nowhere. And everybody on a cruiser would know that the hostile bogey would be within combat range in two months, three days, five hours, and thirty-three minutes. You might as well take it easy and get your rest before the battle. You know the cliché: long stretches of boredom punctuated by brief moments of stark terror.
All Right! I'll Use Decoys!
And to forestall your next question, decoys do not work particularly well either. More specifically, a decoy capable of fooling the enemy would wind up costing almost as much as a full ship.
Just to make sure that we are both on the same page here, I am talking about time frames of weeks to months. Such as found when a task force weeks or months away from their target, attempting to fool the enemey observers into thinking that your are a force of twenty warships, when you are actually a force of one warship and nineteen decoys.
I am not talking about time frames of a few seconds. Such as found when a combat spacecraft, with a hostile heat-seaking missile attempting to fly up its rear, dumps off a couple of decoy thermal flares hoping the missile will be confused.
First off, a decoy needs to emit a similar amount of radiation and heat as the ship it is pretending to be. This means each decoy needs a power source comparable in size to a full ship, the same goes for radiator area.
If the decoy and the real ship thrusts, it becomes worse. The exhaust plume has to be the same, which means both the decoy and the real ship has to have the same thrust. This means the decoy has to have the same mass as a real ship, or it will accelerate faster, thus giving itself away. If you down-rate the decoy's thrust, the dimness of the exhaust plume will give it away.
So if each decoy needs a spaceship sized engine in a spaceship sized hull with a spaceship sized mass isn't much of a decoy. Why not add weapons an make it an actual spaceship?
And you'd better add defenses as well. Otherwise the decoy is nothing more than an unusually expensive, unusually easy to destroy missile.
Isaac Kuo points out that all of this assumes that the decoy and the warship are using rocket propulsion. It does not apply if they are using solar sails, laser light sails, magnsails, or other non-rocket propulsion.
But I repeat: while it is more or less impossible to use decoys to fool distant observers, it may be possible to use something like decoys in a dog-fight to protect your ship from enemy short-range antiship missiles. In the latter case, you are not trying to make a fake image of your ship so much as you are trying to break the target lock the hostile missiles have on your ship's vulnerable posterior.
Dr. John Schilling discusses why the exhaust plume of a decoy will have to have the same thrust as a real ship:
Problem is, the rate (i.e. velocity) at which the plasma is coming out, manifests itself as a doppler shift in the characteristic emission lines of the plasma. As soon as a dedicated tracking sensor focuses on the target for a second or two, the game is up. If the plasma is coming out fast, it can't help but produce thrust proportional to mass flow rate (manifested as luminosity) times velocity (doppler). If the plasma is coming out slow (or fast but in opposing directions), it will be seen to be coming out slow and thus be recognized as not a real engine.
Conservation of momentum doesn't leave much room to hide thrust, or lack thereof, in a visible exhaust plume. If you know how much exhaust there is and how fast it is moving, you know how much thrust is being produced, period. Thrust estimation by observing plume properties is in fact a common procedure in laboratory testing of plasma thrusters, and while it's no substitute for a direct mechanical thrust measurement it will certainly provide the sort of order-of-magnitude values needed for decoy discrimination.
The final step for most people comes when they say "OK, so it will always be detected. I'll just launch decoys."
Unless your decoy has roughly the same mass of the ship it's duplicating, and the same engine, it'll be easy to discern. If it's lighter, and has the same acceleration, the decoy's engine signature (which is a function of the mass being pushed) will be dimmer. If it's lighter and has the same engine signature, it'll be thrusting a heck of a lot faster.
Your best decoy is to run with commercial traffic. He may be able to ID it as 20 ships pushing 0.005 gs with a drive output of 25 GW each, giving a rough mass of 5,000 tons each, but he'll have some difficulty (until they get closer) telling which ones are the freighters and which ones are the warships...
What Sensors Reveal
When the enemy spots your ship by the exhaust plume, it not only knows that a ship is there, it also knows the ship's exhaust velocity, engine mass flow, engine power, thrust, acceleration, ship's mass and ship's course. Not only can it tell a warship from a cargo freighter with all that information, but it can also tell the class of warship, and maybe make a good stab at determining which particular member of that class it is.
In more detail: as mentioned above, propulsion system's exhaust velocity is revealed by the doppler shift in the emission lines, mass flow is revealed by the plume's luminosity, the thrust is exhaust velocity times mass flow, acceleration is revealed by watching how fast the plume origin changes position, ship's mass is thrust divided by acceleration, and ship's course is revealed by plotting the vector of the plume origin.
This means that painting the ship with camouflage in an attempt to disguise its identity is pretty pointless.
Remember the light-speed lag. Light moves quickly, but not at infinite speed. It takes about eight minutes to travel one astronomical unit. So if you are in orbit around Terra and you observe a spacecraft near the Sun with a telescope or radar, you are actually are seeing where the ship was eight minutes ago. By the same token, if you change course it will be eight minutes until the Sun-grazer ship will know.
In C.J. Cherryh's Company Wars universe, ships use both radar and something called Longscan for detection and tactical information. Longscan helps cope with the lightspeed lag of radar.
Ships have two kinds of radar: the ordinary sort which operates sublight; and longscan, which is part guess and part radar.
The way it works is this:
It takes the original information of the jump range buoy and identifies every ship and object in the system, how fast they're going and in what direction. (ed. note: a jump range buoy is a satellite parked where ships emerge from hyperspace. It gives incoming ships an instant update of the location of all known ships. If your universe does not have FTL, you can ignore it.) It calculates a likely track and shows it on the screen as a four colored line. Red is what track the ships will take if they keep on as they bear. Yellow is what they will do if they veer as much as convenient: this is a cone-shaped projection. Blue is their position if they decide to stop.
Human operators rapidly intervene and as the computer priorities them the the fastest-moving ship data, they decide, on the basis of emotional human knowledge, what those ships are likely to do when the informational wave they have just made entering the system hits them (i.e., when the ships learn that you just popped out of hyperspace). If a warship, for instance, it may turn towards them as fast as it can. An operator is assigned for each ship under consideration while the computer handles the slow craft and the other which for various reasons do not need constant monitoring.
In the meantime two things have happened: Their ship has changed course and speed either following or not following the buoy lane assignment; and the other ships one by one pick up their presence in the system and react accordingly.
But this radar image changes constantly, so when the action begins to conform to one of the projections, the computer changes the color codes, assigning red to the most probable and so on down to blue as least. So it is part radar, part computer, and part human guesswork.
The data in the bank is the best information about the mass and engine capacity and turning ability and hostility or friendliness of each ship whose computer number is on that chart; and all ships know to be in space are in that computer memory.
Now, military craft (particularly Earth Company warships) are always making adjustments and honing their turning abilities if only by the smallest degree; this fouls up the enemy's longscan guesswork and can provide surprises. Mallory's Norway for instance, has not recently tested her adjustments to the extreme, and therefore the captain herself does not know just what Norway might do if she has to. And those refinements are only tested to the fullest, of course, when it comes to a situation where a ship either turns tighter than it is supposed to, or breaks apart -- or dies in impact.
James Huff is experimenting with plotting something similar to a Longscan display. He is trying to make a "probability plot" of where to aim your guns, given the target's acceleration, maneuvers, and lightspeed lag due to the range to the target. Mr. Huff generated these plots with a custom C++ program he wrote for generating iterated function systems.
It's a lot simpler if rotation is done independently of burn, but somewhere I recall reading that Outsider ships use differential thrust of their main engines for hard maneuvering, so there will be a significant linear component in addition to the angular component of their acceleration. For a 180 degree turn, the start-turn/stop-turn accelerations cancel out, leaving the ship offset sideways somewhat from its original course, but with the same velocity. For rotations less than this, there will be a linear acceleration left over after the turn. Also, there is a maximum rotation rate that a ship can handle, but I'll ignore that.
For simplicity, I'll assume that the turn is completed before the linear burn is started, no nudging the ship into a turn and going into a full-forward burn during the turn. In addition, I'll confine the movement to a single plane, though in reality, ships would be able to change their axis of rotation mid-turn. I've already got most of the code for this written, in an IFS renderer...
All of these plots are for a period of time equal to twice the time it takes to flip the ship 180 degrees, halting the rotation... burning one thruster to rotate the ship during the first half of the flip, and the other to cancel the rotation during the second half.