So I have in mind a plot point in mind where my protagonist is elected to pilot a giant humanoid robot with flight capabilities - Kinda like a mobile suit from the Mobile Suit Gundam franchise. To pilot this robot, the pilot docks a specialised car into the chest, whose interior transforms to form the cockpit of the robot, with a HUD of the robot's 360° peripheral vision displayed on all of the windshields. This robot is to be used on land, in the air and in space, using its legs alongside thrusters on its calves and upper back to move around, and its hands to manipulate tools and weapons, such as it's characteristic Sonic Sabre™ (which needs a better name).

My character, however, has only ever driven the car, not the robot. And when the time comes for him to do the latter, he has had no training or briefing other than a brief flick through the rather rudimentary 'Beginner's Guide to the Gamma Bull'. But, when he gets going and has a minute or two of practice, he finds piloting the robot to be a piece of cake.

But for that to be the case, the control method for the machine must be really intuitive. Again, my character has only ever driven the car, which is a push-and-go automatic. How exactly could the car's interior transform to make piloting the robot so easy? What additional functions could be made available?

Some things I've had in mind;

  • Avatar style - The mech suits in James Cameron's Avatar have their pilots stand upright in a harness, using the motion of their arms and I presume their legs to make the robot itself move around. This would be perfect and undoubtedly very intuitive, but I'm not sure how well the interior of a car could accommodate this, even after transforming slightly - Remember, the pilot still has to use all of the cars windshields for sight.

  • Mostly automated - The other idea I've had is that the controls remain somewhat as simplistic as those of the car, meaning that a lot of the robot's motion would instead be automated - Pushing the gas pedal, for example, would make the robot run or walk, but the robot itself would be doing all the calculations regarding balance and footfall. The problems I encounter with this concept, however, is how the pilot would control things like arm movements and flight - Could the steering wheel split into two joysticks? Could the PRNDL stick be a joystick?

  • Mental link - As somewhat of an addendum to the previous idea, the pilot could have some manual control through physical controls, but other aspects of the machine's motion would be dictated through a mental link, a la the Alaya-Vijnana from Mobile Suit Gundam: Iron-Blooded Orphans. However, would the mind be able to cope with controlling its own motions while also sending motor orders to body parts it doesn't even possess? Would the unstable mind of someone whose never controlled a giant robot before cope with such stress?

"Why a car?"

Honestly? Rule of cool. I consider it a pretty core part of this system's concept and feel too attached to ditch it. Sorry ¯\_(ツ)_/¯

"What's the advantage of a giant robot?"

See the previous question.

  • 5
    $\begingroup$ Why the car at all? Seems to me, it only makes stuff needlessly complicated instead of bringing any advantage, so why would someone build it this was? In a specialized command center, you could, for example, project a real 360° vision, without any problems with the form of the windshields. I sugest: Dock the car in the chest, if needed, and then take the automated elevator to the command center. $\endgroup$ Jun 22, 2017 at 11:12
  • 4
    $\begingroup$ Step 1: Watch Pacific Rim. Step 2: There is no step 2. $\endgroup$
    – Werrf
    Jun 22, 2017 at 16:59
  • 3
    $\begingroup$ @Werrf MegasXLR Is also an option. It seems specially appropriate, with the car theme. $\endgroup$
    – Mermaker
    Jun 22, 2017 at 17:46
  • 1
    $\begingroup$ As TSar mentioned, piloting a giant robot through a car is literally the plot of MegasXLR $\endgroup$
    – alexgbelov
    Jun 22, 2017 at 21:11
  • 1
    $\begingroup$ This also made me think of Megas XLR. I found a free version of the first episode for you to see for ideas. He pilots the robot from a car that is filled with various video game controllers, and he's so pro at video games that (with a slight learning curve) he became nearly unstoppable. At least, until the show was canceled :( $\endgroup$
    – Cody
    Jun 22, 2017 at 21:14

11 Answers 11


I'm going to go completely the other way from the existing answers, rather than a neural link or sensor suit to translate the drivers body movements to control the robot, make the giant robot have an AI system that controls itself with only command input from the pilot.

The automated system would handle everything from walking, running, launching missiles, to sword fighting. It would provide situation reports and respond to verbal or control button commands from the pilot; The AI could highlight identified threats and recommend actions and respond when the pilot says, "move to position 3 on the map", or, "attack the target I tagged on the radar using the main cannon," and the AI would carry out the orders.

