In a world just starting with developing Mecha, who would make the best "First Generation" of operators?
I know that everyone loves to show Mecha operators as "pilots", but I can't help but feel that taking a Typhoon pilot and slapping him into a Mk. I Walker would be a terrible idea.
Leaves the question of who to use still. If I had to bet money, I'd say Tank Operators would be the best to use, but for all I know they could end up just as bad as the flyboys would. What do you think, /m/?
The controls on a mech would be totally different from any other military vehicle. They'd have to be totally retrained.
I think the best bet would be guys who operate heavy equipment like cranes or diggers.
The use of prototypical limbs is already there.
Kids who spent way too much time playing video games and are good at picking up control schemes.
I'm being serious. No tank or plane or any real-world vehicle experience could prepare you for controlling mechanical bi-pedal motion
I work on the world's first fully-operational mecha at the moment (the Merkava VII, for the IDF) and we're looking at training former Tank Operators (Shirion officers) rather than pilots from the Air Force just like you're saying
(this is /m/ so no one would believe me, but still)
The operation of a fighter jet is actually similar to that of a tank's or ship's. You need people to command/navigate/steer/aim/shoot/reload/communicate.
The difference is that most of these roles in a fighter jet are supported by automated systems or even entirely run by them in some cases.
So it really depends on how your mech is set up.
Don't put down pilots too quickly. If we're talking about those "skating" mechs that dance around the battlefield on jumpjets and hover-feet, I think pilots would make a great recruiting pool.
Just not fighter pilots, but helicopter pilots. They're basically doing the same thing, just in mech now.
Linking back to the type 61 tank in OP's pic >>11676308
From what we've seen in MS IGLOO, the tank could be run by a crew of 2, a Gunner and a Driver, one of which could be the commander. Navigation is aided by electronics, it uses an automatic loader. Traditional tank crew roles have been reduced to 2, close to that of a fighter jet.
The advantage of a fighter jet pilot though, is that he/she would have experience working in 3D space.
Sometimes having the aptitude for a something new isn't about skills as much as it is attitude.
When airplanes were first catching on, the best ones often had histories as race-car drivers or motor speed-enthusiasts
IIRC mobile suits were first used and intended for space so purposing your space fighter pilots would be the best choice for that universe. While the controls might not translate over well their tactics would be much closer than anything else.
So, Drone Operators?
The military would draw from existing pools of armored crewmen and retrain them. The only difference you'd see between a mech and a tank in the real world is locomotion and some difference of tactics because of the mech's profile. Because in our physics-oppressed reality a mech will just be a stompy tank
Apparently jets and other complicated machines have mostly manual controls so in case something breaks they can isolate the problem and work around it.
If everything was automated, it might be harder to work around problems.
I thought jets just had several redundant fly by wire computers.
If the entire system goes down (which is relatively unlikely), it's not like the stick and pedals are still connected to the control surfaces by cables or rods.
I kinda just want that controler so I can find a PC adaptor and see if I can play it with other games. Though, a lot of games just stop accepting buttons after a certain point after like 4, or 8, or 10, or 14 so something like X-padder or Joy to Key would be essential.
However, initial preparation and activation remains manual. This is a layer of safety not dissimilar to the emergency stop buttons and other primary circuitry on industrial robots.
Basically, everything's fly by wire ONCE you go "okay, this can be on".
The alternative is fireballing on BSODs.
Turbofans are very finicky little things. They provide massive power under very very fine control, which lets you do amazing shit, but they're extremely precise bits of machinery.
First, the electrical system. Now we're pretty good at electrically things, but let's not forget that this electrical system is surrounded by extremely volatile jet fuel. A spark in the right place/time gets you thousands of kg of thrust. A spark (or popped transistor or whatever) in the wrong place/time gets you yourself cosplaying a fuel-air explosive quite convincingly.
Hydraulics are extremely powerful shit. Keep in mind those tiny metal tubes or inch-wide reinforced rubber hoses are holding back thousands of pounds of flammable material pressure. A good lesson when working with hydraulic pumps is if you see a "hair" fluttering out of a hose when the pump is running at full power, STAY THE FUCK AWAY FROM IT. That "hair" may very well be a tiny pinprick of leaking oil that will happily slice your fucking finger off when you touch it. Needless to say, hydraulics going bad is a quick way to do some serious damage around the system.
Just these two alone are already cause for extreme caution. As long as they're used properly they offer amazing advantages. But like any great power, the instant anything goes wrong, shit goes very, very wrong very, very fast.
When those blades are spinning at a few hundred rotations per second, a single little fanblade "breaking off" means a supersonic projectile just decided to blast through your ride from the inside... damaging all those nasty-if-fucked-up systems on the way.
A Warthog's ability to limp home after having a missile through an engine and a wing fucking torn off is a testament to to the fucking gods of fucking engineering that fucking put this fucking beast together.
>A Warthog's ability to limp home after having a missile through an engine and a wing fucking torn off is a testament to to the fucking gods of fucking engineering that fucking put this fucking beast together.
Depends what kind of blade you're talking about.
If you're thinking big woodworking circulars, then NO, because those aren't clean at all.
Well, last time I saw it happen, it was very clean all the way till the bone, then more of a tear to the other side where it cracked and got pushed partially out.
But that was about 3200 psi up close.
>A Warthog's ability to limp home after having a missile through an engine and a wing fucking torn off is a testament to to the fucking gods of fucking engineering that fucking put this fucking beast together.
I can't find a single documentation of the supposed instance of that happening, only people referencing it.
An F-15 on the other hand...
one minute of googling gets me, at least, this picture of one that landed with an engine, uh... well pic related.
Damn near any dual engine plane can land on a single engine. It's the whole reason WHY dual engines are so popular for combat aircraft. But the missing wing is another issue..The F-15 pulled it off by the principle that if you strap enough thrust to a brick, it'll fly, combined with the fact that pretty much the entire body of the F-15 is a lifting surface. But I still can't find a one-wing A-10 landing anywhere.
Gundam operators >>>>>> Aircraft operators > Choppa operators > Tank operators.
But, besided the shitty firepower, I really like the Hovercraft operators.
Unfortunately, they fucked just some small shitty aircrafts, nothing more than this.
What a shame, they deserved more.
The A-10 is designed to fly home with one engine, one tail and up to half a wing missing. They've landed with no hydraulics - fully manual, as well as with blown out wheels.
However, the tough little fuckers to my knowledge have never actually taken so much damage as to test this out in combat conditions, although a handful (like three or four) were shot down throughout the gulf war, and some of the fucked-up landers retired.
>Pilots and Tankers placed into the cockpit system of a humaniform weapon for the first time.
Hope they last longer than 8 minutes.
Yeah. I also think astronauts might be good choices, though. Aside from a lot of experience with zero-G movement, there are also lots of robotic arms and stuff astronauts use to make repairs and stuff like that, so they may be better at manipulating the limbs of a mech.
Assuming that it's Infantry who'll be getting the powered exoskeletons, the Royal Mech Corps of the future may very well be recruiting from average grunts
Too fucking stupid to ever use anything more complex than a radio, and even then you've gotta specially train that fucker to not have him fry the damn thing.
Transportation Corps and their glorious walking machines. That's where you'll get your Gen 1 pilots.
pic very much related:
>In a world just starting with developing Mecha, who would make the best "First Generation" of operators?
They'd all have to be trained from the ground up and be relatively young as a result.
So basically it's World War I air combat all over again.
If anything, if it comes to gamers. I would ace gundam pod player's could be great. The skill level in that game is crazy. From blocking, to jumping and flying. Hell you can even kneel, theres 12 things you need to have skill to do with it as ppl are trying to do it to you.
People who actually think this have never learn to drive a manual. They think a mech can be driven by two sticks and a throttle.
>move along the battlefield
>spot an enemy mech
>have to stop and aim the gun because no dedicate gunner
>Amuro, come in
>Amuro, come in. do you copy?
>shut up!I'm trying to focu-
>get shot in the back by another enemy mech because no commander to look out for
Same goes for the tanks, that why we don't make one-man tank.
