85 replies to this topic

[MuKen]

blinkin, on 24 May 2013 - 02:11 PM, said:

that is all a matter of how precise your calculations and sensors are. tighter angles do require more precise measurements and calculations for accuracy, but i am generally assuming that the technology allows for as much precision as we want. this is scifi so we get to use the answer, "because science".

What I'm getting at is the information doesn't exist in the seismic waves at thsoe locations period to be that precise, regardless of technological level. But yeah, if you want to "because science" away all explanations, then we don't need to be participating in this thought exercise at all:P

[blinkin]

MuKen, on 24 May 2013 - 10:52 PM, said:

What I'm getting at is the information doesn't exist in the seismic waves at thsoe locations period to be that precise, regardless of technological level. But yeah, if you want to "because science" away all explanations, then we don't need to be participating in this thought exercise at all:P

our ears aren't that far apart and we can pinpoint an exact direction for targets miles away. each ear only registers magnitude. our brain then rapidly compares the magnitudes and gives us an exact direction. that exact direction can only be achieved by determining the convergence of the two lines. most people can listen to foot steps on a floor and then tell you the exact location of the target to within a few inches. this can be reliably done for several yards. our brain does far more precise calculations than what the seismic sensor does at much, much longer ranges (based on the separation of the sensors to distance ratio). and air is far less effective at carrying waves than denser materials like dirt or rock.

i was using the scifi cop out to bypass the precision of the instruments not the waves, but it probably wouldn't surprise me if we were capable of more than this with our current tech. for information loss to occur the material would need to be soft and spongy. <-spongy being the more important property.

[Victor Morson]

It is really confusing that they'd implement it this way (or at least, call it that) for those very reasons.

Seismic Sensors have appeared multiple times in BattleTech fluff, and every time it's a series of devices left on the ground to act as a sensor net in the area. I don't know if there's actual hard rules for them, however.

It's a ton more logical, at the least. That said, I really like MW:O's sensors, and even if they're nerf'ed I think they'll be quite good. The name, however, is in fact a bit silly. In practice this is closer to an Aliens motion tracker than anything related to Seismic activity.

Edited by Victor Morson, 25 May 2013 - 12:24 AM.

[blinkin]

Victor Morson, on 25 May 2013 - 12:23 AM, said:

It is really confusing that they'd implement it this way (or at least, call it that) for those very reasons.

Seismic Sensors have appeared multiple times in BattleTech fluff, and every time it's a series of devices left on the ground to act as a sensor net in the area. I don't know if there's actual hard rules for them, however.

It's a ton more logical, at the least. That said, I really like MW:O's sensors, and even if they're nerf'ed I think they'll be quite good. The name, however, is in fact a bit silly. In practice this is closer to an Aliens motion tracker than anything related to Seismic activity.

kinda the same thing although i think the aliens tracker sent out an active ping and cross referenced any changes. either way the key portion is listening for the sound input.

and yes i do know that i think too much.

Edited by blinkin, 25 May 2013 - 12:28 AM.

[MuKen]

blinkin, on 25 May 2013 - 12:22 AM, said:

our ears aren't that far apart and we can pinpoint an exact direction for targets miles away. each ear only registers magnitude. our brain then rapidly compares the magnitudes and gives us an exact direction. that exact direction can only be achieved by determining the convergence of the two lines.

I seriously doubt you or I could listen to something on the scale mechs are pinpointing things (whatever the human-sized equivalent of 400m is) and then put a point on a map accurately saying where it is.

And, as I pointed out before, you are IN the wave medium, not ON it, and said medium (air) is far more consistent than the ground.

[blinkin]

MuKen, on 25 May 2013 - 12:30 AM, said:

I seriously doubt you or I could listen to something on the scale mechs are pinpointing things (whatever the human-sized equivalent of 400m is) and then put a point on a map accurately saying where it is.