This would eliminate the main limitation of having the pilots body control the robot, namely that the robot can move in ways and do things that the human body simply cannot do. It's joints can move in different ways and faster than any human, no human has built in thrusters or machine guns, or transforms into a jet. Workarounds for a trained pilot could happen and with extensive training the pilot could use a body sensor input device quite effectively (the AI could also teach your novice how to use it so that the pilot gets better and more control as the story develops), but an untrained person wouldn't initially know how to do anything that a human cannot and is likely to flex their wrist or blink their eyes in the wrong way and accidently launch missiles or initiate the self destruct.

  • $\begingroup$ So, "Drones are better"? $\endgroup$ Jun 23, 2017 at 7:23
  • 1
    $\begingroup$ 'Trust me. Drone better' $\endgroup$
    – Joe Bloggs
    Jun 24, 2017 at 10:33

So, you have a humanoid robot. And you want a human to control it as instictively as possible.

Most humans can control the motion of their own bodies quite well, so let's work from there.

You could strap the pilot into some sort of contraption where he is held around the waist and can move all limbs freely. thus he could simply walk when he wants the robot to walk, extend a hand where needed, and generally just move the way he wants the robot to move. for any extras, like activating, adjusting and directing thrusters, use a joystick.

This way you translate almost all actions direrctly and in the most intuitive way available, reducing specialized control to an absolute minimum. Add to this a bunch of automated and multiply redundant failsafing and stabilizing systems to prevent oversteering and loss of balance, and your setup should be ready to go.

  • 1
    $\begingroup$ That would do it - something like a motion capture suit. Add in VR goggles, and you can become a giant. $\endgroup$
    – tj1000
    Jun 22, 2017 at 12:41
  • 6
    $\begingroup$ but, you limit the robot to the human capabilities. If the pilot gets tired of running, the robot gets tired. Cut out the pilot having to actually move - interface directly with his brain, and you get rid of that problem. $\endgroup$
    – user11864
    Jun 22, 2017 at 13:13
  • $\begingroup$ @AgapwIesu True, that would be better, but not getting "tired" wasn't a prerequisite indicated by the OP. $\endgroup$
    – Neil
    Jun 22, 2017 at 13:46
  • 1
    $\begingroup$ It would be pretty cool to use servo-motors to let the pilot feel some degree of resistance, I believe it would also help them with the sense of balance. $\endgroup$ Jun 22, 2017 at 17:53
  • 2
    $\begingroup$ ...and then the pilot gets an itch on their nose, and your giant robot rocket-punches itself in the face. $\endgroup$
    – Salda007
    Jun 23, 2017 at 7:13

My first thought was what Burki mentions - biofeedback kind of thing. Read "Starship Troopers" and that is a great example of just this sort of thing. But the problem with that is that you then end up limiting the robot to the physical limitations of the pilot. If the pilot gets tired of running, and can't run anymore, the robot is stuck. Plus, how do you handle other functions like flight and firing the robot's eye-lasers? So, I would go with something different.

If I remember correctly, we have something called mirror-neurons. When we imagine ourselves running, the same neurons fire in our brain that would fire when we actually run, but mirror-neurons keep the signal from actually going out to our limbs.

You could have an interface (helmet? telepathic? brain implant?) a-la Avatar movie, which allows the pilot to "imagine" the actions he/she wants and the robot would do them. The interface can induce an imagine/day-dream state that allows the pilot to do everything they want to do, in their mind, without it actually translating into any motion. This allows for total control, and if the interface is sophisticated enough, control could even include flight and use of other non-human capabilities the robot may have - pilot, day-dream away.


A Xbox/Playstation Controller coupled with game-like software and Z Targeting.

I mean, really, who hasn't had a blast in Zelda and made Link do all sort of things using nothing but a regular controller?

If you have a good AI taking care of the hard stuff, like balance, know movements, personal space and pattern recognition, a standard controller with a smart configuration can very well provide a huge array of functionalities. Also, by making it so similar to a videogame the protagonist already knows and plays, you can make it even more believable.

Of course, you wouldn't be able to provide complete control for the pilot for the more detailed stuff, but for gross combat and movement, that should be more than enough.

An interesting note - today, some military vehicles are actually operated by Xbox/Playstation-like joysticks. No, I'm not making this up.