Too simplistic of a control scheme
When mechs are created, pilots will probably be the gen 1 operators because while many of their skills won't translate over, one skill that WILL is their ability to memorize complex input orders and control schemes.
What Tanker or Infantryman could you imagine being able to operate in something like pic-related?
I can see early on mixing and matching, to see what happens -- not nessesarily for the mechanical training but to see which doctrines work best.
It'd be like the early days of MMA with pride: throwing everything at the wall to see what sticks.
A good excuse for this would be the early days of a new energy source, that's allowed humans to actually build these machines. If asked, you could say "Well we've had the rest working for decades now but the power-source has always been the big issue".
In practice, a helicopter gunship has the most in common with a mecha, having six degrees of freedom, a "head" (sensorpod) and an "arm" (turret). In practice, you could make the weapons-pylons pivot and remove the tail and add some folding quadrapedal legs and you've got an early primative mecha that can do VTOL, taking off into the air then when shot at falling to the ground hard and absorbing the shock to hide behind horizon and cover: It'd make hitting it with a tank or even missiles incredibly tricky in BVR.
Anything with that many joints isn't going to tank hits anyway. You should also show how other weapons evolve and change.
Three generations in, fighters and tanks have mecha-like functionality with fighters able to move like helicopters OR go fast and pivot their missiles around easily wheras tanks can now hunker down or climb steep hills with four tracks rather than two, each with a folding foot at the end. Similarly, soldiers evolve and the mecha get smaller very quickly: very rarely bigger than about 7 meters in height.
I can see automation being a big theme too, with one commander unit and three drones making four working in groups of four commanders and one overrulling commander -- or even doing a mission remotely in scenarios where the lag isn't much of a factor.
Lag is a huge factor in CQC so that could be how you introduce close combat with your robots.
I am of course just rambling and shooting the shit.
Yes, because you're going to be commanding the individual leg movements.
More realistically, yes, you'd switch gears based on the terrain type and movement type but it would be an override with the computer trying to do it automatically.
Look at modern fighters. They take the act of flying away from the pilot because a computer can fly the plane better with hundreds and thousands of flickers and twitches that keep an air-frame that should spin uncontrollably flying level at mach 2.2 -- with the pilot only providing suggestions.
Pilots provide decisions. Pilots are decision makers. They see the world through instruments which pre-digest the information into a form that allows the pilot to make better and faster decisions.
Two sticks and a throttle isn't all that unrealistic, especially if there are also thumb-sticks providing other feature.
It'd also have head-position recognition, a display solution closer to something like the Occulus Rift and it would know where your eyes were pointed, allowing you to designate targets and your automatic computer based software gunner to take aim and fire within the confines of acceptable balance disruption based on your movement.
The battlefield you see isn't going to be a raw output from cameras but it'll pass through a computer, building a 3D point-cloud image with a clear fog of war telling you what you can and can't detect. This image will be populated with images and information not only from you but also AWACs, satelites and all units in the ground including your platoon, drones and soldiers. You'll be able to pull out into third person for superior field of view and have faults and errors visually indicated to you.
Disagree with me? Computing and information is our fastest advancing and most useful technology. By the time we have powerplants that can do these things, most of this functionality will probably be doable smartphones or whatever replaces them. You won't even need to be on the battlefield.
Any, provided they have enough training and practice in a simulator. Most anyone can learn to fly a plane. Its not hard: You're just controlling a very sophisticated engine: Once you understand how it works and you've memorised conditional responses (which is arduous, tedious and annoying) and you can recite them even choked full of adrenaline then you're good.
This is stuff like knowing when you cut the fuel line and hit the firebottle or when you dial the engine down in RPM, flush it and then restart or knowing to turn off and shut off broken hydraulics quickly enough not to lose your hydraulic fluid and thus lose all pressure driving your air control surfaces. It is just pages and pages and pages of procedures. It takes years but anyone can learn it.
What's meant to separate pilots from everyone else is they can continue to do this under duress when someone else would panic or choke up.
Now that's tricky, but its valued in any vocation, whether you're a tanker or infantryperson.
The difference is if you fuck up in a plane, you're falling to the ground and your vehicle will be destroyed and you'll be court-martialled.
The stakes are higher but the skills and qualities involved are identical.
A mecha isn't going to smash into the ground if it fucks up and if it does, it has systems designed to absorb the landing usually.
Anyway, another board needs me.
I must go be an embarrassing smug self-righteous all-knowing vapid clueless retard cunt elsewhere.
Duty calls. Toodlepipski, /m/.
Thankyou. I'm also the faggot above my posts who double-spaces and forgot the tripcode.
The idea of MMA-style competition between different doctrines -- taking pilots from not just different nations but weapons and fighting styles and seeing how they mingle and how the forking development of their mecha goes is an exciting one.
On one hand, power-armoured infantry. On the other, fighter-gunships with arms that can VTOL that has some loose humanoid resemblence seeming like something out of Five Star Stories. On the other, big hulking heavies with massive shoulder-cannons, treads in their legs and a love of heavy growling engines backed up by big beefy conal space-shuttle thrusters.
Can you imagine what they'd be like as a combined force with their different styles, all filling inter-complimentary mixed-armaments roles in maneuver style warfare.
The hard part honestly is inventing an enemy and an environment/world that will tax them, meriting their advancement and forcing them to work together. You could gradually start here and grow this into something ridiculous.
Explain how fingers will work then. Not to mention jumping, firing stances, and oh, everything an infantry does, which is the whole point of having humaniform mechs. Explain how it works with two joysticks and a pedal.
not the anon you asked, but you could use pressure sensitive analog triggers on the grip or something like they do in gundam
Which would be more than just sticks and throttles right that gamers are used to, right? That was my argument about the stick and throttle from above. You need pressure sensitive triggers for each finger and variable twist and turn connectors on the sticks and handles just to work out a normal human range of arm movement. Then we come to the legs, which include kneeling, jumping, crouching, prone, before we even touch on propulsion assisted movements. Its just not that simple for pilots and tankers who have "mastered" a stick and throttle game to work a mech, much less gamers
Sure, okay. Lemmie just find some pretentious twaddle to motivate me into writing. This will be covered in everal posts.
Hold onto your butts.
If you think you'll be doing a 1:1 for pressure detection, you're nuts. The machine will have a moving pressure and friction sensitive fingertip that detects how much resistance the material is giving and how much slippage is involved by also calculating changing movements in the fingers and simulating not only the grip of the object but also its balance and the tension NOT the positions needed to hold the object in place, since positional motors slip under stress and are useless for most of what humans actually do: You'd be using some sort of tension detecting drive in every joint, likely either a form of artificial muscle (material which expands and contracts) replicating the fingers or vice-verca.
As for physical control, it might be something like this for very fine tasks
General manipulation will almost certainly be done in software. For example, you won't fire a gun by pulling a trigger-finger: You'll have a socket on the palm that has your onboard telemetry talking with the digital scope on your gun to work out relative motion, wind direction, incline, shakiness, distance, incline drop and point dynamics to computer the ballistics of the shot: Modern tanks actually use a commander's pod which is like a head on a mecha: it swivels round (basically being a cluster of optics) and designates a target. The turret automatically swings into position and the gunner is then responsible both legally and mechanically for taking the shot based on decisions they make, providing feedback to the commander. While the gunner is taking the shot, the commander then designates a new target. What we're essentially doing in terms of ballistics is removing the gunner and replacing them with software and leaving the firing timing and legal accountability to a human being.
Analogue trigger with a central click. Based on what the computer works out is your current scenario or position, the machine reacts accordingly, essentially the button becoming an "apply vertical action" button while the machine works out how.
Prone? Squeezing gently will force you back onto your feet.
Falling over? Gently squeezing will use thrusters to auto-correct your balance, assuming you haven't already set the toggle-switch on your HandsOnStickAndStick already.
Moving either forwards, backwards, left, or right? Thrusters in that direction.
Making a lateral turn but not a lateral maneuver? The thrusters will assist the speed of your turn.
