This weekend, Amazon announced a new delivery method that has many people excited. Called Amazon Prime Air, the service delivers your order in 30 minutes via a "quadcopter."
Take a moment and watch this video that the company uses to show how the service works.
Amazon openly admits that Amazon Prime Air isn't a delivery method that you can use right now. However, the only major hurdle listed on Amazon's website is the fact that the FAA hasn't cleared its quadcopters for flight yet. The company anticipates approval in 2015.
I'd like to start a discussion on the feasibility of this idea. We can all agree that it is theoretically possible, and that it is also not really doable at the moment, but I think it could be fun to delve deeper. Here are a few of my thoughts.
In a 60 Minutes interview that aired Dec. 1, Amazon CEO Jeff Bezos said he expects a 10-mile range with the drones. This means that this option would only be available in a few selected areas. You then have to hope that those areas happen to have the product you're after in their distribution centers.
2. Drop zones.
In the video, there is a nice patio placed conveniently for the little drone to land and drop off its delivery. However, in the urban areas where this is targeted it won't be so easy. Where do you intend to land such a drone when there's no yard, only a sidewalk and a street? What about apartment buildings? What about when the tenant isn't present?
Drones falling from the sky? Drones landing in a yard with a pet? Drones colliding with obstacles? The mountain of liability insurance and paperwork seems like enough to bury this entire idea all by itself.
The drones just seem like perfect targets for theft. Not only the items they are carrying, but the drones themselves are likely to go missing. Running a system like this would be like dumping masses of drones into the trash every day. Normal package delivery has problems with packages disappearing from doorsteps; consider for a moment if an expensive piece of technology attached to that delivery could be swatted out of the air or snatched on landing. While I'd like to give our fellow human beings the benefit of the doubt, a certain percentage of loss should be expected.
Amazon mentions 30 minutes in the video several times. While that does sound interesting, I have to wonder how it plans on getting anywhere near that kind of timing. The time it takes for your order to simply go through processing and packaging should consume several times that on any normal day. Maybe Amazon is just referring to transit.
I think it is a cute idea. At this point though, it is just like Ford or Volkswagen trotting out an amazing concept car that will never see production. Sure, we'll learn some things from it, and some aspects may surface one day, but I doubt we'll see this car on the road. What do you think?
I saw Bezos talking about this delivery system on 60 Minutes and I immediately thought it was impractical because of theft and weather. I can image people trying to shoot them down even. Do these things fly well in all types of weather? I didn't think about the apartment number problem. Interesting. Maybe there will be Amazon chimneys on roofs of houses where the drone can deliever without even getting near the ground.
But do we really need stuff in a 1/2 hour? If so, it must be life threatening then and you should probably spend the time getting to the hospital.
Caleb: You nailed it. This is as cool as it is impractical.
In my apartment complex deliveries typically go to the office with a note left on the tiny mailbox by...the tiny drone man on the drone? GPS won't help a drone find that safe landing spot. And I can see a dog or prankster attacking a drone.
Let see if I need something really, really bad and I am 10 miles from the distribution center--how about a pick up service?
Still, somehow, someday I imagine drones will become commercially viable.
A lot of the "infeasibility" I keep reading about online today regarding this burgeoning technology is the problem of theft and vandalism. I'm not sure if I get this concept. This is still private property. Why would anyone be more prone to shooting down a drone than slashing the tires on the FedEx truck? It's the same crime. Just because it's in the air, doesn't mean it's open season.
I wasn't surprised to see negative comments on other boards, such as Yahoo, but I would've thought this would be pretty positively embraced here on EE Times... perhaps that means it really isn't feasible after all. But deep down, I have to wonder if drones won't be a disruptive technology to the transportation industry in the near future, and that Bezos really is on the frontier of something pretty exciting.
I read in the morning newspaper headline that Amazon getting ready to deliver via drones too...optocopters etc. I guess that would be awesone. Many people who live outside US wants the device directly from the stores in US, just for the quality, it would be just perfect. But wont the delivery cost be more than the product cost...
I believe Pizza Hut in Great Britain demo'ed delivery of pizza by quad copter earlier this year. Also, a town in Colorado is issuing hunting licenses for drones (was just in the news over the past couple of weeks).
A few have highlighted about the risk of the package going missing, from some peeople commandeering the delivery drone, but there are other, darker risks that should also be considered.There were also doubts raised about the availability of a suitable co-ordinate system that gets the delivery exactly where it should go. Apartment buildings and major office complexes would be trickier ones.