And, as I pointed out before, you are IN the wave medium, not ON it, and said medium (air) is far more consistent than the ground.

in or on does not matter. all you need is a point of contact with the medium. as long as the feet are on the ground it should work.

are density changes a great deal with altitude, weather, and temperature. i live in colorado in the U.S.. there are plenty of places you can drive to where the air is thin enough that even walking too fast can make you feel like you just ran a couple miles. air density increases exponentially as you decrease altitude. the density of the air has no effect on the ability to hear unless you reach something similar to a vacuum.

ok could not find a super reliable source for mech height but a consistent minimum number i found was about 8 meters. so if we compare this to a person who is 5 feet tall then the person would be able to hear and pinpoint a target out to 250 feet or 83.3 yards (less than a football field). that seems reasonable to me. also we have the fact that the feet of the mech are further apart than human ears are when things are put to scale.

[AlexEss]

TRO 3026 has the following to say: a seismic sensors is a passive scanner that detect movement on the ground, it can cover a area up to 2 kilometres . (this is the portable version but still)

So there is some attachment to lore but not even there they care to explain HOW it works.. just that it works. The same way that a warp-drive just works,

Edited by AlexEss, 25 May 2013 - 01:32 AM.

[Kageru Ikazuchi]

Google "ground sensor" for real-world examples of current tech.

While this doesn't explain exactly how it works in a moving 20-100 ton mech, or why the rage is limited to 400m ... that's where the "space magic" answer comes in handy.

[blinkin]

Kageru Ikazuchi, on 25 May 2013 - 01:55 AM, said:

Google "ground sensor" for real-world examples of current tech.

While this doesn't explain exactly how it works in a moving 20-100 ton mech, or why the rage is limited to 400m ... that's where the "space magic" answer comes in handy.

the range is just because of game play balance and even that seems to be a little bit to powerful to me. and for realism sake it should go dead any time the mech doesn't have both feet firmly planted on the ground, well unless it is a laser system that monitors ground vibrations (like the lasers we bounce off of windows for conducting surveillance).

[MuKen]

blinkin, on 25 May 2013 - 01:00 AM, said:

in or on does not matter. all you need is a point of contact with the medium. as long as the feet are on the ground it should work.

Sure it does, clearly the accuracy of a sensor goes up the better its contact is.

Quote

are density changes a great deal with altitude, weather, and temperature. i live in colorado in the U.S.. there are plenty of places you can drive to where the air is thin enough that even walking too fast can make you feel like you just ran a couple miles. air density increases exponentially as you decrease altitude. the density of the air has no effect on the ability to hear unless you reach something similar to a vacuum.

I wasn't commenting on how much variation in density there is in various areas. I was commenting on how inconsistent the density is between you and the object. Air is not going to change much. But the ground is going to have dirt areas, solid rock areas, and everything in between. The mech has no way of knowing what the variations are between it and the target.

Quote

ok could not find a super reliable source for mech height but a consistent minimum number i found was about 8 meters. so if we compare this to a person who is 5 feet tall then the person would be able to hear and pinpoint a target out to 250 feet or 83.3 yards (less than a football field). that seems reasonable to me. also we have the fact that the feet of the mech are further apart than human ears are when things are put to scale.

Alright, so let's work with that example. You're standing in a circular area with a radius of 83 yards. Heck, reduce it to 50 yards, half a football field (so the diameter is one full field), to account for the mech's feet being wider than your ears. You wear a blindfold, have something make a noise somewhere in the field. Except you have no idea how loud it is, you only know how loud it sounds by the time it reaches you. And there's random pockets of floating water all around in the air that mess with the soundwaves, and you have no idea where any of those are either. Could you then point out on a map where the sound came from?

I don't know about you, but the best I could manage under these circumstances is general direction, not precise. And I'd have no clue at all about distance. If somebody gave me a picture of the field top-down and said "mark an X where the object was" I'd no doubt be way off.

Edited by MuKen, 25 May 2013 - 02:48 AM.

[Sheraf]

blinkin, on 25 May 2013 - 12:26 AM, said:

kinda the same thing although i think the aliens tracker sent out an active ping and cross referenced any changes. either way the key portion is listening for the sound input.

and yes i do know that i think too much.

It is never too much, sci-fi shapes the way our technologies are developed

[blinkin]

MuKen, on 25 May 2013 - 02:34 AM, said:

Sure it does, clearly the accuracy of a sensor goes up the better its contact is.