Hold the stick all the way down during the turn? One foot anchors down and the thrusters fire aggressively, swinging you around the foot -- the same way an ice-skater or hockey-player does what's called a lock-turn/lock-twist
You could put harsh corrective thrust on one analogue trigger and softer more generalized thrust on another, letting you do hard side-steps one moment then soft flowing maneuvers around large structures without hitting them.
If you ease totally off the softer thruster and squeeze hard, applying the click at 100%, you get a thrust assisted jump. Hold down and VTOL like flight follows.
If your machine allows, shifting a lock-switch on the left stick (responsible primarily for maneuver control) will switch from making you do it automatically to switching your thrust control from being oriented on the Y axis (where Y is up) to a Z-oriented axis where forwards is the primary axis of lift. Forwards on the left-stick becomes gain height, down lose height instead of moving forwards and backwards. Throttling could be achieved with either a mouse-wheel within easy reach or the triggers or the foot pedel: its entirely up to you.
So you're saying you can put the cursor on a target, pull the trigger, and a computer will do everything to pick it up without needing additional input from the user? No need to input that you want to 'pickup' the object instead of 'smash' it? Cause that's what I'm asking, how you can program a mech external utilities to be so simple that a 'master gamer' who uses two sticks and a throttle and maybe 6 buttons to perform all the functions a mech is expected to do realistically.
Given that we're not dealing with a totally humanoid body (it most likely has an exo-skeletal construction, lacks a spine, lacks many of the types of joints we have) we can actually get away with lots of things here that we couldn't with the human body.
For example, provided your hand-weapons are belt-fed, you could easily hold what would on the human scale require two hands to fire, because you've got the benefit of choosing the proportions of your machine and not being limited to the human design. On paper you could easily have one large cannon in each hand, each feeding from a scope on the weapon itself, firing at individually designated targets by the driver.
Target designation, as stated will be based not on head-position but gaze-position: The system is aware of where you LOOK with your EYES inside your helmet, at what distance/range you're attempting to pan to based on how the muscles in your eye move and will apply feedback accordingly. You look, snap designate within the limits of balance-control (which you assign yourself), arms automatically guide, you pull weapons-triggers to render weapons live and the computer decides when to shoot based on when its ballistic computer estimates a probable hit, meaning you'd need less ammunition. Alternatively you slip the thumb-toggle on the right stick and you can manually decide when to shoot, with a progress-bar indicating the statistical likelihood of a hit.
Tanks already do this stuff.
>Firing stances II:
If you need a specialist firing stance, the IK solver looks at the weapon judges its weight based on the resistance in the various components as it handles and moves the tool around. From this, it computes the density and weight of the object and how it balances, works out the weapons output trajectory and computes a solution that places the least strain on the machine or benefits the weapon best. It might present you with several visual options with different groupings of characteristics in diagram, letting you make the selection or even shift postures. Between combat, you could even adjust the posture and the responsiveness: Its not unreasonable to have a few mouse-wheel style dials you can adjust that vary how quickly an arm swings, how much your balance is taken into account verses that of the weapon, how much recoil compensation is applied, things like that -- letting you override these characteristics or key them in manually if the system doesn't know, then it adjust them gradually based on feedback it gets from ongoing testing.
You could also issue direct gaze-override: A big ring following your gaze that the weapons are more likely to be biased to follow, like an auto-designator. You look, provided you hold triggers, it takes aim then shoots. No need to confirm a target selection or review information about it.
You're largely asking me "how do I move this individual joint?" -- the question you should be asking is "what actual considerations do I need to feed into the computer that it might not be able to discern for itself?" The willy nilly stuff is too tedious to enter as and when you do stuff: You figure this shit out ahead of time and show the computer how to know which to use, then give the operator the option to override with their own selection of things the computer can do. Over time, the library of things computer can do is going to get really advanced very quickly based on the needs and experiences of operators.
Again, that's more than just moving a throttle up and down on one axis. It requires a lot more coordination. Also, your explanation doesn't account for the fact that all those inputs are on the same sticks that you are controlling your arms from, which you haven't mentioned how the arm movements will be controlled other than "leave it up to the computer". Like this, you're saying that one stick can do it all.
You missed my point or are not aware of what shooting in a combat environment actually implies. I'm talking about the need to have dynamic movement of the weapon limbs or hands holding the gun to be able to hold the weapon in positions other than right in front or 'aim down sights'. Such as holding it over head or around a corner, and firing the weapon while keeping the main body out of the firing sights of the enemy.
>Again, that's more than just moving a throttle up and down on one axis. It requires a lot more coordination. Also, your explanation doesn't account for the fact that all those inputs are on the same sticks that you are controlling your arms from, which you haven't mentioned how the arm movements will be controlled other than "leave it up to the computer". Like this, you're saying that one stick can do it all.
Legs are likely to be controlled by a pair of analog foot pedals.
>oh, everything an infantry does, which is the whole point of having humaniform mechs
Hunkering down, switching weapons, leaning against and moving based on cover, orienting the self based on the enemy, walking based on the terrain, climbing, grabbing and moving things?
I've basically given you the outlines of some very real solutions to these problems.
Arms? Waldo. You're not going to be kicking people with your legs: You're an 8+ ton robot for fuck's sake: If you fall, you're going to break something unless you're really well designed. Given that materials science is inherently limited by energy technology, if we're powering mecha having huge yields of advanced materials we're not even talking about yet in academic circles as having practical applications will probably be the norm.
Hypothetically, we've right now already got cameras which can see through the skull and watch brainwaves moving. The trick is our computers and our algorythms at this point are too constrained operationally to discern what the output actually means. More important is that people don't generally have an excellent account of their own neural impulses and needs, with muscle-memory and neuron clusters in the body acting as limiters: The brain actually only provides how much force to apply and how much tension the force should stop being applied at to any given muscle in the body. What results in complex motion is those muscle-groups working together.
Sure, and the first think tank that come up with a mech that can perform to my specs will eat yours up.
So its not two sticks and a throttle then.
Personally, I'm thinking that using the current object recognition wire frame overlay and subjective pressure feedback would allow a machine with a bifocal camera to identify an object's shape, receive pressure feedback when it grips, and allow for it to lift the object with a constant rate of speed using what you already mentioned.
To continue my reply to >>11684885
The outlines you gave are in very simple scenarios with clear and definite objectives. It doesn't consider the complexity of the range of movements the body can perform on any one point at any one scenario. For example, you have a running mech. You want it to go into cover behind that house. How do you control the mech to adapt to cover, when there're several options the mech has to chose from, up to smashing through the house, jumping off the house, sliding to the house, stopping at the house and leaping over the house. Saying that you can assign option selection to triggers can't cover the myriad of possibilities of actions the mech can take in any one situation. This isn't COD where a button prompt asks you if you want to leap over the debris. Anything that is fully automatic with build in presets is naturally limited by design.
>Again, that's more than just moving a throttle up and down on one axis. It requires a lot more coordination. Also, your explanation doesn't account for the fact that all those inputs are on the same sticks that you are controlling your arms from, which you haven't mentioned how the arm movements will be controlled other than "leave it up to the computer". Like this, you're saying that one stick can do it all.
You don't understand what a computer is capable of here, do you? Look, as a soldier you don't need to do complex gymnastics. You almost never ever need to punch someone or do MMA in any form. And even if you do, as a machine these kinds of strikes wouldn't actually be that damaging compared to weapons-fire.
Look. Okay. At some point I lost you. Let's go back to basics.A modern fighter pilot doesn't actually control what the flaps on the plane do. The plane makes millions of adjustments a second. The body is intentionally designed to sink like a rock, to be aerodynamically unstable and these corrections keep it in flight. When maneuvering, these instabilities are taken advantage of to allow the high alpha of the plane: its minimum turning circle to exceed that of a stable aircraft because it is out of control in an aerodynamic spin but in a controlled way. The pilot actually only provides suggestions to the computer. "I want to go here, this quickly" and the computer says "okay" and crunches the trajectories and says "I can get you here in 0.3 seconds" or something and then acts upon it.