  • $\begingroup$ The problem here is: You are limited to a handful of action patterns. which the AI actually isn't optimized to exploit, well isn't even coded to care about in most cases, since otherwise it would be no fun for the player. But in military setting, it takes like 2 or 3 battles to realize that the robot only has (lets say...) 37 different move and action patterns. So you jsut had to develop a strategy none of thees patterns would be able to handle and you are done with that OP mecha. You definitely don't want to go for that in an military setting. $\endgroup$
    – Zaibis
    Jun 23, 2017 at 6:11

A mental link is not much of a problem

You ask, "would the mind be able to cope with controlling its own motions while also sending motor orders to body parts it doesn't even possess?" and the answer is "probably quite well, in fact." The human mind is actually incredibly adept at adopting tools into its "body plan" and a number of experiments have shown how relatively easily it is to trick the mind into thinking that things that aren't part of its body actually are. Just think of the phantom limb phenomenon. You may think that that is slightly different because it's a limb that the mind feels like it 'should' have, but further experiments have shown that you can convince the mind that completely foreign objects are part of it.

Further, adapting extensions to our body plans is commonplace. When you use a tool such as a wrench or grabber, it subconsciously becomes part of our body plans, and we sort of feel like we have an extended reach. We're not getting nearly as much sensory information from the tool, just a bit of vibration and resistance, but we still adapt it into our body plan to feel as though it is part of us in a sense. We work by "feeling" things with our tools. Even when we're driving cars we have a sense of adapting the car to our body plan, using information we get from sight and touch from the vibrations moving the vehicle to sense how close we are to other objects, how straight we are on the road, etc.

Experiments in inducing out of body experiences have shown that where we feel like our consciousness is is highly susceptible to suggestion. There's nothing paranormal or supernatural about it - it's just a trick of the subconscious. But people have been able to be convinced that they are actually inside dolls that are larger or smaller than them,, for example.

You see where I'm going with this: If the mental link is providing sensory input from the mech's body, and the pilot's body is not receiving much variance in its sensory input (he's restrained in a comfortable seat), then the pilot will pretty readily feel like he is the mech. And it does help that the mech already has the same basic body plan of the pilot - it is easier to trick someone into thinking they're a doll than a car.

(Likely, this trick of adapting tools into body plans evolved alongside tool use, because it makes us feel much more natural when using them. Any master craftsman who has told you that they treat their tools as an extension of their bodies is not just being poetic, this is a real thing. I wouldn't be surprised if other tool-using animals have a similar cognitive experience, but it's an abstract question to ask a bonobo. Either way this is an aside since your protagonist is a human, presumably.)


I agree with @Michael that some kind of neural interface would likely provide both the best performance (least delay and most accurate implementations of the pilots intentions). But what if we don't have that technology available? As an abstraction of the "tools as extensions of the body" idea, you can see similar results with more generic controls; just look at the majority of video games. Want to swing your sword? Press X. Need to circle-strafe the enemy? Use the two thumb-sticks while squeezing the trigger! All you need is a gamer for a pilot.

I didn't see any mention in the OP of the pilots other skills, but the statement "My character, however, has only ever driven the car" could suggest that he built, or at least modified, the car himself. In that case, he could install whatever combination of controllers/joysticks he's most comfortable with. A few short hours of pilot practice combined with system training (when I hit X, swing the sword) and your mech is up and running/gunning/flying. Related? Wikipedia: Megas XLR

  • $\begingroup$ Welcome to WorldBuilding.SE Shadow! If you have a moment please take the tour and visit the help center to learn more about the site. Have fun! $\endgroup$
    – Secespitus
    Jun 23, 2017 at 7:09

Complete mental link only.

Since Your mech is obviously inspired by Daimos, I suggest You follow the pattern set there to its ultimate version. The pilot's suit/chair/helmet hijacks the impulses from the spine and feeds them back there. The pilots looses the non-autonomous control of his body completely, and the mech becomes HIS body. Note that the 'non-autonomous' part is very important, else his heart would stop. Anyway, the pilot feels what the mech feels, and moves it as his own body because it IS his own body now. His biological body is completely inert and would probably have to be kept very tight by the pilots chair.

You cannot get more instinctive from that, and as a bonus this is a solution which is realistic in terms of our own world's physics and biology. A further bonus is in that this solution offers a possible explanation for the origin of the 'glowing eyes' which are one of the staples of many robot series, including Gundam franchise.

