Lily Drone Folds After Funding Dries Up

[Marco]

The whole point for this one was not to be "RC" but fully autonomous with barely any user interaction required. Even the biggest players aren't quite to that level yet.

Maybe a ToF based direction finder for a 433 MHz beacon would help, forget about image processing on the drone.

[CraigHB]

It takes some innovative technology to do what the Lily set out to do, not something particularly cheap to manufacture.  Even so, drones or multi-rotor RC aircraft with cameras are a dime a dozen anymore.  The difference is their flight paths have to be pre-programmed or manually controlled.  Even if the Lily had accomplished what it set out to do (follow the user autonomously taking video), these things are typically made cheaply as possible and tend to have reliability issues.  They could run into lawsuits where people are getting hit by drones following them around or other damage to life and property when a drone flies off out of control.  It happens all the time with drones sold currently and was the imputes for the FAA to impose regulations on them.  The whole idea was wrought with pitfalls so it's no surprise to me it didn't make it.

[helius]

Seems to be another case of "AI is easy" handwaving. (In 1966, Marvin Minsky assigned "vision" to an undergrad as a summer project.)
The Lily was to be an autonomous UAV with no operator backup. Just the legal liability issues involved are crippling.

[Kilrah]

Maybe a ToF based direction finder for a 433 MHz beacon would help, forget about image processing on the drone.

GPS is enough for locating, vision recognition is just there for additional precision and image smoothness. The big issues for something you "launch and forget" are obstacle detection and positioning reliability.
Technologically the first is becoming possible, the 2nd is getting better but still a bit of "spray and pray"... but the state of the art in both fields combined costs >3x what Lily was hoping to sell for, anyone with a foot in the industry would call BS, completely unrealistic.

My stance is that anything that needs breakthrough technology should never be crowdfunded or even announced before being ready. All they achieved is reveal the interest in this so that the big players could invest all their know-how and WAY bigger resources into competing with them.
It's not a bad thing per se, but anyone going this way should be aware of it. Maybe they were. Make a big buzz so that someone else does the complex job that's outside of your capabilities and creates what you want... IF they actually do refund all preorders I would consider them geniuses.

The lawsuit developments will also be very interesting for anyone considering a tech project. Nowadays if you don't do what they did you have 0 chance of getting attention. If that gets punished... it would be a giant blow to innovation and small-medium scale developments. If when you try to do something good all you gain is being sued for doing what everybody does, why bother. I'd rather sit on my chair and watch the others in their mess than try to do something, even if I've got great ideas.

I'm NOT condoning the stupid "not actual footage" or whatever warnings that would make it legal. The only solution is for everybody to be prohibited from including any footage that would warrant such a warning... but it would force a "downgrade" on so many big players there's no chance it would happen. Actually, maybe with Trump's councelor pushing down on them that could be a thing...

[Marco]

GPS is enough for locating

Yeah I guess you're right, at least as long as you calculate position from the same subset of satellites.

[PlainName]

Just the legal liability issues involved are crippling.

I wonder if that might be the real reason. I've been involved with a novel (at the time) invention which worked well and got to the point of selling hand-built pre-production units, and then got dropped because of the legal and insurance potential. The current crop of rules regarding drones didn't exist when they started, I think.

[PlainName]

If you read the actual complaint : https://regmedia.co.uk/2017/01/13/lily_complaint.pdf

It appears that they made deliberate attempts to disguise the fact that the demo footage was shot from a GoPro, and there are emails to prove it. Seems like a pretty good case for false advertising.

I read the complaint and am not that sure it was false advertising (bearing in mind this complaint isn't going to offer up any evidence for the defence). Yes, that video is kind of incriminating, but only out of context. The on-screen text ("Lily shot | Follow", for example) I would normally think to myself as having the subtext "this is what it would look like". Typically, this type of advertising would say "Actual footage" or similar, so I don't think it's that clear cut.

Possibly the decider would be the Kickstarter blurb - in that, did they say they need the dosh only to get manufacturing started or to finish development? If the latter, the punter should suspect it can't be real footage.

I imagine the same applies to the pre-sales. That is, whether it was sold as basically in production but not shipped yet, or 'soon to be ready' kind of thing.