With all this talk about Amazon and other less serious drone delivery concepts out there, there has been a large amount of discussion regarding the feasibility of delivering a package via drone for same day delivery or general delivery. While I don't think this is really all that feasible from a logistical or safety point of view, the idea is interesting.

What would be more realistic in my mind is a better system where you have autonomous ground vehicles carrying packages and UAV's. The autonomous ground vehicles being developed by various companies like Google's self driving cars are getting better and closer to integration into society, although I'm sure the regulatory hurdles will be just as painful.

The basic concept is you have that autonomous ground vehicle, probably just like a UPS truck outfitted with sensors, carry packages around the city to delivery locations much like is currently done using people. Simple enough but that alone wouldn't get the package to your door as I know I wouldn't want a van driving through my lawn up to my door to "fling" a package at my house or door. Instead, the autonomous UPS/FedEx/whatever truck dispatches a mutlirotor of appropriate size with your package to go from truck to door. Ground navigation is hard enough on the road but getting around a cluttered yard or sidewalk is going to be really hard. With a quad/hex/octo, you could easily have it take off, fly over all the obstacles, drop the package off near the door, and head back to the truck to essentially charge itself back up.

A delivery truck is going to be large enough to carry all the sensors it needs for autonomous navigation as well as packages and delivery "drones". As seen just a couple of days ago, Tridge and the team have also demonstrated a milestone in getting ArduPlane running and flying an airplane using Linux and ArduCopter likely is close behind. With that comes all sorts of possibilities like running SLAM or other computer vision algorithms, which would aid in a multi dropping a package off at your door.

If you want to get around the FAA regulations too, tether each multi to the truck! Then it's considered a tethered object and no different than a kite! You'll need a way to keep the tether from getting to the blades but there are enough people out there like Jack Crossfire that could whip something like that up and take care of that issue. It's one way to get things rolling without waiting for the FAA to get their act together.

With Amazon's announcement of drone delivery, much of the speculation has focused and assumed a drone is going to go from a warehouse to your door, which rightfully so, with just one or two images and a video of an octo, that's what people are going to think. However practicality suggests that's really unrealistic. It's going to take something more than just some quads to implement a system like this but I get the feeling it's going to be a combination of autonomous UGV's with UAV's that will get the job done.

Views: 950

Comment by Daniel Lukonis on August 27, 2014 at 12:02pm

Good article. I'm sure automated trains, trucks and drones are going to be intertwined in product delivery.

We (humans) are about land a robot on a comet flying thousands of miles per hour, millions of miles away. Hopefully we can land a robot on my front door mat in the near future.

Comment by Gary McCray on August 27, 2014 at 12:02pm

I doubt that Multicopter delivery of packages is ever going to become a common place occurrence.

At most it will be used for either very special and important packages that need to be delivered under very singular circumstances, or for hype and entertainment but also under very tightly controlled circumstance.

The safety considerations alone are just too significant for more generalized use.

Personally I would be horrified to see squadrons of these flying over head and would probably spend a lot more time indoors.

Having said that though I am very glad that Amazon has decided to take on this venture with so much enthusiasm.

Clearly they will be at the forefront of actually determining what is feasible and what isn't and in aiding the FAA, the government and the public to "see" that as well and that really is of benefit to us all.

As for the Automated delivery truck why not just stop and have the automated delivery robot wheel it up to your door - still a lot safer than a flying lawn mower especially if my kid is out playing in the yard.

Best Regards,

Gary

Comment by Gary McCray on August 27, 2014 at 12:39pm

Also the tethered object thing is wrong, the FAA provides no specific exemption for tethered objects and treat them exactly like untethered ones.

And that includes balloons and kites.

They may eventually decide to do otherwise, but at this point in time it makes no difference whatever.

They haven't done much enforcement but that isn't because they can't.


T3
Comment by healthyfatboy on August 27, 2014 at 2:02pm

Gary,

You make good points and I agree with you. This is just an idea of how one might do air delivery of items, granted it may not be as efficient in some cases. Maybe if you had a case where the home was a tenth of a mile or so from the road, it might be faster, but not necessarily better.

I do think having a lot of these flying around is a safety hazard no matter how you look at it, although I wouldn't necessarily say there will be squadrons of these flying around. In fact, no more than your general delivery truck. Either way, I wouldn't trust having something like this flying around me or my kids without much more safety flights and until reliability is much better.

I didn't know about the tether. I just made an assumption but it does make sense. I had to do a tethered balloon experiment once for work and it was a bit of a paid to go through some of the hoops but it's because we used a 500' tether.

Thanks for the comments and insights.

Comment by vf171 on August 27, 2014 at 7:14pm

University of Cincinnati has been working on this concept already...

http://www.uc.edu/news/NR.aspx?id=19929

Comment by Cliff-E on August 28, 2014 at 11:05pm

Tethered is a fading option according to the FAA since it (quadcopters, etc..) is a vehicle that must provide thrust as means for flight. FAA realized a lot of folks, even military starting using tethered as a loop hole to balloons.

I too have direct info I cannot post from the FAA, but here's the idea:

http://www.phantompilots.com/viewtopic.php?f=4&t=7822&start=20

Balloons do not provide thrust, they are naturally lifted by the fact they use helium. And local gov't typically have [restrictive] rules for balloons too--for instance Los Angeles does within city limits (permits!)...

Autonomous ground vehicles are also going to have issues due to the manned vs unmanned aspect. The mayor Los Angeles here wants autonomous cars (Uber, UPS/FedEx) within 5yrs. BUT having attended his tech sessions, they are seriously pondering separate autonomous car lanes since a bunch of sensors are not going to prevent accidents as much as we think--too many edge cases. And that makes sense, much like zones for flying. I think the key is going to rely on the car to car communications, but I can sure tell you that's hard stuff to make it robust.

Comment by Stephen Wuest on September 10, 2014 at 7:38am

I can't see the desirability of complete drone delivery from a warehouse (I wouldn't want low level drone highways and their sky pollution). Robotic vehicle delivery seems feasible, but I don't see how drones would fit into that picture. It's not a matter of knowing where the front door is, but of situational awareness to handle all the things that could go wrong. That is core AI research, and everyone has been focussing on merely the physical design of flying airframes.

Delivering a package to the front door doesn't make it safe. Someone following the autonomous truck can steal them all with a quadcopter. Right now, a locking mailbox is much safer. 

If Google had focussed on building technology that would disable any airframe coming within my fence line, unless it had encrypted permission, that would be a step forward. But it still would not address the AI problem. 

Comment

You need to be a member of DIY Drones to add comments!

Join DIY Drones

© 2019   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service