Author Topic: Neurohelmet feedback  (Read 9546 times)

0 Members and 1 Guest are viewing this topic.

Offline shyrkonflex

  • MechWarrior
  • **
  • Posts: 284
  • l33tp0intz: +7/-0
Re: Neurohelmet feedback
« Reply #30 on: February 14, 2011, 09:50:33 PM »
My idea was not that your thinking like "Hey mech, do this n' that" and then "he" just makes something up out of it like punching or walking, but that the limbs where felt in a rudimentary way (we know the neurohelmet can insert stuff into your brain and sarna states that the position of extremeties is felt) and "mind control" was just used to do that tiny bit of dynamic motion that the mech requires in addition to the normal joysticks and rudders.
So if you wanted to puch you'd rotate your torso yourself with the stick, then "feeling" the arm of your mech and just guiding it to the target a little.
Same thing goes for walking around on bumpy ground.
I mean the "mind control" in an extremely limited way, far away form direct neural control ... just like an enhancement of movement, "humanization" if you will.
Thats kind of what I understood concerning the neurohelmet.
That would also explain mechs moving "stiff" without it.

Offline Kelmola

  • Star Captain
  • ***
  • Posts: 841
  • l33tp0intz: +62/-1
Re: Neurohelmet feedback
« Reply #31 on: February 15, 2011, 12:47:39 AM »
The relevant part in TechManual says that to punch you either choose "punching mode", point at a target with your crosshairs and pull the trigger, the 'Mech's computer does the rest; or alternatively, (especially on older 'Mechs with hand actuators), either sensors in the MechWarrior's gloves or separate "waldo gloves" fixed in the cockpit will tell the 'Mech to mimic the MechWarrior's hand motions. No neural link is used for this.

However, the "modern" - Jihad-era in the context of the quoted fluff, as opposed to Succession Wars era - neurohelmets can "clarify" intentions, as in you want to pick up something fragile (such as a human), you concentrate and think this very hard when doing the task, and the hand actuators will then use minimum force required to carry out the task. However, it says "clarify", not "completely control", implying operating the regular controls at the same time: if you mash the trigger while wrenching the joystick it will obviously produce very different results than gently pulling the trigger with feather-like hand on the stick, so it's questionable how much is done by the action itself and how much by the "thought". Also, it is implied that even this takes practice, and you still do not control the movements directly.

There is no feedback from "tactile and kinesthetical senses" of the 'Mech unless you have the Clan neural link version and are very skilled, and even then the amount of feedback is purposefully limited in order not to cook the brain.

What the neurohelmet mostly does really is just the maintaining of balance. Imagine sitting in your chair and suddenly it starts to fall over; you would instinctively try to maintain your balance, and by the power of handwavium, that is what the 'Mechs computer interpretes into movement of the 'Mech limbs. Of course, without this input, the 'Mech is still able to move but it will move clumsily and awkwardly because the computer's power of calculation obviously cannot match the imprecise sense of falling inside your brain.

In an universe where interstellar starships have armour less than one millimetre thick, such small details hardly stretch the suspension of disbelief...

Offline [IPA] Thalamus

  • Lance Sergeant
  • **
  • Posts: 460
  • l33tp0intz: +28/-0
Re: Neurohelmet feedback
« Reply #32 on: February 15, 2011, 12:53:22 AM »
Hey bud,

I'd like to reply to that befor I'm too tired to :)

Let me just state that the most difficult problem isn't actually the computing power needed to process all that neurological information. The problem is rather the limitations that the brain has, anatomically. These set certain limits to what you can get out of a brain and how you can improve its functioning.

My idea was not that your thinking like "Hey mech, do this n' that" and then "he" just makes something up out of it like punching or walking, but that the limbs where felt in a rudimentary way

To a certain degree, transferring feelings is actually possible. The difficult part is that the both incoming and outgoing information are (mostly) processed in and routed through the same brain regions. Mech and body sensation as well as control over mech and body will both overlap; I'll come back to that later.

Quote
(we know the neurohelmet can insert stuff into your brain and sarna states that the position of extremeties is felt)

Compared to other aspects of BTU, neurohelmets come relatively close to reality. But the way their working is describend is - in some ways - simply the opposite to what the brain works, and thus what's possible. If we're elaborating the idea of neurohelmets, I'd consider it waste to ignore both psychobology and neuroanatomy. There are ways to implement neuro-interfaces realistically in BTU, but some ideas neet to be revisited in order to do so.