Just hear me out. I'm explaining a fairly complex process here. It took me a while to get. Then I'll deal with questions individually.
For this reason, pilots actually have what's called a stationary control stick: It doesn't move inside a socket like a joystick does. Why? Well, ever play racing games and over-compensate for a turn and end up darting left, right, left, right in a zig-zag trying to find your line? That's called fish-tailing and its the biggest loss of speed when racing. In a plane you don't have that possibility: you lose too much speed and the pilot induces a stall: its literally called a pilot induced stall. This new stick which measures pressure in any given direction but has zero travel time back to its center doesn't bob back and forth: you let go and the stick is totally neutral and the plane snaps to that angle. Traction control works the same way: Computers can drive machines better than you can.
Why am I telling you this? Do you drive every single joint of your legs when you're walking? Try doing it now while walking, moving the joints individually in sequence. You'll either walk normally because you're not actually doing what i told you to or you'll fall flat on your ass like QWOP. When you want to go somewhere, your brain is doing the same thing as that computer. You say where you want to go. Your legs work out how to do it. Its one of the first things we ever learn to do without any complex thought under auto-pilot -- and that software system in the plane is based on the way we as humans walk.
See, if a pilot has to focus on every single motion of the joint, they're so busy driving the machine that they don't have time to make strategic decisions. They'd be fishtailing every single joint that the machine would fall flat on its ass. QWOP. We've found the same is true of fighter pilots, helicopter pilots, tankers and even people: olympic athletes develop that "autopilot" mentality after about 400 hours of doing something and within 10,000 hours, they're considered "masters" at what they do. Now imagine instead of needing to train the machine, we feed the computer simulations based on weight measurements. Eventually it will not only know how to walk with its own weighting but it will know how to compensate for any differences in proportion, weight bias, height, grip, tension and strength. And that software can be in every single machine on the battlefield.
The special cases you're talking about where you need special action are quite rare. A joystick and throttle with the joysticks able to WALDO a bit if you're doing repairs or something tedious would be more than enough here. Why? Well, even though two sticks and two pedals doesn't look like much, it can be triggered contexturally. Context here is the difference between computers with flippy switches that have to be programmed by hand every time and computers with proper user interfaces.
If you're asking me "why don't we make the machine follow the motions of the pilot?" its because our bodies aren't weighted like even a 1 ton robot, let alone a 5 ton robot or a 50 ton robot. The way we deal with balance and correct ourselves is totally different not only due to the design specifications of the machine (which will be able to do things you can't -- and because there are BETTER shapes than people to be, even if they are loosely humanoid: No spines, big rugged joints, small heads and you already have totally different shapes from a human).
As a direct consequence of this, you're running into lots of problems. And even if a pilot does learn to walk like this? It'll be like the way sailors learn to walk on ships: they'll be wonky on land because they can't adapt back for a flat non-moving structure which is largely responsible for why everyone always thought sailors were drunk. They were to a degree (becoming their reputation is something we see throughout human history) but primarily they were unable to re-adapt quickly enough. The same happens if you flip human vision upside down using lenses: The brain compensates and people learn to see "upright" within a week. Then, when you take them off, everything is upside down again while they re-adjust.
The other answer you're all thinking of giving me is "why not do it neurally if direct feedback doesn't work?"
I'll get to that in a second unless you'd rather do some more back and forth first.
You don't understand soldiers from what you say. In the modern world, there is no such thing as a 'simple' soldier. Do you think rappelling from a helicopter to a rooftop ISN'T complex gymnastics? While carrying a 60 pound bag and having a rifle in one hand?
Anyway, back to the topic, yes, I understand what you are saying that computers can perform the control of the plane and make automatic adjustments of each individual part so that the pilot only has to move the stick to move the plane as intended.
But using your analog, a mech would require a magnitude more computing power just to move an arm, since there are more control surfaces to calculate for compared to the flight surfaces of a plane.
>The outlines you gave are in very simple scenarios with clear and definite objectives.
Yes, constrained solutions in engineering are the solutions which work the best. Occams razor is the backbone of coming up with any answer to a problem.
>It doesn't consider the complexity of the range of movements the body can perform on any one point at any one scenario.
There's this thing called the 80/20 rule. 80% of the time you need less than 20% of the features available to you. If you bring new features in contexturally then you can start looking into making sure the right 20% iof the other 80% is available whenever you need it or that its easy to call up the other 20% of the time.
>For example, you have a running mech.
Right, say we're 8 tons running at 20-30 mph or so over terrain that's as stable as concrete.
>You want it to go into cover behind that house.
Right. This is a software problem, not an interface problem. This would, seriously, be part of your 20% that you'd need most of the time.
Sorry, 4chan keeps clipping my posts. will finish this in a second.
Yes, they were experimented with because truck couldn't get through the thick jungles of Vietnam. When the US pulled out however they stopped research into the idea.
And while the computer can automatically compensate for all the balancing and 'unconscious' correction the brain does, it still requires the operator to provide instructions on the specific action it needs to take. Again, this comes back to the myriad of options menu I mentioned above, which a operator has to select with the stick and throttle.
Yes, this is that 20% part, but the thing is that you need to be very quick to provide that order to the comp in that window that you are about to hit that house. I'll wait for your next post.
>How do you control the mech to adapt to cover, when there're several options the mech has to chose from, up to smashing through the house, jumping off the house, sliding to the house, stopping at the house and leaping over the house.
Well, moving all the joints individually you'd be less efficient than the computer, so you'd fuck up the action and do the wrong thing here: using too much force and toppling the wall or too little and fucking things up. The trick is to feed the suggestion into the computer with good solid interface design here.
Ask yourself, for example, how it works in games. I know this sounds stupid but bare with me: a game is a very very simple simulation where you use inputs to suggest actions to an object ingame. The object doesn't handle realistically but its a very efficient way of seeing how interfaces work and how they don't. I do agree: Anything fully limited is inherently constrained. But at the same time, you do want an outcome that IS constrained: when you want something to happen, there are also things you don't want to happen and the system would have to be aware of this.
In practice for example, you could use gaze recognition and the onboard sensors to recognise where surfaces and shapes are, probably using sonics or basic spectrum tests to determine the hardiness of the structure and how much it can endure, relative to your own weight/mass.
In english, this translates to a wireframe overlay of what the computer thinks are strategic elements in your environment. Say for example, you're shooting but you don't want to cause collateral. You designate some of those elements, but you can still override with that extra click sensor when the firing trigger is at 100% (aircraft use something similar).
Good grief Half-chan's post length limits are a headache. Making another post.
Sorry if I come off as pompus: I have a good idea of how this works but explaining it is hard and we've had things that already do this stuff for decades.
>You don't understand soldiers from what you say. In the modern world, there is no such thing as a 'simple' soldier. Do you think rappelling from a helicopter to a rooftop ISN'T complex gymnastics? While carrying a 60 pound bag and having a rifle in one hand?
It isn't. Repelling is a simple controlled fall down a line. Even being in full battle rattle with a semi-stable mount doesn't change that.
>what is fly by wire
>what is macros
You don't actually think air plane pilots have an individual lever for each control surface, do you?
Most pilots with decades of experience wouldn't be able to fly that thing. I fail to see your point.
Now the trick here is to look at the object and for the right contextural option to be available to you and to alter your movement behavior accordingly.
Here, to use cover we'd want something like an adherence option, where the computer is reading your proximity to the object and you have to try harder to leave the imposed "gravitation" of the object in software, so to speak. Similarly, if you try to ram into the building, the software steps in and won't let you make that move. This would be similar to how DARPA are playing with software for drones that recognises walls and uses adherence like code to say "don't hit this" -- and while the instructions from an operator can tell the machine to do some pretty aggressive maneuvering, the software steps in and prevents the collision. We already have this in traction control: Its called bounding.
>Anything that is fully automatic with build in presets is naturally limited by design.
>Saying that you can assign option selection to triggers can't cover the myriad of possibilities of actions the mech can take in any one situation. This isn't COD where a button prompt asks you if you want to leap over the debris.
