Hacker News new | comments | show | ask | jobs | submit login
Insect-inspired robot can fly a kilometer on a charge with flappy wings (techcrunch.com)
112 points by confiscate 10 days ago | hide | past | web | favorite | 25 comments

Matěj Karásek, Florian T. Muijres, Christophe De Wagter, Bart D.W. Remes, Guido C.H.E. de Croon: A tailless aerial robotic flapper reveals that flies use torque coupling in rapid banked turns. Science, Vol 361, Iss 6407, 2018. DOI: 10.1126/science.aat0350


Thanks for the direct link!

This video shows more of the DelFly Nimble's flight maneuvers: https://www.youtube.com/watch?v=CEhu-FePBC0

I lost $1,200 sponsoring something like this on Indiegogo. It failed and nothing got shipped to us backers. Sadly, they didn't open source any of their work. https://www.indiegogo.com/projects/robot-dragonfly-micro-aer...

Sorry to hear that. Is the prototype demonstrated in their video a fake? Or something that could not be mass produced?

I'm imagining a thousand of these released in a city. They could spread out through every single street and map an entire city in a matter of minutes. Do it at dusk and you'd never even know it was happening, you'd just think bats were out feeding.

Considering the size of these, and their incredibly light weight design, that's probably a bit unlikely. With current technology, they are not capable of carrying much of a payload.

It's great research, and very cool design (both mechanically, as well as the driving software), but frankly, these types of flying mechanisms just don't scale well and don't have much use. (But please don't take that of a criticism of their research, I think what they've done is awesome and I really look forward to see what they do next.)

We already know they scale quite well, your average duck goes 100 mph and will fly incredible distances on a relatively small load of fuel. It's just that we haven't been able to re-engineer them to the point that we can achieve that kind of efficiency, but we have solid proof it is possible.

Scaling flapping wing devices up is much easier than making them this small.

No more than insect eyes scale to bird eyes. The problem for wings is physics doesn't scale.

e.g. 1 Consider animating falling, scaled at 1 pixel per km vs 1 pixel per meter. The former is much slower.

e.g. 2 An ant can carry several times its own weight, but you can't.

> your average duck goes 100 mph

Apparently this was a record-setting duck; we wouldn't want to say "your average human goes 23 mph", even though Usain Bolt has done so.

Ok, we'll swap the Duck with a Peregrine Falcon.

Please take the 100 mph duck with an appropriate amount of salt (vegetarians, feel free to substitute (the duck, not the salt) with an equal amount of Tofu), and instead read it as 'birds are very fast'.

That would be a dive, not flying. Humans can reach great speeds while falling too! :)

But if we can build a replica duck that matches the record-setting duck's performance we can create lots of copies of it

It doesn't scale well is a bit vague. Care to clarify? What doesn't scale and why doesn't it scale? My mechanical engineering skills are absent...

- Is it that it cannot be miniaturized? However, very small insects seem to use flapping as well.

- Is it this particular mechanical setup that won't scale?

- Is it that vision cannot be miniaturized?

- Is it energy density that is not high enough?

- Is it material strength?

- If the application goal is to for example land in a tree or circumvent all kind of tiny obstacles very fast, are there better mechanisms?

I'm always really interested in miniaturization. From watches to chips. So if someone tells it won't work in a particular field, that's important to know.

Scale way in the upwards direction. As you noted, it's great at small size. We can build ornithopters at the scale of this and bird sized devices, and they fly pretty well, but getting to man sized and airplane sized, would be quite difficult and inefficient.

I'm sorry, but I'm too lazy to answer all of your questions at the moment, though they're good questions. But generally the answer to all of your questions is "yes." (with the exception of the last question, which is "I don't know", but if you've seen a hummingbird or dragon fly, you know exactly how insanely maneuverable they are, and I don't know if I've ever seen better.

For more information, see: https://en.wikipedia.org/wiki/Ornithopter

Mapping is a noble application. There’s probably more money to be made in the 24/7 monitoring of private citizens, including in their own homes.

Can we have one of these man-sized with straps so I can put it on my back, please?

It looks like quite a small bill of materials and standard servos. It would be interesting to see a proper set of plans for making one. Perhaps someone in Shenzhen could start selling kits on Ebay.

Thanks, that looks very interesting.

I think things like this show that very light weight solar panels will enable a huge variety of drone applications that aren't feasible now due to limitations of recharging.

Could solar cells for wings provide enough power?

These wings are obviously very thin, light and flexible. The battery is also a lithium-polymer one, which can deliver a lot of power for a short duration (eg: 20 watts for 5 minutes). With current tech, it wouldn't be possible for this robot. I also don't think it would be possible for this robot to generate enough power to keep flying continuously, even if you had 100% efficient ultra-light solar cells.

All of that being said, there are solar-powered drones with large static wings out there, typically based on glider designs. We can also imagine a design where a drone lands and recharges using solar power for ten minutes, and then flies for a short one-minute trip. With the fly-inspired design in this video, it could probably take off vertically.

No. Not enough ratio of surface area to volume.

One of the reasons these systems (and their insect inspirations) can fly so nimbly, while a jumbo jet or elephant is more plodding in comparison, is the "cube/square law" - an object increases in weight much more quickly than it increases in area, and it in reases in area much more quickly than it increases in length.

You can get a decent ratio of volume (weight and chemical power storage) to surface area and heat dissipation/respiratory potential with tiny systems. If you want a lot of area (solar power insolation), you need to go bigger and slower.

I don't know that it would be possible with this design. The wings are a kind of plastic sheeting. They kinda look like bits of shower curtain, actually

Applications are open for YC Winter 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact