Musing on detection information for alien active sensors

[Steve Walmsley (OP)]

I have been wondering whether to change the amount of information you get when you detect alien sensor emissions. At the moment you know the range and resolution, which makes it easy to avoid them. It may be more realistic and challenging to just detect the existence of the sensor and its strength. You would just see 'Sensor Strength 2500' as part of the contact information. This would make things a lot more tense as you wouldn't know the range and resolution and whether a given ship would be detected at a given range. You would still know the sensor strength so you would have an idea as to long-ranged or short-ranged but not the specifics.

Comments?

Steve

[waresky]

STeve,u know my old school,same as your:) Traveller.
Fog-of-war r the most amazing and intriguing in Space.

So,deepest u coding the "Fog" and "Unknown" and deepest r the Intriguing on space traveller and survey on a new SOlar system.

An red,alone,unknow "Red Beep" r most interesting thing into a void space:))

[Elmo]

I'm all for more "fog of war".

[ZimRathbone]

Yes, this sounds good

[Shinanygnz]

I concur, sounds like a good plan to me.
How about adding a function so you could then feed that strength in and some resolutions in (perhaps limited to the tech you know) and tell when they might detect you?  So "Captain, detecting strength 3000 sensor.  They might have us at 20m klicks, but if their res is x, then we can get to 10m."  You can then have some tense times as you try to sneak in closer.  Or get bitten because the aliens changed their tech and suddenly can see you further away than they could before with the same overall strength.

Stephen

[Hawkeye]

Sounds like a plan  

[sloanjh]

Ok, so it sounds like I'm the lone dissenting voice.

When the resolution stuff was put in, the technobabble associated with it was that there was some difference in the signal (e.g. pulse repetition rate) that the emitter gave out that caused the resolution differences.  This difference in signal should be detectable by the target, i.e. it is reasonable that the target should know the resolution of the emitter.

At the same time, I don't like the crisp certainty of knowing exactly when the bad guy can pick you up on his actives.  So how about this:  Add the EM (passive) level to the design of an active sensor, with the range increasing like sqrt(sensor_rating/lowest_tech_sensor_rating).  What this represents is that there are two components to an active sensor: the emitter (active strength tech) and the receiver (EM tech).  Better passive tech should mean that you can build better receivers, hence the range bonus.  In reality, signal should drop off like the fourth power of the distance (inverse square law each way: out and back), but that's such a steep drop that it probably wouldn't be noticable; the sqrt is a compromise between the fourth root and a linear drop.  (If you wanted to be drastic, you could have the range be proportional to the passive rating - that would make it even more uncertain.)

With this suggestion, two good things happen (from my point of view):

1)  The target has a feel for the gross characteristics of the active emitter, but still is fuzzy on exactly where the detection threshold is.  This is my impression of the way things currently work both for radar and sonar.

2)  The design choices for active emitters become richer, and there are other ways to improve actives rather than just cranking up on the active sensor research track.

One other thing:  I think that my response to this thread is in alignment with my desire to make passives able to identify ships/races (in another thread).  If you look at current military signals intelligence (e.g. radar and sonar) it seems like it's all passive, rather than active.  Submarines trail other ships trying to get a passive "fingerprint".  Elint planes fly around just outside excercises trying to detect characteristics of active emitters.  In addition, submarines can launch attacks on other ships without ever going active.  Passive intelligence techniques for identifying emitting platforms are very sophisticated; I'd rather see their abstractions in Aurora move in the direction of gaining more information about the emitter, rather than less.  On the other hand, a big drawback in passive techniques is range uncertainty - it would be nice if there were an easy-to-code and easy-to-represent way of putting that location uncertainty into passive contacts.  At present, I think that abstraction is handled by the requirement that you need an active contact in order for its location to be "good enough" to fire at.  Since we don't have another way to abstract the fuzziness, this seems a reasonable way to require people to include active sensors in their designs.

John

[MoonDragon]

From the realism perspective, distance should be easily detectable by passive sensors. All you need is a stereo sensor in order to triangulate the location of the active signal. So, perhaps, that may be a requirement for more info displayed? Need double sensors (or triple if you want to triangulate in 3D)? Or maybe this can be something you add to the sensor itself. Make it so that its mass can be doubled, but that gives you distance info to the detected target.

[Hawkeye]

From the realism perspective, distance should be easily detectable by passive sensors. All you need is a stereo sensor in order to triangulate the location of the active signal. So, perhaps, that may be a requirement for more info displayed? Need double sensors (or triple if you want to triangulate in 3D)? Or maybe this can be something you add to the sensor itself. Make it so that its mass can be doubled, but that gives you distance info to the detected target.

Don´t see the need for two sensors. Put sensor on outer hull, look at target, roll ship. We _are_ in space, you know  

On the other hand, how large has the baseline to be, to triangulate a target, say, 100 million km away?

[Father Tim]

Submarines tend to need to move two or three nautical miles to get a large enough baseline to triangulate contacts more than forty thousand yards away.  That's about a ten-to-one ratio.

[sloanjh]

On the other hand, how large has the baseline to be, to triangulate a target, say, 100 million km away?

It depends on the angular resolution of your sensor.  A naive estimate of this for wave-like phenomena is going to be wavelength/antenna_size.  So if you're looking at 1mm waves with a 10m antenna, you've got roughly 1 part in 10^4, which is ~10,000 km if the thing is 100 million km away.  1m waves with a 10m antenna, on the other hand, gives a 10:1-ish number similar to the submarine example, although I suspect that more complicated stuff is going on for submarines (like weird focusing effects in the water).  Caveat: these are just dimensional analysis estimates; they could be off by an order of magnitude "in reality".

Amusingly enough, I just googled for "ccd angular resolution", and all the hits I looked into were giving roughly a few*10^-4 resolutions.

Hmmm - this means that the Aurora abstraction of getting  a pretty good idea of the location of a passive sensor contact after a few 10s of seconds probably isn't that bad, assuming that the passive platform is moving at a few kps.

John

[mrwigggles]

Would be getting estimation of resolution strength in the alien intelligence screen?

[plugger]

Goodaye,

[edit] While this post focuses on Passive sensors the same factors apply to Active sensor transmissions in the context that I mention them in. Passive is more interesting as that's where all the gray areas are but given the distances involved in the Spacial sensor war it's probably a closer fit for Aurora's active signal environment. [edit]

As already pointed out the most acute use of passive sensor technology in today's military is subsurface. Aurora's sensor model is a pretty good analogy for current sub/sub & sub/surface sensor warfare.

Passive sensor detection in the underwater realm is dependant on triangulation as per previous posts. However a key point that may be overlooked is that even with triangulation the narrowing down of a target to a specific point in a 3 dimensional grid is very difficult and time consuming. Any change in speed, course or depth of the target necessitates a restart. While it's true that you could quickly narrow down the target's position to a cube of say, 2 x 2 nautical miles, this is a huge potential target space in underwater warfare and not terribly useful.

The task of pinpointing a target with passive sensors (subs usually have a range of different passive sensors that they can use) is made more difficult by two additional factors.

The first being the underwater environment itself. There are such things as thermal layers, ducting zones, convergence etc. that significantly complicate the process. You may think that your target is only 10 Nm away when in fact it is 100 Nm distant. A lot of effort is put into identifying the anomalies in the underwater environment that may be encountered in specific geographic locations.

The space analogy would perhaps be that in deep space vacuum there is little apart from nebula dust clouds to impede passive detection but within a specific system there could be all manner of localised interferences that distort or diminish the transmission of passive sensor signals. In an unknown and unsurveyed system you would be - to an extent - flying blind as to the reliability of any passive sensor abilities.

The second factor would be the use of the knowns. By this I mean that every thing that floats or swims on planet earth is currently residing in a number of military databases with an accompanying acoustic signature, propeller profiles, operating parameters and full sensor suite information. The navies of the world put great emphasis on keeping these databases up-to-date. The reason being is that once you are attempting to track a target by passive means you are essentially solving a series of equations with multiple variables. The task is made immeasurably easier by being able to tap into your database and pluck out a bunch of these variables in order to figure out the ones you don't know.

Flipping back to the world of Aurora and contact with unknown Aliens it is apparent that you aren't going to have any advantages of foreknowledge of a range of key variables. This alone makes the task extremely difficult.

Submarines of today minus their databases of underwater conditions and ships and subs would find that passive localisation of targets more of an art than a science. Throw in differing levels of technology that they may encounter and you end up with a very shaky process indeed.

Summary

My take on interpreting Alien sensor data would be based largely on the above. I'd be leaning heavily towards a thick and murky fog of war.

Perhaps something like this...

[edit] with the provisio that I'm now talking about active incoming sensor signals, not passive.

1. Initial sensor data is presented as a range rather than a specific value. Eg. Instead of saying 'Captain we have picked up EM transmissions of strength 200' you would present it as '...strength between 100 and 400'. Alternatively the signal strength is a given (probably more likely as it's an easy thing to measure) and the potential range and/or resolution is used to provide the band of potential values. In the interests of clarity I'll stick with signal strength

2. The breadth of the range presented would be determined by two factors. The first would be the difference in technology levels. The more advanced in tech you are relative to the sensor being picked up then the narrower the band. The further behind you are the wider the band reflecting your poor understanding of advanced technologies.

The second factor would be the human one. As you are dealing largely with an 'art' then it makes sense that a good operator is going to be more effective than a poor operator. Perhaps ship experience levels could influence the depth of the band.

A final factor (yep, I said two but there's always a but...) would be your knowledge of the system in question. If you have conducted a grav survey then your knowledge of the local transmission conditions would be a plus and thus narrow the band. Conversely if it's an unsurveyed system then the band widens.

3. The breadth of the band intially reported is gradually refined as more information is gathered. At five minutes increments I would update the band and it would narrow by an amount in proportion to the ship's experience. Eg. a sharp operator can infer a lot given enough time and information.

4. An optional idea to enhance the tactical possibilities would be to factor in the recieving ships (that would be you) proximity to system bodies. Provided you are close enough to a big chunk of rock or gas your ability to be detected would diminish as would your ability to figure out the bad guy (eg. your band of possible signal strengths increases in breadth). This, in effect, is simulating the tactical terrain of submarines where they have the ability to pop-up  over / under the thermal layer. It would probably involve an additional algorithim factoring in the distance to the nearest body and the size of that body to determine if any interference exists. Don't know whether this is practical code wise or not.

So overall the situation on encountering an alien active signal would be to either immediately vacate the premises or maybe go into lurk mode in the hope you can get a more accurate fix on what you are up against over time. The go/stay decision could be quite challenging given that your signals officer is initially telling you a fairly broad range of possible strengths. How much can I rely on Mr Signals? How much do I know about this system? What is my best guess as to the Biggo Baddo Bugs level of tech? Can I risk it for another five minutes?

A few thoughts to add to the mix.

Cheers,
Plugger

[Steve Walmsley (OP)]

From the realism perspective, distance should be easily detectable by passive sensors. All you need is a stereo sensor in order to triangulate the location of the active signal. So, perhaps, that may be a requirement for more info displayed? Need double sensors (or triple if you want to triangulate in 3D)? Or maybe this can be something you add to the sensor itself. Make it so that its mass can be doubled, but that gives you distance info to the detected target.

Don´t see the need for two sensors. Put sensor on outer hull, look at target, roll ship. We _are_ in space, you know  

On the other hand, how large has the baseline to be, to triangulate a target, say, 100 million km away?

I am assuming that staff officers handle the target motion analysis for passive sensors and provide a location for the target. It's fun to handle that type of detail in a sub simulator but I think it would be more work than fun in a fleet game like Aurora.

Steve

[Steve Walmsley (OP)]

Ok, so it sounds like I'm the lone dissenting voice.

When the resolution stuff was put in, the technobabble associated with it was that there was some difference in the signal (e.g. pulse repetition rate) that the emitter gave out that caused the resolution differences.  This difference in signal should be detectable by the target, i.e. it is reasonable that the target should know the resolution of the emitter.

At the same time, I don't like the crisp certainty of knowing exactly when the bad guy can pick you up on his actives.  So how about this:  Add the EM (passive) level to the design of an active sensor, with the range increasing like sqrt(sensor_rating/lowest_tech_sensor_rating).  What this represents is that there are two components to an active sensor: the emitter (active strength tech) and the receiver (EM tech).  Better passive tech should mean that you can build better receivers, hence the range bonus.  In reality, signal should drop off like the fourth power of the distance (inverse square law each way: out and back), but that's such a steep drop that it probably wouldn't be noticable; the sqrt is a compromise between the fourth root and a linear drop.  (If you wanted to be drastic, you could have the range be proportional to the passive rating - that would make it even more uncertain.)

I like the suggestion in principle. I think similar suggestions have been made before and I decided against it. Not sure why I like it more now , possibly because it fits with the idea of making enemy sensors ranges more uncertain. Perhaps an easy mechanic to use, and for players to understand, is to divide current active sensor ranges by 10 and multiply them by the current EM tech. This would halve current starting sensor ranges but they would be about 10% better than at the moment by the time you research the 8000 RP level of EM sensors. 3x better at 120,000 RP and 7.5x better at max tech. Or perhaps I change the Thermal and EM tech to match the progression of active sensor tech and use the average value of Active/EM to determine range for active sensors.

With this suggestion, two good things happen (from my point of view):

1)  The target has a feel for the gross characteristics of the active emitter, but still is fuzzy on exactly where the detection threshold is.  This is my impression of the way things currently work both for radar and sonar.

I agree. The trick would be finding a way to display that uncertainty. Perhaps the display of enemy active sensors could accept a user-provided parameter as the estimated EM sensitivity. Similar to the current display of thermal ranges.

2)  The design choices for active emitters become richer, and there are other ways to improve actives rather than just cranking up on the active sensor research track.

Also agree.

One other thing:  I think that my response to this thread is in alignment with my desire to make passives able to identify ships/races (in another thread).  If you look at current military signals intelligence (e.g. radar and sonar) it seems like it's all passive, rather than active.  Submarines trail other ships trying to get a passive "fingerprint".  Elint planes fly around just outside excercises trying to detect characteristics of active emitters.  In addition, submarines can launch attacks on other ships without ever going active.  Passive intelligence techniques for identifying emitting platforms are very sophisticated; I'd rather see their abstractions in Aurora move in the direction of gaining more information about the emitter, rather than less.  On the other hand, a big drawback in passive techniques is range uncertainty - it would be nice if there were an easy-to-code and easy-to-represent way of putting that location uncertainty into passive contacts.  At present, I think that abstraction is handled by the requirement that you need an active contact in order for its location to be "good enough" to fire at.  Since we don't have another way to abstract the fuzziness, this seems a reasonable way to require people to include active sensors in their designs.

In many ways, making passive identification possible would solve a lot of issues within the game, particularly with regard to interrupts for passive targets. It is something I have avoided in order to provide a little more suspense regarding the identity of a target. However, in many cases you 'know' what the target is based on past active contact so the lack of an ID is more irritation than suspense. Even if you do detect a new alien class based on its thermal signature, you will still only have limited information so I think most of the suspense would remain. As you say, requiring active sensors to target a contact reflects the abstraction of not knowing the exact location of passive contacts.

Assuming we move to a passive ID model, thermal would definitely provide an ID. Perhaps EM detection of shields would also provide ID due to unique frequencies or some other technobabble. Even EM detection of active sensor emissions might be possible depending on identification of the characteristics of individual sensors. It is probably easiest to have detection based on all passive information rather than change the current active / passive split to active + some passives / other passives.

Steve