Well while using a computer normally you have a fuckton of options tucked away, imagine if the computer knows the situation its in, knows the possible results and is offering solutions that you control the perameters to as they're executed.
its role isn't to "do everything": Its your motor cortex. You assign action and perameter and adjustment and correction and it provides a simplified data in/data out that strips out the tedious shit.
To make a truly comprehensive UI would take a long time and would be very very modular -- but a UI like this will always beat some idiot shooting at you with a goddamn WALDO running with his own legs.
Post-limit again. brb baked potatoes.
Sorry, lost my orientation. Where were we on this?
Am I answering questions, doing back and forth or trying to explain the concept?
I like discussions like this because I usually learn a few things and the concept grows and improves.
>Most pilots with decades of experience wouldn't be able to fly that thing. I fail to see your point.
The SR-71 is also an over-engineered super-prototype from a time before computer miniaturization would allow all those switches and gauges to be hidden.
>Yes, this is that 20% part, but the thing is that you need to be very quick to provide that order to the comp in that window that you are about to hit that house
There is no window. It presents three options: Smash my ass through, full stop/don't penetrate and adhere. Going OVER it is as simple as firing thrusters with the left analogue trigger which begins with a thrust assisted jump, more thrust then when you let go you fall and the legs adjust to absorb the landing shock while keeping you moving automatically.
How you smash the wall, which side of you hits it and the velocity is all decided by the speed you strike it at. If you want not to damage the wall or to adhere, the system steps in like an auto-pilot. You still have adherence but its slowing you down and preparing to make a full stop both with thrust and with leg anchorage and body balance correction.
Eventually it doesn't even need to present the options: You have a collision mode. Depending on what its toggled to on mousewheel determines how hard you want to hit something, with zero being stopping. If near the object, you can then call up what I'd call an motion/action style key and tell it to do the adherence thing which keeps you moving near the wall. A clever machine sees the collision mode at zero and adheres to the wall automatically instead of trying to pass through it.
This is of course a concept. A better qualified engineer and good solid back and forth with pilots would invent a better solution for this problem.
As far as I can tell, it is. It's a fairly standard shooter game with a slow pace, whose main selling point is the level of fiddly intermediate stuff sitting between the player and the game. You could say its selling point is having overcomplicated controls, but that would be obscuring the fact that overcomplicated controls can be quite fun.
Btw, sounds like its becoming a auto verses manual control debate. And you don't sound pompus, you just know your shit. I'm trying to suggest that input controls might not be the way to bring out the full efficiency in a system that has the same potential as a human body.
Yes I agree that leaving the computer to handle the adjustments is just more efficient. And I already mentioned the wireframe detection system above for the computer to understand the environment and predict the options the operator might use. It boils down to which is the most efficient manner of getting the instructions to the computer. I favor a body waldo feedback device for the arms and mid legs, with the computer actually handling the automatic adjustments of the operator movement to the mech movements. This way the operator is able to move naturally and instinctively without having to hunt for options, while the comp is assisting the operator to interpret the commands and feed it to the mech.
I think this started over whether two sticks and a throttle is enough to control a mech.
Maybe not in the day in calm weather. Try it at night while its raining and the damn rope is half off the roof most of the time. You'll be swinging that booty with all your might to get down without broken legs, or without missing the roof totally.
>And while the computer can automatically compensate for all the balancing and 'unconscious' correction the brain does, it still requires the operator to provide instructions on the specific action it needs to take. Again, this comes back to the myriad of options menu I mentioned above, which a operator has to select with the stick and throttle.
I don't want to sound condescending but google HOTAS: A good stick alone has 16 switches as four hat switches alone. In much the same way a computer has shift, control, and ALT modifiers, a good HOTAS also has modifiers. We could even simpify the number of switches and make all these modifiers circumstantial and contextural. The specific tools you need when you would need them.
Having a set of "breaks" could be useful: A button which tries to slow the inertia of the machine. Targetted breaks could mean "stop me before I hit this", which is your breaking applied to a context. You could have an anchor button, which keeps you moving in formation automatically following something (like an automatic match-speed feature on fighters where the throttle then becomes adjusting your speed up or down slightly from that of your opponent, letting you move in formation easily and fall behind someone in a dogfight) which when used with a stationary object could trigger adherence.
Again, I'm just coming up with simple contextural inputs that would be controlled algorhythmically via standard pilot input. You enter an operational bias, not an operational mode. You still HAVE control over your machine but your controls are biased toward specific outcomes.
The other mode is running a mode and letting a user make adjustments to the mode itself, rather than their standard operational movements.
>The battlefield you see isn't going to be a raw output from cameras but it'll pass through a computer, building a 3D point-cloud image with a clear fog of war telling you what you can and can't detect. This image will be populated with images and information not only from you but also AWACs, satelites and all units in the ground including your platoon, drones and soldiers. You'll be able to pull out into third person for superior field of view and have faults and errors visually indicated to you.
It really bothers me how many of your posts are basically science fiction, and yet you act like they're inevitable developments. At least be honest and admit you're being optimistic.
A tank is just a swivelling gun on wheels. Considering that both tanks and mechs have to accomplish the task of 1) moving from place to place and 2) orienting their gun and firing it accurately at an enemy, there may be more commonality between the two than you think.
I agree with your analysis of the option selection, and I do know how to operate a HOTAS, which is why I'm debating against it in a mech. The reason is because the dynamics of flight is actually relatively simple compared to the dynamics inherent on the ground. In the air, you either dive, roll or yaw, while as mentioned on the ground there are many, many contextual based options, and variations of combinations of orders. You will need to really simplify the options into a very few preset groups (go through, stop) which like I mention limits the orders the mech can handle at any one event window. Your explanation handles the contextual options of go through or stop (leaving out the jump option, with is another can of worms since there's jump Over or jump Away), but say you want to follow up with additional variations to the end command, such as bash through and prone or stop and prone, you still need to add that input in as well. Add in a third set of variation like weapon alignment stance, and you get a load of inputs you need to give to the comp while you're sliding up to the house. That's why I mentioned a window to give the commands.
Well the arguments against manual are massive.
Fatigue (how many of you can go full retard with a wii mote for 7 hours solid?)
Pilot muscle manipulation skill
Mechanical differences between pilot and system (if they don't exist, you have an inferior robot. If they do exist, your pilot will find the system highly unintuitive at best and imossible at worst)
Tedium/boredom of repetition
Massive inferiority to ballistics, traction control, maneuvering, thruster control and other computer control
VERY high learning curve (you're looking at raising pilots from children like a goddamn circus act or something)
The only place it makes sense is as something where the machine itself is small and wraps around a wearer not a pilot in a seat. The driver leads the machine, it follows. It still has ALL the flaws above but for highly tedious/constrained problems like lifting and loading things or doing repairs or stuff like that, its golden provided you give a pilot a button that says "move when held" and let them release it when they have it how they want and pass some of the balance off to the computer.
Let me give you a recent example.
Pulling out into third person? Computer vision capable of visually identifying hostile elements? Considering Land Warrior failed abominably, considering we have the tech to highlight friendly forces on an infantry level and yet we don't use it....excuse me if I'm skeptical.
Furthermore my issue is the tone. When discussing future scenarios and technologies that don't presently exist, the proper wording is "should", "ought", "could", not "will", "is going to be". Using the latter instead of the former gives a false impression of legitimacy.
Fatigue is a biggie. Did anyone see Mark Hunt v Faricio Wedum last night? MMA fight. Okay in a nutshell: Mark is the more skilled fighter. He started out as a slugger with a focus on striking without any grappling or ground game to speak of, with a chin of iron -- the man fucking eats punches. Through time, he's grown up and become a much much better fighter, better rounded with a brilliant ground and grappling game who's now a VERY technical fighter (where you put the punch, which you use, when you choose to strike, faux, exchange choices, etc). They call him the Super Samoan because Samoans dominate contact sports: They come from a tiny island with a warrior caste that lived on fish for thousands of years and fought like fuck. Brown people, kinda chubby who never got used to carb heavy diets, looking like hawaiians with tattoos down one of their arms. You'll see them everywhere now I've mentioned it. They're practically super human physically because their cardio (heart/lung performance, oxygen blood exchange) endurance is insane.