  • $\begingroup$ I hadn't even heard of Daimos until you just mentioned it :P $\endgroup$
    – McOwen
    Jun 23, 2017 at 8:48
  • $\begingroup$ You shouldn't add the part about this only being a comment. That's a very specific reason for deleting a post as "Not an answer - this is a comment". Remove this part and make sure that your answer can stand on its own. It's okay to make an answer that draws inspiration from other answers (it would be appropriate to link them in the text) if the new answer summarizes the most important points and adds important details that were not discussed previously. $\endgroup$
    – Secespitus
    Jun 23, 2017 at 8:53
  • $\begingroup$ @Secespitus - You now have me very confused. How is adding details to the previous answers allow it to stand as a new answer? I, quite honestly, cannot imagine how this is supposed to look like. Can You give an example? The more I try to participate in stack exchange, the less i understand it... :( $\endgroup$
    – Maciej
    Jun 23, 2017 at 10:23
  • $\begingroup$ @McOwen I do not know if it is the first anime ever to use this system, but definitely one of the earliest. And come on, car as a pilot cockpit? Pilot who never piloted a mecha before real combat? 100% Daimos :D Honestly, check it out - tough I dont think trying to watch every episode is a good idea. More like... a few here and there... to get the feeling for what history of mecha anime looks like. $\endgroup$
    – Maciej
    Jun 23, 2017 at 10:26
  • $\begingroup$ There's also MegasXLR. $\endgroup$
    – jo1storm
    Jun 23, 2017 at 11:13

Highly sophisticated AI combined with a neural link. (note, the AI does not have to be sentient. It is likely just a very powerful computer)

The neural link operates at a high level. It intercepts the "desires" of the user (eg walking forwards), forwards this to the AI which then, with regards to the robot's situation, actually makes the robot move.

The neural link also has another function. It blocks commands from the body, preventing the user from actually walking around and injuring themselves inside the cockpit. I suggest you attach the neural link at the neck so that he can still smile and talk. Or maybe somewhere down his spine if he should be able to move things with his hand.

With an advanced enough AI you could forgo the neural link, but in doing so you would lose significant control. A steering wheel and gas pedal could easily be translated into the robot walking, but then you have no way to tell it to kick.

On a side note - real life robot boxing:



The simplest way is a suspended full feedback body suit.

You see what the robot sees and you feel what the robot feels and the robot responds to your body's movements. You move your arm and the robot's arm moves the same. See Haptic body suit for VR

Weapons target based on what you are looking at and everything else is voice command or automated.


As machine learning & AI becomes more and more robust, I've learned that the idea of having a human control something robotic is the fastest way to slow down the robot.

Machines run tons of calculations per second. If you have a robot that has a stabilization method in order to walk bipedally, then it's already running tons of calculations to equalize it's legs to stabilize itself.

That processing can get extended to aiming and firing weapons with pin-point accuracy in split-second timing (aim-bots, basically). To react to situations with dodges and maneuvers in split second timing, etc.

Adding a human to the sitation..

1) you now have sometihng soft and squishy the robot has to worry about protecting

2) that soft squishy thing has a brain and body that can only withstand so much g-force, which limits the reaction speed of the robot

3) if the robot is waiting for the human to react.. human reactions, regardless of how fast, are far slower then machine / computer reaction..

Basically, humans are good at making decisions

Machines are good at enacting will.

So, I would second the notion that the human "pilot" would actually be a co-pilot just along for the ride. And, the robot is there as an exoskeleton to protect the human while giving them a lift to some place the human needs to go to because "human decision required".

With enough technology, robots are going to be able to react, fight, etc, far better then humans, so funneling all of that through a human brain just slows down the reactions and capabilities exponentially.

Machine learning is even getting good at "reading" things beyond the level humans can.

EG: a human eye only has so much color / light processing capability to detect the differences in things. Meanwhile, a computer can detect the differences in colors at the lumens scale (such miniscule differences in color that the human eye can't detect it). This is how machine learning algorithms are able to spot and highlight things in photos that the human eye normally wouldn't.

After I got into data science and machine learning, I looked at sci fi a totally different way.

When I see shows where the humans say "give me manual control!" to insinuate the human will be far better at accomplishing a goal rather then the computer that has split-second reaction times and far better sensors.. it makes me laugh.

My idea of what pilotable robots also changed. The robots would pilot themselves. We're just along for the ride while the robot acts like a walking APC protecting us inside.


This is gonna come off as a little rude or dismissive, but I don't mean it that way AT ALL!

And what I want to say is: if you're going to do something so nonsensical, why bother thinking the rest of it through? Why spend so much time trying to make it realistic when you're doing everything else around it in such an unrealistic fashion? If you're doing stuff just for "cool-factor", then just swing for the fences. Make up something similarly non-sensical and don't worry for a moment whether it's realistic or not!


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .