I was interested to see this video of a new jumping robot. The blurb reads:
Salto-1P is a small monopedal jumping robot capable of continuous high-power hopping. We demonstrate a new control algorithm that can land Salto-1P's foot at particular spots on the ground like jumping on stepping stones or playing one-leg hopscotch. We call this "deadbeat foot placement hopping control". Precise foot placement enables Salto-1P to jump on surfaces like furniture. This work will be presented at the 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems.
Here's the video.
What I want to know is, what's it for? If it's just a demonstration of a new control algorithm that allows greater precision in maneuvering the robot, all well and good. However, in the real world, what function might need a hopping motion like that, or great precision in executing it? Could one build it into an artificial intelligence pogo stick, so that one always kept one's balance?
(Of course, there's always cat toys. Our two would go absolutely ape if they saw that thing!)
Peter
15 comments:
Jumping allows a machine to traverse terrain that gives rolling stock fits. This tiny thing can climb stars. No rotors required. Think of the ability to maneuver, say, inside a collapsed building or across a boulder pile. Wheels or tracks would have to be inordinately large, and flight requires much more attention to weight vs power.
Just some quick observations off the top of my head.
If they could be made cheaply enough, they could be fitted hardened steel exteriors and used as targets. Hopping randomly around in a prescribed area between the berms would be perfect
Mine clearing, for instance.
Shaken, not stirred?
Too easy. Ever used a "Jumping Jack"? It's a gasoline-powered single foot compaction machine for construction, weighs about 120 lbs and bounces up and down about 3 inches on a single 12" X 14" foot to compact dirt, backfill, crushed stone, "crusher-run" (coarse granules and gravel), etc. Very effective at compacting, absolutely brainless and extremely irritating to operate.
Add some weight to this thing and some simple programming and let it do that job automatically. Make it cheap enough and it can just be left in the trench and backfilled over.
Strap two of them together, and you've got the foundation of a mech.
Technology, one step at a time.
Put a claymore on the front of one, and send it to a target painted spot or a gps coordinate.
Bouncy bouncy bouncy BANG!
It's all about balance and accuracy of 'jumps'...
Plus there are tremendous savings with the other three legs no longer required.
Put a small bomb on it.
Have swarms of them hop towards an enemy position.
Hopalong Asteroid/Moon explorer. And if it can hop hard enough, just so... why stop at one asteroid?
I'm with the bomb folks, one or more, attached droppable devices that it could hop in and spread around an area and with a bit of luck hop back out to be reused. If nothing else that would be a great distraction to unleash a few seconds before attacking with troops.
I'd worry about sand, mud and ice throwing off the hopping but it still would be great in a lot of situations.
I'm commenting late, but one-leg locomotion algorithms are a key building block for dynamically stable two-leg locomotion algorithms. Running, jumping, climbing stairs, etc., all rely on accurate single-leg placement.
That is absolutely amazing tech. Rapid analysis of trajectory, terrain, and capabilities is a cornerstone of any fast motion. Better optical terrain mapping needs to go along with it to make it hugely practical for much, but this is a really cool first step.
Post a Comment