Wedum is a grappling specialist. Not anywhere near as technical as a striker but a great grappler. The first guy to beat THE Fedor Emelianenko in mixed martial arts competition. Both are a huge fucking deal, heavy weights and badasses.
Wedum gets to Mexico City and starts altitude training for the fight four months ago. He fits the fuck up for the fight and cuts weight for the weighins like everybody does. Hunt has had some bad news, we think his wife has left him or one of his parents has died. Next, he got the call to the fight -- HIS title-shot with only three weeks to train. He cuts weight because he's bloated and unlike Wedum hasn't got the benefit of altitude training (less air forces your body to be more efficient with oxygen).
Sorry if I'm going off on a tangent. You'll understand in a moment.
If I haven't replied to you yet, could you ask again? The threads getting long and I'm losing track of who's asked what.
Fight kicks off. First round, Hunt's on the aggressive. But he's gassed. His cardio's shitty and its showing. He's technically the better fighter and got some REALLY nice hits. But Wedum, an inferior fighter, is chipping him down. Round 2? Wedum lays into Hunt trying to get him to the ground. Hunt isn't having it. Finally two strikes to the jaw with spinning heel-kicks connect. The second sends him down. Wedum climbs ontop and starts wailing. The referee panics and pulls the two apart, calling it Wedum's fight, even though Hunt has proven time and time again he could go another two rounds.
Cardio. Physical endurance. This shit is a huge deal. We make machines so our pilots don't have to be able to lift a 2 ton cannon and be the only one loading 200 pound shells. Our benefit over other animals is we make better decisions. We're intelligent. We have tools to be strong, to be fast for us.
Why would you go backwards?
Sorry if I come off as pompus or ignorant or if I haven't answered you or that I'm acting like I have all the answers. I've thought about this lot and I have my answers. They might be wrong, they might not be, but its a good solid set of answers I think. I really appreciate you taking the time to read my rambing horse-shit: I know how much of a total fuckwad I can be in speculation threads and how many of you utterly cannot stand me for them.
>Pulling out into third person? Computer vision capable of visually identifying hostile elements? Considering Land Warrior failed abominably, considering we have the tech to highlight friendly forces on an infantry level and yet we don't use it....excuse me if I'm skeptical.
We have software existing today that you can feed 20 photographs to and it will shit out a 3d photograph. It isn't instant and its fault tolorance is terrible but it will be realtime soon. http://www.youtube.com/watch?v=6aU2s85Zw3A
We have cameras today that can now sense depth and know your heart-rate, your height, estimate your weight based on your movement and recognise thermal changes in your body that you can buy. Lots of people even have them in their homes, though not as many as Microsoft would have liked.
We have software which eats medical journals (which would take 200 hours of reading a WEEK for your doctor to stay up to date on) and do really tight assessment on patients. The average (not median or mean) diagnostic rate of an oncologist for lung-cancer is about 60% accurate, requiring consultants and differential diagnosis. Watson is 97% accurate and can do differental on its own. The question is how long you're willing to wait for your answer.
I think its a matter of time before these technologies are converging. Television and computers converged, giving us the graphical user interface. Molecular engineering and computer design had sex and we got integrated micro-circuits.
Hang on, the idea is that the comp would take care of all the automatics, so the operation just has to say make a grab at the object without squeezing pass the feedback to have the mech pick it up. Squeezing pass the feedback crushes it.
The arm controls can be detachable from input, maybe by voice command or a preset when say the operator wants to scratch his nose.
The body of the operator is still locked into the seat, so it's not like he has to move anything other than his arms. The legs are only mobile at the knees, the feet have to be semi stable to operate the pedals. Its not like he will be running in mid air to make the mech run. The part we want the most articulation, the arms, are the only free moving objects, so the stress on the operator is minimized, and stamina drain is reduced. Effectively, he's just sitting in a bucket while waving his arms and curling his legs and tapping his feet. Simple, with the actions matching the human range of motion where its needed, and allowing for complex orders to be input to the comp via the operator's movements.
I get you, and we're hashing out which way is more efficient to operate a mech, so we're still within the sticks debate. And this is 4chan, its enough that you're providing replies with a clear foundation of the logic, and you know your facts. As mentioned, I'm not against your logic, but I am debating against the final conclusion you draw from them, since I have a different point of view and concept of human movement interfacing.
We have the technology to do a lot, but there are so many gaps, so many limiting factors, so many devils-in-the-details that predicting that anything WILL happen is arrogant in the extreme. Ultimately my problem isn't with your opinions, it's in the obnoxious way you phrase them, so I'll duck out for now. I've made my point and any more would be inappropriate.
With software, its a bit trickier. When the patents expire, the processors get faster and the way we write software grows up a little (I mean fuck we're still punching CODE into computers with no visualisation of what the code actually does structurally: programming languages have changed but programming itself as an act has not changed in 30 years) and the ways in which computers work (neural processors particularly being especially promising for execution and quantum very very specifically for optimizing software on the execution level so you can achieve 1:1 asm like results that made the 8800 do full motion video and music -- and the SG-1000 run SNES-like visuals in demoscene wankery) we're going to run into some very interesting stuff.
Unless society comes crashing down catastrophically, this stuff is coming.
Think of the electric motor and what it meant for factories. We started off with a big giant steam engine running an entire factory, driving all the machinery called the cog-wheel. The cog-wheel meant only the very very rich could afford to build a factory. That factories could only build specific things.
The fractional horsepower motor came along. The electric motor. Suddenly, every workstation in the factory has its own motor which can be varied in speed, torque, size and output for different jobs for complex tasks like milling. Metal didn't need to be cast anymore, leading to huge savings and the kinds of things that could be made got way more complex.
Having been in the unit that field tested Land Warrior, I can tell you why it failed: Weight, Bulk and Battery life for the individual soldier's systems not the vehicles.
At the start of field testing, Land Warrior had each member of the unit carrying a disassembled 2004 laptop and 2 heavy spare batteries in order for the computer to last roughly 1 day. This was in addition to all normal mechanized infantry gear.
A hardened modern smart phone with an expanded battery pack could run everything that the laptop was in 2004.
Not to mention, hooking up a 3d printer to a CT scanner allows for replication of the internal skeleton or a frame for organ tissue growth in current medical research. So we'll be 3d printing organs using cells aka Fifth Element. Science fiction is now. People though having handheld computers with touch interfaces were fictional as little as 30 years ago.
The same then happened with computers. We start off with big giant mainframes, big things that fill entire floors of buildings. People have these dumb terminals. They log in and they get to their stuff. They get the illusion that they have total access to the machine because its lots of power spread across lots of users.
Then Motorola and Xerox come along with their concepts for the personal computer, teaming up with Bell Labs who create what we know today as UNIX, the backbone of all operating systems that's special because it deals with things as files and directories and seperates live execution memory from long-term document storage memory but if you run out of execution memory it can make a "page file" on the diskette or harddrive of your machine letting you multitask. The people running the companies "don't get it". IBM loses out because they didn't get it and want to keep building mainframes. The 8800 hits the world running MSDOS. Apple follow up by stealing Bell and Xerox' concepts for the UI. Microsoft copies Apple for 3.1 then goes back and looks at Xerox and Bell for Windows 95. SGI show up prior to Windows 95. Suddenly networking is massive.
And networking is the next thing which is going to become fractional horsepower.
Then its going to be the ease of entry into programming, the same way user interfaces made general use of a computer easy.
In 50 years, you can code if you really want to but you'll probably suggest actions and then correct when the computer misbehaves if your input isn't highly constrained. You will teach the computer to do the thing. It will learn what should be done and what shouldn't and eventually do it properly.
How do I know this? Google's already doing it with their self-driving car which has something like 4 million miles on the clock in all weather conditions (even hellish ones) and only three collisions -- which were ALL caused by human error on the part of other drivers or a human driver overriding the Google Car.