Presumably the drone is going to return the carry box to Amazon's fulfilment centre. It could be carrying something unpleasant for the workers at that centre, or if the drone is cyber captured, could end up targeting another destination altogether. That gets really scary.
Wouldn't you assume that there is a camera on board? And a wifi receiver? (Temperature and humidity maps could be educational too) I am sure there would be a few prize winning photos of the happy faces of the skeet shooters. What a gold mine of video and wifi infomation these things could generate, leaving the google car in the dust.
While not in any way dismissing the importance of the several excellent criticisms in this article, I would at the same time point out that for every genuinely new option there is an often deliciously amusing history of dismissing that technology by comparing its many disadvantages to methods more familiar at the time. The technology then goes on to plow some often very unexpected path that never looks quite like what everyone was expecting, and so does not fall precisely under the criticisms (or overcomes them in unexpected ways).
Take for example this 1800s conversation (from my imagination only!) about that newfangled 'telephony' idea...
"Posh! This silly new 'telephony' device doesn't even leave a record of what was sent, like any good telegraph system would! And what is the point of using imprecise voice instead of unambiguous clicks to talk between two telegraphers? Yes, a few non-experts have talked about ordinary people who don't know Morse code using these devices, but they should be ashamed of themselves for pandering such nonsense. Would you try to string incredibly expensive telegraph lines between every house in a community, just to share some kind of silly gossip?! Simple math shows that the resulting square of the number of lines used would bankrupt any community beyond a handful of people! So for those who understand technology deeply, it is all too apparent that this 'telephony' idea is nothing but a passing fancy, a bit of entertainment with which to impress naive non-technical folks who do not realize it has no more depth or practicality than feasting on rabbits pulled out of a magician's magic hat."
Seriously!? I can't believe you guys are taking this publicity stunt seriously. There are far better means to deliver this stuff. One would be to beam the product to your house or appartment à la Startrek. Another way would be to get a 3D printer and get the manufacturing code online and print it in the comfort of your living room.
And it will kill retail businesses!? Really? Having a warehouse with everything in stock within 30 minutes of drone flight from your house? Guess what, you might as well open a retail store and attract the traditional shoppers too. Just watch the Colbert Report that aired last night (12/2/13).
So the issue with drones is that there are only two ways to fly (in the US). You are either using Visual Flight Rules (VFR) or Instrument Flight Rules (IFR) which requires a transponder and a filed flight plan.
For VFR the safety practice is to "see and avoid". However, drones don't do that as is, so there's a hazard to flight for the rest of the flying community (and the folks who live under them). Is it OK for a drone delivering my pizza to put a 787 with 300 souls aboard into the ground? Probably not. It's also not fair to put the burden to avoid on the non-drones and will be (is being) totally resisted by the flying community and their substantial support in Congress. So now the drones need pilots to do the see-and-avoid, or significantly more sensors (which ups the costs, lowers the package payload and generally kill the value proposition).
For IFR the requirement to file flight plans, carry transponders, and communicate with controllers is going to significantly up the requirements, as well (i.e., costs & weight -- same issues as above). The flight plan filing is likely trivial. The rest of it is really an issue of pushing the responsibility for safe operation off onto the National Airspace control system. Which brings up the question: is Amazon signed up to pay for the additional controllers that have to be hired and the improvement in the radar systems around the US to actually see in all the radar-blind low altitudes that are away from airports and currently uncovered?
In the end is this possible? Yes. Feasible? Only if you push the true costs off on everyone else. Because of our grand tradition of allowing people to make money by socializing their costs and privatizing their profits, this idea must be bound for glory.
Drones are not cute- they are a very serious threat to privacy by an already intrusive government. The amount of air play and press coverage this asinine idea is getting is an obscenity to freedom and democracy, desensitizing a naive and stupid public to a tool of evil. Drones are not cute, they are spy tools. Bezos is a provider of compute farms to NSA and the government and was among those corporations who willing opened his komono to NSA. Amazon is playing everyone for a fool with a really stupid idea for parcel delivery, but a brilliant way to get a drone to intrude onto your property and airspace. The camera? Why it's merely for navigation and collision avoidance, Citizen.