Quote
"mind control" was just used to do that tiny bit of dynamic motion that the mech requires in addition to the normal joysticks and rudders.
So if you wanted to puch you'd rotate your torso yourself with the stick, then "feeling" the arm of your mech and just guiding it to the target a little


You can't let the brain solve those tasks, unfortunately. There is basically just 1 instance that directs visuospatial movement. While you might be able to perceive visuospatial information (tracking others) independently from your own movement control, you can't control 2 bodies at once (your mech and your own body). In order to do so, you'd have to seperate your brain from your body (not feasible); other than some obvious side effects this may actually make you unable to feel the mech as well. Short story: you can transfer some diffuse sensation onto the pilot, but you cannot let the pilot feel & control both his body and another thing at once.


Quote
Same thing goes for walking around on bumpy ground. I mean the "mind control" in an extremely limited way, far away form direct neural control ... just like an enhancement of movement, "humanization" if you will.

That's just the way neuro-information cannot be used.

Things as "having equilibrium" or "walking on bumpy ground" have to be done by computers; a pilot can either focus on his own body, or the mech, but not both at the same time.

The next point is that neuro-information is not able to "smoothen" or "humanize" mechanic movement. One aspect of the problem is the aforementioned resource problem: the brain simply can't split itself up. The only thing that comes close to this are humans who can actually can control their left and right half of the body independently. Unfortunately this is always due to brain damage or surgery which obviously has severe side effects. In order to be able to use your motor cortici independently from each other (for example to let your hands do different tricks; or to use one hemisphere to control [half] your mech and one hemisphere to control [half] yourbody) you have to disrupt communication between both hemispheres of your brain, which is only possible by surgery (or, as said, brain damage). You wouldn't want that :)


You can however use the neurofeed to get 'crude approximations' about your intentions (see below).

Quote
Thats kind of what I understood concerning the neurohelmet.
That would also explain mechs moving "stiff" without it.

I know that's the way BTU represents neurohelmets. But those helmets seem to be based on assumptions about how the brain works which are simply not true. But by fixing these assumptions we'll get slightly different neurohelmets (or neuro-interfaces, which would be more accurate and better working) that are no less awesome.


What would a neuro-interface then be able to do?


It could operate within and remedy as much of the brain's natural limitations as possible. Look, you can't let your brain control multiple instances at once. But you can extract a lot of attentional information of one's brain, which is actually a lot more valuable than fine-tuning the mech movements (which can be done by computers much better anyway).

Just imagine the following situation:

You are currently sitting in your Timber Wolf, fighting a fleeing Bushwacker. Suddenly, to your right, a poptarting Shadow Cat pops up behind a hill. Because you're surprised and frightened at the same time, your amygdala fires an intense stress response at the same time that your left motor cortici are processing the Shadow Cat's movement. Also, you think "SHADOW CAT OMGWTF".
   Of course, this shows up in the neurofeed. The computers of your Timber Wolf realize that you've been frightened by something on the right side and that you directed all attention to it and that your brain "wants to get there". So within the fraction of a second, long before (!!) you could make that decision and perform that action consciously (!!), it swings around the torso. As you concentrate on the Shadow Cat, the neurohelmet seels to suppres currently irrelevant stimuli like the fleeing Bushy or distant combat sounds. Additional activation is stimulated in the frontal cortici (to increase your attention), to lower corex areas (to keep your arousal at the idea level) and to both the (sensori-)motor and occipital cortici so that you can track and predict movements like young god. Stimulated that way, you can focus solely on the Shadow Cat; perceiving time as running in slow-motion, you fixate your eyes on a vulnerable pre-damaged spot on the mech. The neuro-interface tracks where your attention and your eyes are going; based on some probability calculations he identifies roughly the same spot on the enemy spot as you. Less than a second after the Shadow Cat popped up, you push the trigger button and rip the Shadow Cat into pieces.

At that very moment, while you are still totally consumed by the Shadow Cat, the Bushwacker turns around. Normally it would take you seconds to know what is going on, but the sensors of your Mad Cat have tracked its movements. As there is nothing relevant else on the battlefield right now, the neuro-interface stops helping you to direct your attention to the Shadow Cat. Still irrelevant information (like the SCat exploding and parts flying around) is suppressed again; the neuro-interface detects that you directed your attention to the engaging Bushy on your left... and the whole thing starts over.



That is the way I view the neurohelmet. You can't really use a neurofeed to humanize and refine your movements etc. (at least not without the costs of that outweighing the disadvantages of that greatly), but you can combine the neurofeed with tactical and targeting information to create the ultimate war machine. It's all about sharing attention, especially visuospacial attention.
While most pilots will be distracted by a lot of irrelevant stimuli, you won't anymore because you can direct all your attention to tactically relevant stimuli. While most pilots will take at least 2-3 seconds to even crudely initiate a conscious response to a sudden threat, your action will be there immediately with only the delay of electronic processing (-> light speed). Normal pilots will crap in their pants, but you have a neuro-interface that keeps your arousal constantly at the optimum level and prevents stress symptoms that impair performance. While others have a hard time hitting rapidly moving targets, the neuro-interface melts your own spatial processing with the tactical feed so that you will literally hit anything.



That's how I perceive the neurohelmet idea. The neurofeed can't magically turn a machine into a human or lets you feel as if you were the mech but it can maximize your brain performance by watching where your attention goes, comparing that to tactical information and helping you to hit hard, fast and precisely with the power of 31st-century pentium xxxvvi ;).



Summary: Neurodevices wouldn't make you feel your mech, they would make you see and know through its sensor feed. A little bit different, but as awesome and imo even more powerful --- and, at least, to a certain degree, based upon scientific evidence  ;)





dit: I forgot one little detail. That neuro-interface I talk about can also help the pilot to "zoom in" on certain stimuli. Not literally making it bigger, but if it notices that you focus on a distant target it can "highlight" relevant aspects of it. Think of it as "Enhanced Imaging"... might be a clanner thingy ;)
« Last Edit: February 15, 2011, 01:18:28 AM by [IPA] Thalamus »
The Independent Pubbers' Association

Offline [IPA] Thalamus

  • Lance Sergeant
  • **
  • Posts: 460
  • l33tp0intz: +28/-0
Re: Neurohelmet feedback
« Reply #33 on: February 17, 2011, 04:37:38 PM »
Guys I've got the feeling that my wall of text destroyed this discussion... come on, I just wanted to shift the perspective a bit  :-\
The Independent Pubbers' Association

Offline shyrkonflex

  • MechWarrior
  • **
  • Posts: 284
  • l33tp0intz: +7/-0
Re: Neurohelmet feedback
« Reply #34 on: February 17, 2011, 04:55:42 PM »
Haha, no prob.

I hear you guys, guess my view on it was a bit to science fiction ...
Sounds very reasonable, especially the sensor thing.

Thalamus, you said the brain couldn't handle human body + something else, I think thats kind of an underestimation.
If I remember correctly (I'm not sure) we are already able to put electrodes on somebody's head and have him move a PC mouse courser in addition to his body with like a few hours of training (21. century lol).
Or brain is (with practice) able to adapt to pretty much anything subconstiously, and its pretty damn powerfull.
The big problem of moving the mech directly would then be the sheer multitude of limbs, it would end up in clusterfapp I guess ...

Also mechs aren't supposed to be Transformers style agile robots but much rather walking tanks, so that fits with that quite well.

Offline [IPA] Thalamus

  • Lance Sergeant
  • **
  • Posts: 460
  • l33tp0intz: +28/-0
Re: Neurohelmet feedback
« Reply #35 on: February 21, 2011, 01:37:09 AM »
Haha, no prob.

I hear you guys, guess my view on it was a bit to science fiction ...
Sounds very reasonable, especially the sensor thing.

Thalamus, you said the brain couldn't handle human body + something else, I think thats kind of an underestimation.
If I remember correctly (I'm not sure) we are already able to put electrodes on somebody's head and have him move a PC mouse courser in addition to his body with like a few hours of training (21. century lol).
Or brain is (with practice) able to adapt to pretty much anything subconstiously, and its pretty damn powerfull.
The big problem of moving the mech directly would then be the sheer multitude of limbs, it would end up in clusterfapp I guess ...

Also mechs aren't supposed to be Transformers style agile robots but much rather walking tanks, so that fits with that quite well.


Yes, these devices exist. It's possible to move a cursor on screen with neuro-feedback, but it's not really "ready for the sales" yet. These devices have a chance 0f 60%-80% to guess in which direction the user wants the cursor to move. That's already a lot more than chance, but still far from working reliably. And this kind of precision was achieved in laboratory experiments, totally differing from real-world application. It took participants weeks or months to calibrate the device; also it takes enormous amounts of concentration to produce the same thought over and over again, until it works, suppressing everything else. It will get better with advances in neurosciences and technology, though.

Yet what I find more impressive are experimental technologies to let blind people see. I read about experiments with blind people (who were not cortically blind but whose eyes were broken).  In these experiments a devices was implanted that tried to mimic/emulate the excitation that would normally come from the optical nerve; in a way it tries to re-establish communication between eyeball and occipital cortex. Results were promising: participants could see some vague schemes and light. "See" is maybe the wrong word, it's probably rather a 'sensation of light', but as we learn more about the brain, that implant will perform better.


It's a lot of science fiction still, but if we can keep that pace up, we might even see some deus-exish gadgets before we die :)
The Independent Pubbers' Association

Offline CHH Siege

  • Lead Texture Artist
  • MWLL Developer
  • Lance Captain
  • *
  • Posts: 564
  • l33tp0intz: +44/-0
  • Inveniemus Viam Aut Faciemus
    • Clan Hell's Horses: Gamma Galaxy Reborn
Re: Neurohelmet feedback
« Reply #36 on: February 21, 2011, 01:47:03 AM »
In the books there's a few references to the pilots crying out when hit...

I do that when playing MWLL sometimes. Surprise is surprise, even when there's no pain: add in anger and/or frustration and even the sound and visual effect on a screen can be enough to elicit a yell or an anguished wail, etc., and I don't see why the same wouldn't apply in a mech, where immersion is swapped for reality and you have all your other senses fully engaged.



Proud designer of the Mk II E 'Siege Engine' and the 'Perseus'...
Lo-Wang say, "Animated .gif is worth one mega-word."

Offline CGB [CoffiNail]

  • Star Colonel
  • ****
  • Posts: 1895
  • l33tp0intz: +101/-0
  • Khan of Clan Ghost Bear International
    • Clan Ghost Bear International
Re: Neurohelmet feedback
« Reply #37 on: February 21, 2011, 08:44:40 AM »
From the star league source book. It talks about the first ever battlemech run.  It was pitted against 4 or 6 tanks... freaking crushed them.  The document is a transcript from one of the scientist's logs and describes how much joy he felt in accomplishing their goal, yet fear for making such a destructive machine.  During the trial, the last tank for added flair got crushed under the bi-pidal warmachine.  The scientists logs describe how the pilot remote controlling the tank at the HQ had tried not to hide his fear when his RC died, that he was so in fear and shock he was in such a state he had tears, and did his best to hide the fact, he had wet his pants.

;)  I get VERY adrenaline rushed when playing this or most other games.  When i was 10 playing Super Mario World on the SNES, during boss castles my heart would pound so much and i would get so caught up in it that my mom and dad joked about me having a heart attack at such a young age.  Emotions can very easily drive a verbal response, we are human, how we work.


Offline ~SJ~ Atlessa

  • Star Captain
  • ***
  • Posts: 991
  • l33tp0intz: +80/-1
Re: Neurohelmet feedback
« Reply #38 on: February 21, 2011, 03:29:53 PM »
In the books there's a few references to the pilots crying out when hit...

I do that when playing MWLL sometimes. Surprise is surprise, even when there's no pain: add in anger and/or frustration and even the sound and visual effect on a screen can be enough to elicit a yell or an anguished wail, etc., and I don't see why the same wouldn't apply in a mech, where immersion is swapped for reality and you have all your other senses fully engaged.

this ;)


also

It's possible to move a cursor on screen with neuro-feedback, but it's not really "ready for the sales" yet.

http://www.youtube.com/watch?v=iDV_62QoHjY

granted it's still experimental technology, but it makes me feel that it will be prly hit the free market this decade...


In soviet clanspace, trinary bids you.

Offline Nebfer

  • MechWarrior
  • **
  • Posts: 292
  • l33tp0intz: +28/-0
Re: Neurohelmet feedback
« Reply #39 on: February 27, 2011, 02:01:54 AM »
Well I'm not all that versed on neural helmets

From what I understand is their mostly their do deal with balance, though early on before the universe was fleshed out all that much their dose seem to have been some thought they where used more than where they turned out to be.

Melee attacks are done mainly via hitting a switch and aiming a crosshairs on the target and hitting the trigger, though as tech manuals says and mentioned earlier in this thread, neural helmets can help with this.

Though it dose seem that older and more skilled mechwarriors can get more out of their neural helmets than your typical mechwarriors. So while their their mostly for balance control they can be used for other things.

Offline CHH Siege

  • Lead Texture Artist
  • MWLL Developer
  • Lance Captain
  • *
  • Posts: 564
  • l33tp0intz: +44/-0
  • Inveniemus Viam Aut Faciemus
    • Clan Hell's Horses: Gamma Galaxy Reborn
Re: Neurohelmet feedback
« Reply #40 on: February 27, 2011, 02:39:56 AM »
Yeah, I also remember the 80's where people were sure and certain that if you died in a video game while wearing a pair of VR goggles you might die in real life ;)



Proud designer of the Mk II E 'Siege Engine' and the 'Perseus'...
Lo-Wang say, "Animated .gif is worth one mega-word."

Offline ~SJ~ Wolf

  • Penalty bench
  • Star Colonel
  • *****
  • Posts: 1341
  • l33tp0intz: +148/-2
  • The HNIC of Toast is too baller for avatars.
Re: Neurohelmet feedback
« Reply #41 on: February 28, 2011, 02:26:12 PM »
Well I'm not all that versed on neural helmets

From what I understand is their mostly their do deal with balance, though early on before the universe was fleshed out all that much their dose seem to have been some thought they where used more than where they turned out to be.

Melee attacks are done mainly via hitting a switch and aiming a crosshairs on the target and hitting the trigger, though as tech manuals says and mentioned earlier in this thread, neural helmets can help with this.

Though it dose seem that older and more skilled mechwarriors can get more out of their neural helmets than your typical mechwarriors. So while their their mostly for balance control they can be used for other things.

This^

Is exactly my understanding of them as well.

If a pilot of screaming out in pain after a shot it's because something in their cockpit hurt them no some neural feedback.

Offline Blu C

  • MechWarrior
  • **
  • Posts: 219
  • l33tp0intz: +9/-0
Re: Neurohelmet feedback
« Reply #42 on: March 03, 2011, 05:42:36 AM »
Well I'm not all that versed on neural helmets

From what I understand is their mostly their do deal with balance, though early on before the universe was fleshed out all that much their dose seem to have been some thought they where used more than where they turned out to be.

Melee attacks are done mainly via hitting a switch and aiming a crosshairs on the target and hitting the trigger, though as tech manuals says and mentioned earlier in this thread, neural helmets can help with this.

Though it dose seem that older and more skilled mechwarriors can get more out of their neural helmets than your typical mechwarriors. So while their their mostly for balance control they can be used for other things.

This is basically completely correct.  In effect the NeuroHelmet mostly is there to take up the slack from the controls being limited, allowing the 'Mech to glean some idea of what the pilot's intentions are.

The older and more skilled MechWarriors getting more out of their Neurohelmets may be a result of the 'Mechs own battle computer.  It has a limited ability to learn to predict what the pilot will want to do by observing past patterns.  Thus someone who has been in the same 'Mech awhile will end up getting a lot of help from the battle computer just because it understands their patterns so well.

From page 43 in the TechManual:

Quote
This is why it takes so long to produce a good MechWarrior.  Such warriors have to learn to how to think for their machine and yet learn how best to use the machine's own intelligence.  Yet the 'Mechs own intelligence is why novices learn to operate 'Mech's quickly.  Pardon the heresy, but it probably also accounts for incidents like the Black Pearl's last combat, where her 'Mech fired weapons even after the pilot had died.  Well-trained 'Mechs can make some surprising independent decisions.

Offline ~SJ~ Wolf

  • Penalty bench
  • Star Colonel
  • *****
  • Posts: 1341
  • l33tp0intz: +148/-2
  • The HNIC of Toast is too baller for avatars.
Re: Neurohelmet feedback
« Reply #43 on: March 03, 2011, 06:07:42 AM »


This is basically completely correct. 

Been saying it since page 1....

Offline Blu C

  • MechWarrior
  • **
  • Posts: 219
  • l33tp0intz: +9/-0
Re: Neurohelmet feedback
« Reply #44 on: March 03, 2011, 06:50:10 AM »
Been saying it since page 1....

I know.  More people should listen to you.