Eventually it'll be energy. When the engineering issues are solved, building a thorium reactor the size of your head that can power your street or an electrolytic cell which uses nanocapacitor glass with to form half hypercapacitor half battery will be something kids will do following instructions from books with their 3D printers which will democratize manufacture and be the fractional reserve factory.
It will happen eventually. Or legislation will stop you. Actually, that's guarenteed. There are hundreds of organisations in the US alone which step in to stop you from mucking about with electromagnetism in unusual ways, which tell you not to write certain kinds of program, to not breed plants or do certain kinds of chemestry or to manufacture and sell certain fabrics or plastics. I'm not a conspiracy theorist: These are just simple facts of life today in the modern world. Given how disruptive a lot of this would be, I think it'll have to be sanctioned and you'll need permits to do any of it. Not because of anything to do with the ruling class but because you could very easily hurt yourself with this stuff.
Sorry if I'm not answering you individually. I'm looking for sentiments and trying to come up with answers (or really, glorified opinions built on evidence) to try and satisfy you. I would answer you all directly in constrained ways but 4chan's post cooldown and post-count (really, half-chan? 8 goes to 50,000 characters, not 2000: I post entire logs on my own private board) make it really really hard.
Do any of you use IRC or would you like to do a back and forth over email? I actually learned almost all of this for you guys. I didn't come to these conclusions on my own: I'm just a big sponge that's been here half a decade mostly lurking and soaking up YOUR discussions.
Again, I'm really enjoying this back and forth.
The problem with democratization of anything is the same problem democracy itself has run into: qualification. In an ideal world, you wouldn't have to worry about rednecks or idiots, but neuroaugmentation is probably even further off than all the other technologies we've discussed so far.
The main problem with that view is one of the major justifications given for mecha is that they can do things tracked and wheeled vehicles can't, thus requiring a much more complex go from A to B system
I think the next big scifi thing is going to be brain interfaces. I know that's a very blanketing term but I have some pretty specific ideas about what it would be like. While current systems look for specific events in the brain and do piss poor noise compensation, future systems will be very error corrective. In addition, I don't think they'll be driven by thoughts of movement or memories or smells or flavors like modern systems are which require huge amounts of discipline.
I think they'll be driven by intention. Like a terminal line composed of some simple language objects. Programming the computer with desires.
Say for example.. You're in a big heavy 4 ton robot designed for lifting cargo. You want to pick up the spoon. This overly sophisticated forklift then assesses the spoon and looks into ways of acheving this optimally, doing its own live simulation and problem solving instead of just doing highly constrained backs and forths within programming boundries.
Egg and spoon race?
I want to run here as fast as I can. But I don't want to drop the egg. You're not focusing on driving every joint. The computer is doing that. You don't focus on not dropping the egg. The computer does that.
But you could in the other hand have a rifle. Your elbow and shoulder joint playing recoil absorbsion. Again, you decide you want to shoot something. It throws up a UI. For reasons of accountability, you do HAVE to give it physical designation that it can, telling it weapons free then holding the trigger. It crunches trajectories and ballistics. Boom, bulls-eye. You didn't even have to think about it unless you didn't want to follow the default firing mode (centre of mass, as per military doctrine).
It's been fun indeed, but its 4am here and I need to get to work at 8. See you guys went I get back if the thread's still alive.
We're a very, very long way from reading intentions. Last I checked the lab was still in the 'read electical impulses in certain areas and activate A or B' point.
But you can also layer in... I'd say like a limit or a ... What's the word? An action you want to avoid. Like a boundry or a loop edge. Ah, constraint recognition!
You tell it what you want to happen. You tell it what you DON'T want to happen. It works out the best way to fit in the middle of all this. It also has an idea of some general directives like friendly fire so no nasty YF21 type stuff: You can't override that unless you do it with your hands and thus you become legally accountable for your actions (which again, this physical switch accountability I think is huge since we don't always have total control over all of our thoughts).
Want to net a basket? You think about it. It knows you've seen people doing it. Sees if its already baked a solution for this problem. If not? Comes up with a solution (either locally or remotely on a big server beaming the solution back) and then it acts. Slight lag the first time but it can now do it regardless of your gait, velocity or distance.
The problem comes when you start piling up these kinds or problems. You will need to also issue what I'd call a resolution of action: how much control you have verses how much you let the computer have. If you want to occupy a fixed pose, that's all you. If you want to intercept a blade with your own without being knocked off balance in such a way where in your next two moves you're not left open on the opposing side (think fencing, which is both constrained and unconstrained) that's a solid mixture of both your experience and the computer's own guesswork.
The question then is providing neural feedback, to make this feel "natural". If the computer thinks it will make a better solution, it might even just be polling your experiences in combat (which you've either had naturally or its written back to you specifically to improve your learning speed and reduce the learning curve), ignoring you all together based on some higher mission directive.
With psychoactive compounds or inducing these states, we could have you the pilot issued with lots of scenarios and answering all of them so when they happen (two moved down in your fencing game) you've already made your decisionin your OODA-loop, making your response an immediate snap. This lets you get inside your enemy's OODA loop (orient, observe, decide, act, fucking fascinating stuff go read Boyd's work you cunts <3) and therefor start second-guessing your opponent(s). People have been experimenting with psyllacibin for decades but you're not 100% lucid or ... What's the word? Cognitively responsible? But then its like we have disassociatives which make you psychologically view yourself from a third person perspective, with yourself as the other. This allows you to do things you normally wouldn't and in some cases can allow you to be painfully objective with yourself in ways your ego ordinarily wouldn't let you be. Unfortunately they're ordinarily very numbing, making you ... well. Drunk. Tired. Exhausted. Mellow. Non-operational.
I was lucky enough to get some Pregablyn/Lyrica about a year ago when I was being evicted and my place was a dump because someone had stolen my bin and my toilet wouldn't flush, the house was full of trashbags and I was in a pretty bad place emotionally (my mother and my best friend had just died and I was basically a drunk) so I tried ignoring it, ending up with some hellish conditions. Worse, we had someone stay over who had very questionable hygine to help with the rent. Cleaning crews gave an estimate of about a week to clean it with four people. I took Lyrica. Two days of no sleep wizz by. Everything is pristine, my clothes are filthy and my hands are raw. My husband tells me I bleached the walls in the bathroom where the damp had set in until my hands bled. I fixed the piping. I painted the walls. I fixed the electrics getting shocked on six occasions. It feels now like I watched someone else doing it. I remember I'd tell myself to do things, then I fell asleep and I'd still be doing them when I woke up, following very simple tasks.
Psychoactive compounds are very very powerful and it drives me nuts that we haven't explored them further -- not only to cure conditions like PTSD with MDMA or alcoholism/addiction with LSD/LSA but productivity, like how LSD is basically responsible for the idea of converging the computer technologies I mentioned in my last post during the 60's. Aside from maybe doing MDMA once in a theraputic setting during trials at the hospital, LSD with my husband and drinking in my early teens, I don't really do much of anything. I can't remember the last time I took anything: those days are long behind me.
I'd do the wide-eyed "wow, can you imagine?" thing but I'd imagine you're already thinking of potentials. Is it ethical? Probably not. Given our coming ability to fuck with neuroplasticity and even instill action-addictions in mice (grooming habits, mazes, etc)
Anyway. In theory, we can get even weirder than this eventually, assuming we hit Zone of the Enders like singularities in technology. Brain scan. Run in a computer faster than the brain. In theory, we can trade longevity of a simulation for accuracy and speed. On paper this means you'd have to do new scans every few minutes, compiling into a new software model. Picture you're about to do something. Your machine does it for you seconds prior. It updates you with what happened, writing the reward/learning process back to you.
I'd imagine it'd be incredibly draining to do the back and forth learning but without it, pilots don't improve and would be consistent. It'd really fuck up dopamine and seratonin regulation and potentially even oxytocin if engineers decide to take advantage of our peer bonding.
Actually, peer bonding has been taken advantage of. Remember the spartans, the guys in 300? Giantic faggots. Man love everywhere, so they could try and impress their partners and fight harder for eachother and impress eachother physically (meaning they stayed fit). Convenient how the movie pushed that aside.