When I read through some of the comments by my Engineer colleagues, I'm truly surprised at how stupid they assume the Engineers at Amazon actually are when they performed the feasibility studies: Do you think they lack the common sense one would expect from a well-trained Engineer?
Here is my Facebook post to a general audience on the subject of the drone delivery:
Of course, there are catches: • In the version demonstrated, it can only carry up to 5 lbs (~2.3kg), which according to Bezos is 83% of their shipments;
• It has a range of only 10 miles radius from the distribution center. At present, Amazon has 92 of these worldwide, so they would need to add more, smaller ones in urban areas (but see below);
• Wind: As anyone who has flown model aircraft will tell you, these small vehicles are sensitive to winds, both gusts and steady breezes. It's one thing to compensate the flight path for wind, but it drains power, shortening range. In addition, urban areas have tall buildings, which creates all sorts of weird air currents;
• Temperature: To be "environmentally friendly" these drones run on Li-ion batteries, which like to be warm to extract the most energy. Although a propeller becomes slightly more efficient in cold weather, the loss in battery energy more than offsets this. [There's a reason why we'll be depending on petroleum as a transportation fuel, especially for aircraft, into the foreseeable future: It holds much more energy per pound than can be stored in batteries.]
As I somewhat alluded to above, this would need much more R&D to work in an urban area, for things like delivery points, additional wind currents, lack of urban warehouse locations, and so forth. If you bothered to watch the entire 60 Minutes segment on Amazon, you would have noticed that the warehouse they showcased was 1.2 million ft² -- That's about 30 acres under one roof; and about double that in land area when you factor in loading docks, truck staging, etc... -- You're not going to see this in downtown NYC or Chicago any time soon.
On The Other Hand, suburban small parcel delivery is much more feasible, at least in the beginning, as it solves the issues of single family dwelling delivery in low crime areas, and of industrial park siting of the distribution warehouses, which incidentally would be somewhat smaller as they would be limited to items weighing under 5 lbs.
My initial observation is that this would be exceedingly useful for medical supplies in accidents. Once medical personel get to the accident scene, they often don't have supplies that could be used on the victums, and in major traffic, it takes a long time to get victums to hospital. With ER doctors able to access the patient vitals remotely, this gives a mechanism to quickly deliver drugs before the patient reaches the ER.
Think of trauma victums which could benefit from life saving medicines as soon as possible, or even during large events where there are medical personell, but no fast mechanism to get the patient to an ER - large stadium. Those initial minutes can be crucial for heart attacks/strokes/...
Recall that wireless transmission was ridiculed until it was used by the Titanic to radio for help. Who would have thought everyone would need to talk on the phone in their cars while driving! And electronic tablets to read books and access information - why would anyone need that when we have paper books and libraries?
Thanks for writing a sensible piece on this idotic idea. If you have ever flown model aircraft or helicopters you will realize this is complete nonsense. Air currents near the ground and buildings are horrendous and include lots of rotors that will just dump the vehicle.
Bezos clearly is no expert on fluid dynamics or vehicle flight.
The harshest criticisms take the form, "If this system won't work everywhere and always then it can't be used anywhere ever." Quite absurd logically. Even conventional airplanes are restricted where and when they fly. Obviously only some customers could benefit from this idea, but why not let then do so?
Some of the objections can be overcome. For example when requesting drone delivery, the customer would have to verify that there is a landing spot and that someone will be standing by to collect the package when delivered.
Interference with manned aircraft need not be a problem, either. Even light planes are not permitted to fly below a certain altitude. If the quadricopter had to stay below that height there should be no collisions.
Weather can be an obstacle, but with modern forecasting the vendor could revert to comventional methods when the weather is unsuitale. The customer simply understands that drone delivery is contingent on conditions.
As for the drone snooping, the nosy operator would do much better with a dedicated vehicle for that purpose. It would carry a much better camera and be able to linger at the target.
Most of the newa about drones these days is their military application for spying and killing. This, however, is far from the complete picture. Drones are already helping conservation efforts, locating missing people, assisting fire and rescue operations, and conducting agricultural surveys. A couple of comments have suggested that these little quadricopters could deliver emergency medical supplies. Thus. even if Amazon's package delivery doesn't work out, the experiment would certainly advance the art for these other useful and even vital functions.
I remain a little skeptical, but I say give it a try.
Caleb, respectfully, I couldn't disagree with you more.
It's interesting that as humans we have an inherent optimism bias toward ourselves and an inherent pessimism regarding the outside world (some survivalist goodness there, of course).