I wasn't commenting on how much variation in density there is in various areas. I was commenting on how inconsistent the density is between you and the object. Air is not going to change much. But the ground is going to have dirt areas, solid rock areas, and everything in between. The mech has no way of knowing what the variations are between it and the target.



Alright, so let's work with that example. You're standing in a circular area with a radius of 83 yards. Heck, reduce it to 50 yards, half a football field (so the diameter is one full field), to account for the mech's feet being wider than your ears. You wear a blindfold, have something make a noise somewhere in the field. Except you have no idea how loud it is, you only know how loud it sounds by the time it reaches you. And there's random pockets of floating water all around in the air that mess with the soundwaves, and you have no idea where any of those are either. Could you then point out on a map where the sound came from?

I don't know about you, but the best I could manage under these circumstances is general direction, not precise. And I'd have no clue at all about distance. If somebody gave me a picture of the field top-down and said "mark an X where the object was" I'd no doubt be way off.

my point is you don't need to know the volume if you have 2 offset sensors. each one gives a direction relative to it's position. you simply follow those 2 lines out to the intersect. that is how our brain determines the direction of noises. without that you wouldn't be able to even point at a target making noise reliably.

our modern submarines use this technology for guidance, and salt water density varies greatly. (after this point i increased the font size and underlined my own words to separate them from the text i quoted)
http://en.wikipedia.org/wiki/Sonar :
4388 + (11.25 × temperature (in °F)) + (0.0182 × depth (in feet)) + salinity (in parts-per-thousand ).

Submarine navigation

Main article: Submarine navigation
Submarines rely on sonar to a greater extent than surface ships as they cannot use radar at depth. The sonar arrays may be hull mounted or towed. Information fitted on typical fits is given in Oyashio class submarine and Swiftsure class submarine.

unless the interfering material is capable of reflecting or absorbing the sound, all of the necessary information is maintained. dense materials like dirt or rock do neither.

bats use active sonar to replace their need for eyes. they even hunt small insects with it.

what you have been referring to is scattering, BUT scattering primarily redirects noise back at the target producing it:
Scattering

When active sonar is used, scattering occurs from small objects in the sea as well as from the bottom and surface. This can be a major source of interference. This acoustic scattering is analogous to the scattering of the light from a car's headlights in fog: a high-intensity pencil beam will penetrate the fog to some extent, but broader-beam headlights emit much light in unwanted directions, much of which is scattered back to the observer, overwhelming that reflected from the target ("white-out"). For analogous reasons active sonar needs to transmit in a narrow beam to minimise scattering.
Target characteristics

The sound reflection characteristics of the target of an active sonar, such as a submarine, are known as its target strength. A complication is that echoes are also obtained from other objects in the sea such as whales, wakes, schools of fish and rocks.
Passive sonar detects the target's radiated noise characteristics. The radiated spectrum comprises a continuous spectrum of noise with peaks at certain frequencies which can be used for classification.
Identifying sound sources

Passive sonar has a wide variety of techniques for identifying the source of a detected sound. For example, U.S. vessels usually operate 60 Hz alternating current power systems. If transformers or generators are mounted without proper vibration insulation from the hull or become flooded, the 60 Hz sound from the windings can be emitted from the submarine or ship. This can help to identify its nationality, as most European submarines have 50 Hz power systems. Intermittent sound sources (such as a wrench being dropped) may also be detectable to passive sonar. Until fairly recently,^[when?] an experienced, trained operator identified signals, but now computers may do this.
Passive sonar systems may have large sonic databases, but the sonar operator usually finally classifies the signals manually. A computer system frequently uses these databases to identify classes of ships, actions (i.e. the speed of a ship, or the type of weapon released), and even particular ships. Publications for classification of sounds are provided by and continually updated by the US Office of Naval Intelligence.

[Theodor Kling]

Caleb Brightmore, on 23 May 2013 - 05:00 PM, said:

If an Atlas falls in the forest and no one has seismic sensors does it make a sound?

Yes it does, if you got a microphpne on your mech.

Noesis, on 24 May 2013 - 07:41 AM, said:

I was thinking more to do with wave interference behaviour with multiple Mechs as opposed to the idea of pinpointing wave sources. i.e. interference noise potentially confusing some interpretation as a result of multiple enemy sources.

That should not be too hard ( or maybe it should for BT computer systems that obiously have problems with math going above simple calculator level). Run a fourier analysis over your sensor input and you get all the original frequencies and their relative amplitudes. At least in a homgenious medium, with the ground this is of course more difficult.

blinkin, on 25 May 2013 - 12:53 PM, said:

unless the interfering material is capable of reflecting or absorbing the sound, all of the necessary information is maintained. dense materials like dirt or rock do neither.

bats use active sonar to replace their need for eyes. they even hunt small insects with it.

what you have been referring to is scattering, BUT scattering primarily redirects noise back at the target producing it:

I think that is flawed. A solid block of let´s say coal, or marble to take something harder, might not reflect waves, but the ground is usually not uniform. And unlike sea water the density fluctuations are not gradual, but often pretty abrupt. So you get an interface layer on which waves ( of any kind) can scatter and reflect. To make thingseven worse: Different materials in the medium mean different absorbtion, so unless you know the exact composition of the ground around you the amplitude your sensor reads and the amplitude at the source are hard to correlate with each other.
But then again, even getting rough reading could probably tell you: movement behind that ridge..all you realy need to know.

Victor Morson, on 25 May 2013 - 12:23 AM, said:

Seismic Sensors have appeared multiple times in BattleTech fluff, and every time it's a series of devices left on the ground to act as a sensor net in the area. I don't know if there's actual hard rules for them, however.

I think I noticed them once or twice as a mech based systems as well, but only working when standing still yourself. Mulptiple ground devices with radio contact to the mech would get better range and resolution ( increased spacing inbetween sensors) and be permanently available ( like it was already entioned: JJs arenot mech based sensors greatest friends).

Edited by Theodor Kling, 25 May 2013 - 01:57 PM.

[Noesis]

Theodor Kling, on 25 May 2013 - 01:46 PM, said:

That should not be too hard ( or maybe it should for BT computer systems that obiously have problems with math going above simple calculator level). Run a fourier analysis over your sensor input and you get all the original frequencies and their relative amplitudes. At least in a homgenious medium, with the ground this is of course more difficult.

Yes, I can see that helping to potentially approximate positions sufficiently with a timed based sampling process.

[blinkin]

Theodor Kling, on 25 May 2013 - 01:46 PM, said:

I think that is flawed. A solid block of let´s say coal, or marble to take something harder, might not reflect waves, but the ground is usually not uniform. And unlike sea water the density fluctuations are not gradual, but often pretty abrupt. So you get an interface layer on which waves ( of any kind) can scatter and reflect. To make thingseven worse: Different materials in the medium mean different absorbtion, so unless you know the exact composition of the ground around you the amplitude your sensor reads and the amplitude at the source are hard to correlate with each other.
But then again, even getting rough reading could probably tell you: movement behind that ridge..all you realy need to know.

my point is you don't need a range reading from either sensor. all you need is a pair of directions and then you calculate where those intersect.

if you have range and direction then you obviously have exact location, BUT if you do not have exact range then you can use a second offset direction to determine range. this method is also used calculate the distance for stellar objects that are (relatively) close. the primary measurement is the parallax second or "parsec" we use the total change from one end of our orbit to the other and measure the visual change in position. our planet is roughly 8 light minutes away from the sun so the diameter of our orbit is approximately 16 light minutes a parsec (the primary measurement we use) is 4.243 light years. the ratio of 16 minutes to 4.243 years is far, far, far, smaller than the ratio of a mech's stance to the 400 meter range of the seismic sensor.

[General Taskeen]

Sneaking is OP. Nerf Sneaking with more modules.

[MuKen]

blinkin, on 25 May 2013 - 12:53 PM, said:

my point is you don't need to know the volume if you have 2 offset sensors. each one gives a direction relative to it's position. you simply follow those 2 lines out to the intersect. that is how our brain determines the direction of noises. without that you wouldn't be able to even point at a target making noise reliably.

