Watch: Pibot, a robotic pilot to fly planes that were designed to be operated by humans


#1

[Permalink]


#2

Do they not have autopilot in Korea?


#3

Sure, it seems to be competent. But it’ll need a lot more work on the packaging and a friendlier name before it’s ready to puff itself up and stand up to the competition.


#4

Finally, a pilot we can be sure won’t be raptured away mid-flight!


#5

Is visual data the best we can do for landing and take-off? Shouldn’t this thing have an array of avionics for this instead?


#6

Planes already fly themselves for most of the trip. Eventually we will do away with pilots all together. Military jets will be even more streamlined with out a canopy jutting out.


#7

Not in small planes which aren’t commercial airliners, no. Just like everywhere else.

The problem I have is they’re clearly trying to give it a human form, which is needlessly inefficient. In the same volume of space you could build a much more mechanically reliable robot with less points of failure and better control, with the only downside being that instead of a vaguely-human-in-appearance machine flying the machine, you have a machine that looks like what it is: an industrial tool.


#8

Immigrate robots gonna took yer jerbs!

Thanks but I am already Jewish.


#9

Yes, but only for old people.

This post sponsored by Smirnoff’s Home for Elderly Memes.


#10

I did not know this.


#11

I think they should go all out and make it look even more like a skeleton.


#12

“…give it a human form, which is needlessly inefficient.”
Thank you for stating this. I say it all the time when seeing technical “human-like” demos for robots. Just about every situation (climbing stairs, running, ocular parsing, etc) can be accomplished much easier by a robot that isn’t built in the human form factor (Hence why the fastest runners on the planet are not upright, for example).

Instead of creating what is essentially an autonomous replacement for the human controller, why not spend the time/money on removing the human-form components entirely and replacing them with sensors/controllers fitted to the task, rather than fitted to the limitations of the human form and its sensory inputs?


#13

Upon reflection, the idea, I think, is that planes aren’t standard - they all have the same basic controls, but the layouts can be wildly different in terms of physical space.

And while you can account for that by specifically programming the software to accomodate for the differences (upload the flight software for this or that model of airplane, for example), if the hardware itself can’t adapt to all the different configurations, it won’t matter.

Now, that still might not necessitate the degree to which they’re giving it a human form factor, but it does mean they can’t just use a design that is a custom fit for a single cockpit design, and have to have something that can manipulate controls in different places, or of differing sizes, or of differing sensitivies.


#14

Anyone else remember that old Looney Tunes episode “Hare Lift?” When Bugs activates the automatic pilot during an emergency, one of these gizmos comes out of a storage compartment, grabs a parachute and high-tails it out of the doomed plane.


#15

Well if you buy an autopilot you can have or buy a small plane with one built in, it makes staying at altitude and heading much easier and cross country much more fun.


#16

Valid… but my point was, for all but the smallest planes which still use some sort of direct cable controls, it would seem more efficient to just connect a computer which controls hydros/servos, etc. in place of the control mechanisms that are meant to interpret human tactile input. In other words, instead of replacing the human that interacts with control inputs meant to interpret human signals, why not just supersede those input controls with direct “by wire” inputs from a central computational source… since anything but the smallest planes is all “fly by wire (or hydro)” now anyway. Move the override one step deeper in the path than between the stick/rudder and seat. :smile:


#17

I would assume the point is that this way you don’t have to physically modify and customize each and every plane you want to fly remotely - you can stick your robot in the pilot seat of any individual compatible plane and off you go.

And when you’re done, you can just remove the robot and put a regular pilot back in the seat without having to re-convert the plane back to a human interface.


#18

This topic was automatically closed after 5 days. New replies are no longer allowed.