The modern macho image of bravado owes a lot to ancient Greece. If you wonder if the guys on your football team (lol b& why the fuck are you here if you're under 18) seem a bit gay, the simple answer is yes: It creates better peer bonding, but they'd never admit it in a million years (ironically out of fear for being seen as weak).
When people survive scary shit together, they release oxytocin, that same peer bonding hormone. That's what "feeling in love is". Its why you like petting your cat and its why your cat likes being pet. Its why people kiss and touch and hold. And its why guys who take their dates to see scary movies instead of romcoms actually get laid.
Yeah but provided we don't wipe ourselves out, its basically inevitable provided there's interest in it. It'd be kind of funny if the reason we ended up getting into it as because greedy marketers wanted to know how to sell shit to people better. Then it ended up in military weapons. Ha.
Sorry if I'm rambling a bit or getting too anecdotal or veering off. Peer-bonding in physical practice could enable some pretty legendary feats. Or even in a narrative could be really fascinating to exploit, to mess with pilots and make them like eachother.
I think inevitably its exploitation is going to be something every military does within the next 30 to 100 years -- whether its sexual or not. Yes: The army is going to become a hugbox on the inside but probably a ferocious beast on the inside. I mean shit, we have "crime of passion" for a reason: If you can leverage that part of the brain, that's a really scary weapon. I can't help but wonder if the CIA experimented with it within MK Ultra or some of its other black projects. Kennedy, monster that he was let some really fascinating (and unethical) research happen.
Another good justification for mecha on paper is that they can correct their balance in ways something like a tank or an aircraft can't and they can point entire groups of systems in different directions.
Imagine for a moment a robot is just a box with two (or four) engine pylons that end in landing gear and two turrets and a sensor pod. Even on a very abstract level, it can already do a shit ton of stuff a plane or a tank can't, especially if you solve the energy problems and material problems involved.
For one, it can land harder than a harrier on really unstable terrain because its landing gear has extra joints. It can also lean into corners if its skimming the ground like a hovercraft/rollerblader/ice hockey player performing some pretty majestic feats of movement.
Second, it can fire at two independent targets. With weapons that are effective getting smaller and lighter, this is going to become a pretty big deal early on: You don't need a huge gun to deal the same hurt because you have smarter bullets. Better yet, retooling this sucker for a different mission doesn't require any kind of special hanger gear: It can pick up and mount its own weapons, theoretically even in the thick of it.
Now unfortunately (this is opinion mostly) nothing with this joints as I say so often here is ever EVER going to tank hits. Its too fragile and too complex. But there are lots of ways to survive an enemy.
You can not be seen. You can see and shoot him before he shoots you. You can make yourself not in the trajectory of the incoming shot. You can put a thing between yourself and the shot.
In practice, this could be something like burst maneuvering. AC4's quickboost but not quite so extreme. You get some nice lateral motion and the shot passes you and you're already returning fire or moving into cover.
You can deal with the acceleration and the deceleration without wobbling off your own feet because you HAVE feet. Quickboost in theory isn't all that unrealistic if you look at something like Orion and the use of pusher-plates which even with a barrel sized example had 15 Gs of acceleration for about 4 seconds with a non-nuclear example in testing. Pusher plates is a bit silly but there might be some way to do it with thrust. idk. Its a starting point, I haven't thought on it much.
Actually, you know one thing that gets on my nerves? The robots arrive all very advanced, spiffy and developed... And other weapons haven't changed beyond cosmetic changes. Tanks are still tanks. Planes are still planes. There's no convergence of these technological benefits to those platforms. Mecha aren't going to be the be and end all unless they cam some how do something those platforms can't. Personally I think that'll come when we start making really small mecha that can go places tanks and planes can't that are light enough to be able to fly by themselves quite well. Given that ability to swing entire engines, they'd out-dogfight fighters the same way the Harrier made a mockery of the F15 outside of BVR standoff engagements during wargames.
Anyway, I'm spent. If you'd like to chat, drop a server and a username on IRC or an email address and I'm more than happy to ramble aimlessly slothing pathetic twaddle like this for hours if you're happy to listen.
I'll be checking up periodically for anything like that if I can't sleep so expect a response within the next few hours.
OP, keep writing. A writer writes. I hope some of this has been helpful to you.
I honestly think the legged tanks from Gungriffon Blaze will be the future of tanks. This is because they can walk and have wheels on the bottom of the legs. The main problem is making legs durable enough and the ability to have a wheel on the bottom of each leg that allows the tank to go as fast as a tank with treads.
You know, bendable treads with feet on the end, as opposed to legs with wheels on the end.
An item of fantasy as far as I know, but intuitively more satisfying than Tachikoma-style wheels.
I'm not seeing it. Could you knock up a crude illustration side on in mspaint or something?
Do you mean like a curved flap on a hinge that comes down to become the feet that normally sits ontop of the tread instead of the bottom?
Something like this, I guess.
>>11685236 >>11685338 >>11685361
I tried, anons, I tried. Now I just want my time reading all this crap back. This feels like discussing economy with my wife.
At least with an AT, the engines seem to be situated squarely in the giant feet, with not much space for anything at all in the cockpit.
Well that is the problem that could hamper it but I think in the future speed, maneuverability and damage output are going to be the most important things in a tank. Armor will be less important.
For skating mechs, I'd imagine they'd first pick up speed and momentum by running, and then start skating both for increased manouverability and to keep the thing steadier for firing on the move than if it was running.
I think the best would be if the mech could sort of "slide" in any direction, not locked to just slaloming forwards or backwards. A mech's big advantage should be being able to quickly change directions on the go, I think.
The reason we have three people in a tank is because if two get hurt, the third can drive them home.
You actually have a commander, a driver and a loader. The commander is usually the gunner. The loader is now beaten by technology for the most part. If the commander could both drive and aim some how through better situational awareness and only designate while the tank takes the shot, you'd only need the commander.
A lion doesn't have three brains, even when in three minds about something.
Pointless and totally unrelated cereal packet quotations aside, I heard them cackling like an hour ago mang.
Keep in mind though, my timeframe is the next 50 years, not the next 20 or even 10 years.
Hell in 50 years, there won't BE people in tanks unless EPFC generators get really popular.
Usually the tank crew positions are commander, gunner, driver
Obviously a loader if you don't have an autoloader.
Autoloaders are also usually slower than human loaders, at least for the first few minutes. For sustained fire autoloaders do beat out humans
>Autoloaders are also usually slower than human loaders, at least for the first few minutes. For sustained fire autoloaders do beat out humans
I'm pretty surprised we haven't managed to deal with this particular engineering challenge yet, but I suppose the military-industrial complex would rather spend money on fighter jets or something.
This. Astronauts are also used to learning newly-developed new tech in very short times.
I'd say engineers with astronaut training would be the best you can get to pilot the prototypes during initial development of units.
First of all they'll make customs of actally weapons, like tanks, planes and power suits for soldiers.
Eventually the faster mechs and bigger will take the lead, mobile armors as tanks, ships and planes, and small MS flag-like to air fights
Its thought to in the future be a way of making an destructive ECM/EMP warhead about the size of a grenade that could be airburst near a target or used as a mine that would cook microprocessors.
>the best ones often had histories as race-car drivers or motor speed-enthusiasts
Not sure about other air forces, but I know that for the RAF a lot of their first recruits were transfers from the Cavalry. Whether Cavalrymen made good pilots, I have no idea.
Less of them making good pilots and more that there were a lot of Cavalrymen finding themselves without jobs as time marched on.
Despite what popular histories will tell you, the Armies of the world didn't march to WWI with any delusion that Cavalry would play a major role, so everyone cut back on their Cavalry numbers. And, since you've got a lot of already disciplined troops without jobs, might as well put them to work on a new steed.
>A modern battlefield tank needs at least a crew of 3 to efficiently operate the machine.
There are two-manner MBTs on the drawing boards right now, mark my words.
Driver and Commander/Gunner
That's all you need if you can automate the fuck outta everything else.