Yes, but directional sensing is BASED on distances. Because of the fact that there is varied stuff in the way, your direction sense will never be a line, it will always be an angle of uncertainty. You are assuming perfect direction measure, which is impossible without knowledge of the intervening materials.

You have a ring of sensors, and whichever one is 'closest' will pick up the vibration first telling you which direction the vibration came from. But, if there are different materials in the way of the closest sensor and the line to the one next to it gets around some of those materials, maybe that one next to it or even further away gets the signal first even though it shouldn't have. Because of this, no matter how good your technology, you will always have a slight +/- variation in the direction sense. Technology can only compensate for this if it knows what is in the way and where that stuff is.

And if we had scanning tech to see what is out there, we wouldn't need seismic sensors, now would we?

Which brings us back to the diagram I posted before. If you have two far apart directional sensors, they can compensate for a variation in the direction sensor. But if they are near each other, that tiny directional variance becomes a HUGE variance in the extrapolated position of the origin point.

MuKen, on 24 May 2013 - 01:54 PM, said:

Yes with two separate sources that indicate direction you can get a pinpoint location without a distance measure. And the further apart those two sources are, the more accurate your location will be, the closer they are the less accurate, that's the point I've been making.

Say you have directional accuracy of +/- 5 degrees. You have two points that are 100m apart, and you are trying to locate a third point that is somewhere in the general area of a regular triangle (i.e. 100m away from each of them). Then you draw the intersection of their measured directions (which comes out to two cones of error), and the intersection of those two cones is pretty small, meaning that we have a pretty accurate guess of where that third point is.

Now say that those two measurement points are only 5m apart, trying to find a third point 100m away. Draw two +/- 5 degree directional cones pointing at the third point. What is the intersection? It is by comparison HUGE because the two directional cones have a lot of overlap. These two points are not able to draw a very accurate picture of where the third point is, despite having the same individual accuracy as in the first example.



This image illustrates what I'm talking about. All four sensors are looking at a point at about the same distance with the same accuracy of directional measurement. The green area shows how accurate they are with far points; they are able to tell where the third point is with pretty tight accuracy.

The red area shows how accurate they are with nearby origin points. While they know what direction the third point is in, they have relatively no idea how far away it is. They cannot get an accurate location.

Regarding your discussion on active sonar, sonar works because it isn't just a directional measure. It's also a distance measure. You know WHEN the pulse started, and you know roughly how fast you expect it to propagate, so from the return echo you measure distance as well as direction. There will be minor variation in that measurement as well, but since you aren't using it to make a bad extrapolation like you would be in the two-distance-measure example as above, that minor variation will not get blown up into a huge variation.

Edited by MuKen, 25 May 2013 - 06:18 PM.

[blinkin]

MuKen, on 25 May 2013 - 06:04 PM, said:

Yes, but directional sensing is BASED on distances. Because of the fact that there is varied stuff in the way, your direction sense will never be a line, it will always be an angle of uncertainty. You are assuming perfect direction measure, which is impossible without knowledge of the intervening materials.

You have a ring of sensors, and whichever one is 'closest' will pick up the vibration first telling you which direction the vibration came from. But, if there are different materials in the way of the closest sensor and the line to the one next to it gets around some of those materials, maybe that one next to it or even further away gets the signal first even though it shouldn't have. Because of this, no matter how good your technology, you will always have a slight +/- variation in the direction sense. Technology can only compensate for this if it knows what is in the way and where that stuff is.

And if we had scanning tech to see what is out there, we wouldn't need seismic sensors, now would we?

Which brings us back to the diagram I posted before. If you have two far apart directional sensors, they can compensate for a variation in the direction sensor. But if they are near each other, that tiny directional variance becomes a HUGE variance in the extrapolated position of the origin point.





