Unicycle Drones!

Well, that’s what they’re calling them:

And here I thought this was going to be about a new selfie-videography tool in your arsenal.

Website seems long on vision and short on engineering reality. I’m quite familiar with self-contained electric wheels - they are fine at speed, but not so great for stop and go… or forklifting packages. Plus they are only “self-balancing” along the travel axis - the side-to-side remains up to the rider, something these allegedly autonomous units seem to lack a substitute for.

Making an autonomous unicycle is the easy bit of that project. I’ve pondered the idea of a robot unicyclist myself - side to side balance would work in the same way it does with a human powered uni, by steering the wheel under the centre of mass (and if necessary idling whilst doing so). Simply use a gyro - or indeed any mass which you can rotate on the vertical axis, accelerate it in one direction or the other to steer using Newton 3.

a wheel is not that obvious for a robot, a ball makes much more sense from a control engineering perspective. See this example


The example that they show here is more a bicycle than a unicycle since the two units need to coordinate. The only thing is that the frame is the box itself.

this is also a funny example of a hopping bot


It’s a pretty poor concept at this point imo. They don’t have anyway to protect the package in case of rain or a crash. Also it takes some pretty expensive electronics to make something like that so they would have real problems with people stealing them. With battery tech where it is right now there’s no way to put enough power in them to move at any kind of decent speed as well.

Neither control electronics nor battery capacity are issues - you should see how physically small the pack on a human carrying electric wheel is - they actually get away with mounting it on one side, not even balanced!

People kicking them over to steal the motors and batteries (not to mention packages) is on the other hand…

And making such a unit that can self start and recover from outside a narrow range of balance is not as trivial as some seem to think - though adding deployable outrigger legs would help a lot.

I disagree. If the application is to deliver stuff with these specs :

  • unicycle drones that swarm together to transport heavy parcels

  • Transwheel drones use a single self-balancing wheel and a robotic arm to pick up and carry packages…

  • The electric robots would work alone for small deliveries or operate together as a swarm to carry larger loads…

  • They would use GPS to help them navigate and facial recognition to confirm the identity of a recipient…

  • Designers say the robots could allow customers to receive deliveries at times that are convenient to them…

The darpa challenge has resulted in spectacular robotic cars but to build a robot that would be cheaper than a guy in a van would be a formidable challenge. I think it will take some time before we see this kind of technology.

This is where it’s necessary to draw a big distinction between practicality and economics.

ie, if you can do it, unit cost will end up being a small fraction of “a guy in a van”

The question is more can it be done at any price. If it can, the per-unit electronics to implement the idea are actually fairly cheap once you get beyond the prototype stage.

Having robots navigate today’s streetscape is pretty challenging; a potential future one where every powered vehicle is networked might be a bit better, but it’s still not clear that a self contained unit with a single wheel is advantageous - it might be slightly more efficient in travel, but what is it going to do during the rest of a delivery, hop up curbs and steps? And cruising along riderless and so low enough to the ground to be invisible to drivers won’t work well either.