This is an amazing story and an even more amazing reaction to me: This technology exists today and the ONLY thing preventing its implementation is regulatory and safety concerns. That's amazing to me. I'm not minimizing them, but we've turned a historical corner when magic is commonplace and human embrace lags. We used to wish for magic; now that we have it, we don't know how to deal with it.
As innovators and engineers, let's celebrate the achievement and convince the world that this stuff will work, will work profitably and safely. Bring on the octocopters. Bring on the autonomous vehicles. Bring on the goodness of IoT.
Only engineers would get into a discussion about the feasibility of this :-) How much free advertising did Amazon just get by being in the news on the busiest days of shopping?
BTW, I heard this on CBC radio this morning, so I'm not taking credit, but given the timing, you have to think this is a very elaborate piece of marketing.
Meanwhile, back to the feasibility discussion!! I think we need to add cloaking http://media.utoronto.ca/media-releases/thin-active-invisibility-cloak-demonstrated-for-first-time/ and you make the package appear when you are ready to pick it up! Sorry, just wanted to push some local technology :-)
I agree. The legal ramifications and theft will be too much to support such an idea. And with a 10 mile radius, why not just jump into your car and drive to the Amazon 'will call' window and simply pick up your purchase.
Flight time is about 10 minutes and the radio range on a fully charged battery is about 1/2 mile or so. Now you add some weigh like a small package and flight time is really short. What sized box will it lift? It would take less time and money to just deliver the package the old fashioned way.
First, everyone, please: Don't ever assume that privacy will take care of itself. It won't. It is a right that everyone must pay attention to and treat carefully, and it's a big part of what real democracy is all about.
Privacy will also require some innovative thinking as we get more and more connected, since in many cases the real issue is not how much information is available, but who gets access to it. If only a few have access, you and your society are in trouble, since such highly concentrated power will inevitably be subject to human foibles -- and most humans really, really like power.
The James Madison et al concept of balance of power and cross checks very much comes to mind here, but it will likely need to be broadened in new ways. Physicist and science fiction author David Brin advocates a sort of open-source privacy model in which everyone has equal access to public information. It's a pretty unsettling thought, though arguably not that different from what many people are already doing with Facebook. It's also a lot less scarier than the idea of some very small number of people getting exclusive access to everything. Open privacy in contrast lets everyone be part of the free press -- and that in turn makes it a lot harder for some small number of folks to do really stupid or greedy things without getting caught.
With all of that said, blaming drone cameras as somehow magically putting us over some critical horizon of loss of privacy is, to the say the least, a strangely myopic view of the problem. For example, what other IT capabilities should be banned based on worries about what you could do with them?
Well, the entire Internet of course. Cars too, because they can carry cameras, look at Russia! Of course people can carry small cameras too, so I guess you have to ban all small cameras... so no more smart phones, stop that right now! No bank cameras of course, you know how this stuff creeps in starting with what sounds like a reasonable exception... Facebook? Ohmigosh, I guess you would just sort of have to arrest everyone there, at a minimum? And LinkedIn too? And Twitter? And any social medium? I mean, these nefarious folks are seducing the entire population of the world into giving away all their secrets and enjoy doing it! It's almost like one of those, what do they call them, free-market economies where everyone trusts each other enough to advertise who they are and what they can do freely, so that everyone can make new innovative connections and build new forms of trust.
Less cynically: Ludditing does not enhance privacy. Bezos is trying hard to show how drones can do useful things, and I applaud him for it. By doing so, he will eventually help address the very privacy concerns that folks are worried about. Why? Because he's showing how positive, free-market style options for sometimes scary new ideas can benefit everyone in positive, growth-generating ways. That in turn will produce the political will to ban the truly negative uses of the same technologies, something that's hard to do when people can't see the full spectrum of possible uses.
I think it's obvious that Bezos is putting UPS and Fedex on notice. Keep your pricing lower than it costs me to build millions of tiny helicopters!! Effective whether or not it actually gets implemented. At the same time he builds so much loyalty with those of us in love with technological advances. Quite brilliant and bold as usual.
The way to use the drones is to couple them with Google's autonomous vehicles. The self-driving vehicle drives up to your house and a drone delivers the package to your front door. It then calls your phone and tells you the package is there.