Regarding your discussion on active sonar, sonar works because it isn't just a directional measure. It's also a distance measure. You know WHEN the pulse started, and you know roughly how fast you expect it to propagate, so from the return echo you measure distance as well as direction. There will be minor variation in that measurement as well, but since you aren't using it to make a bad extrapolation like you would be in the two-distance-measure example as above, that minor variation will not get blown up into a huge variation.

i was focusing fairly strictly on passive sonar since active sonar has it's own set of rules and does deal with interference issues. and also since passive sonar is basically how a system like this would have to work. i generally only mentioned active sonar for clarification and comparison.

i will admit the bat thing probably wasn't very useful or needed.

i see the image you posted, but the fact that stellar parallax works, debunks it entirely. imagine seeing a flashlight drop from about a mile away with standard human eyes (as far as sensor separation ratio) and then being able to tell the exact location of that flash light and then you will just be getting into the easy stuff with stellar parallax. in those cases we don't know the brightness of the stars either.

for a little bit of reference the "second" part of "parallax second" refers to change on a 360 degree scale. each degree is split up into 60 minutes and as you would expect each minute is then split into 60 seconds. so a second is 1/3600th of a degree. in fact because nothing is as close as 5 light years we usually divide that second into (i think) tenths of a second. these mech sensors get to use whole degrees which are 3-4 orders of magnitude LESS precise.



seismic sensors give two pieces of information.
1. direction the tremor came from because motion always radiates from the source
2. magnitude on arrival

this is a single sensor with a single free weight. this can be as simple as a pendulum hanging over sand and it will still give both of those pieces of information. the directional information will always point to the source, but the magnitude measurement will change with distance and material as you have described. that is why you need 2

if we have #1 and we have 2 sensors that are separated we can draw lines from each sensor and find the point where they intersect. motion will always be parallel to the closest line radiating from the source. if we have #2 as well it MIGHT be possible to even determine the tonnage of a mech (this is where ground variation could cause problems).

[Ralgas]

blinkin, on 25 May 2013 - 07:42 PM, said:

i was focusing fairly strictly on passive sonar since active sonar has it's own set of rules and does deal with interference issues. and also since passive sonar is basically how a system like this would have to work. i generally only mentioned active sonar for clarification and comparison.

i will admit the bat thing probably wasn't very useful or needed.

i see the image you posted, but the fact that stellar parallax works, debunks it entirely. imagine seeing a flashlight drop from about a mile away with standard human eyes (as far as sensor separation ratio) and then being able to tell the exact location of that flash light and then you will just be getting into the easy stuff with stellar parallax. in those cases we don't know the brightness of the stars either.

for a little bit of reference the "second" part of "parallax second" refers to change on a 360 degree scale. each degree is split up into 60 minutes and as you would expect each minute is then split into 60 seconds. so a second is 1/3600th of a degree. in fact because nothing is as close as 5 light years we usually divide that second into (i think) tenths of a second. these mech sensors get to use whole degrees which are 3-4 orders of magnitude LESS precise.



seismic sensors give two pieces of information.
1. direction the tremor came from because motion always radiates from the source
2. magnitude on arrival

this is a single sensor with a single free weight. this can be as simple as a pendulum hanging over sand and it will still give both of those pieces of information. the directional information will always point to the source, but the magnitude measurement will change with distance and material as you have described. that is why you need 2

if we have #1 and we have 2 sensors that are separated we can draw lines from each sensor and find the point where they intersect. motion will always be parallel to the closest line radiating from the source. if we have #2 as well it MIGHT be possible to even determine the tonnage of a mech (this is where ground variation could cause problems).

Given it only has range over 400m though, it not impossible that said battelmechs legs are far enough to triangulate. Add a semi "magic/realism" approach it not hard to also think the system would be attuned to individual mech resonances within a tight tolerance ala passive sonar.

[Theodor Kling]

Wether they are far enough apart or not, ground material fluctuations will complicate triangulation anyway: Like Raglas already said: One sensor might get a signal before the other, despite the souce beeing exactley the same distance fro boht of them. The propagation speed of sound varies greatly inbetween materials. And even if you coudl get exact directions from your sensors, their data ight conflict due to refraction of the wave. You "see" only in a straight line if there is no interface between different media in the way, or you look perpendicular to that interface. And the gound often is full of interfdaces. First example that coes to mind are forest colony and canyon maps, where you have sand or earth to stand on in the valleys, but rocky ridges between you and the enemy. so a pressure wave goesthrough sand, rock and sand again to get to you. Assuming we can aproximate the rock as a solid uniform peace.