FAA issue is simply keeping them from crossing airfield with control towers (commercial airfields). The general aviation airfields is just another problem to work through. This is what the FAA is working on. The operating height is not an issue since the general aviation height requirments and this would not overlap except in the area of a general aviation airfield. Navigational issues are another non-issue. GIS is a major technological inovation that every level of govnerment has invested in. So getting to wherevere is not the technical problem that some of 'engineers' here think it is. The end user is expected to be at the point of delivery so the criminal conduct has automatically escelate past petty theft to robbery. I doubt I am going to see an increase this activity as well as someone stepping out spur of the moment and discharging there firearm another felony. No, the disruption factor will be in the area of keeping high mainteance vehicle usage minimized. What is the head count and expense of operating one package truck. I'm certain Bezos already knows this. QVC, EBay, and company won't be that far behind. Including UPS and FEDEX.
The real difficulty in near surface navigation is utility lines, guy wires, etc. This is a general difficulty of personal flight applications. Look up sometime at the maze of urban connections. My neighbor was killed several years ago in a helicopter accident. The pilot became fogged in and was following a highway. Landing skid snagged an electric wire. End of the road.
Short of Google mapping the power, cable, and telephone grids, I'd like to see how they solve this one. (We don't have the imaging capability. There are thousands of changes daily. Even the utility companies don't have complete and accurate information.)
There may be a functional limit on homeland drone incursions after all...
Amazon will need to use differential GPS. Probably with the user doing a "self-tag" for where he wants the package delivered. An app could be developed to run in a cell phone to read offsets from a local Amazon beacon transmitter and the internal GPS receiver in the cellphone. Send all the appropriate data in an encrypted email to Amazon. Or maybe Prime users can mount a beacon device on an elevated delivery platform to guide the copter for the last 100 feet. (All good so far.)
Viola, before you knew it, another cracker attack vector. Making and operating a counterfeit Amazon delivery drone to harass people will become the new "swatting" technique. Or a way to drop mini-bombs or bags of burning dog-crap on your hated neighbor's front door step.
Amazon octocopter maintenance better be excellent or people will be finding dead ones in random places all over.
If I find one in my backyard, do I have to send it back to Amazon on UPS or FedEx? Or will Amazon send another "tow-truck copter" to fly it back home?
What if my dog chews the hell out of it? Anything in the backyard, on the ground is the dog's new chew toy. He loves toys that make whirring noises.
If I didn't order anything from Amazon, can I use a baseball bat to knock the thing down for violating my back yard's airspace?
I use Amazon a lot but I don't like the idea of thousands or hundreds of flying delivery obstacles in low airspace. It's bad enough I have to keep an eye on where I put my feet. WIll I have to keep an eye on the surrounding airspace as well?
"The telephone will never be practical. You would have to run wires to everyone's house and that would be impossible."
Actual quote from the mayor of NY after a demonstation of the telephone in 1876.
The point is, what looks impossible now, may not be in the future. Drones could have counter-measures built in, like a mosquito, to avoid would be swatters. It could even have artificial intelligence to select best landing sites.
I used to read about flat-screen TV in the 60's but I never really thought it was practical. You would have to run million of wires internally to each and every pixel. How would you be able to do that and assure every connection was perfect?
But engineers found a way. I still can't believe it and I have one!
Jeff Bezos, in my opinion, is a visionary that far exceeds all others. I'm sure he is well aware of the issues and will find a way to overcome them. Whenever he sets out to do something, he usually succeeds so I'm betting drones will be routine in a decade or so.
My first concern when I heard of this idea was not the technological challenges (which can be met) but the cost factor. A signel quad delivers a single package and it takes time. So, you can only deliver enough packages in the day as you have quads available, so perhaps ten or so trips per day per quad. And given the cost per quad, it will have to be a premium delivery charge for this kind of air mail to make it cost effective for the company. You also have a limited area you can serve, so your available market is highly limited. Payload capacity is low, further limiting your market.
I would imagine, then, that this kind of delivery would only be requested when the purchaser absolutely needs delivery within an hour or two. That, too, restricts the market size. The fortunate thing (for Amazon) is that customers having such urgency are usually willing to pay for the quick response.
I personally would think that hiring someone with a bike to courier the package to the destination would be more cost effective. But you never know. If the cost of the quad itself can be made low enough, the economics shift and the approach could actually become economically viable, but still only for a very limited market. Don't expect to see these things being used frequently, if they ever do get used.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.