Feds give Google OK to test Project Wing drone deliveries | Hacker News Reader

Feds give Google OK to test Project Wing drone deliveries

engadget.com

62 pointsRileyKyeden38 comments

> Though the FAA released its final rules for drone operations in June, deliveries still aren't allowed. However, Alphabet has reportedly been testing them in the US anyway, skirting the regulations via a NASA waiver

You see, all American's are allowed to be innovative. Some are just more allowed than others.

That's always been true. Figuring out where the rules are flexible and bending them is part of the game (in almost all games, really).

Something of a difference between finding the flexible nasa waiver rule and actually obtaining a nasa waiver.

From my perspective, it's pretty clear that "Delivery by drone" is going to happen. It might not be this year, or next year, or even in the next 5-10 years, but eventually it's going to happen.

I think that if the FAA doesn't keep up with the pace that other countries are setting (in terms of regulation, laws, and rules surrounding out-of-sight drone usage), then eventually they will be pressured to rush some solutions which will probably have some bad consequences.

From my perspective I think it is pretty clear that drone delivery is never going to happen and is just a fad.

Firstly, let me just say, technologically drone delivery could definitely happen. There's no technical reason why not.

But if we look at the history of freight it has always been about efficiency, and new competitors have always competed by finding ways of being cheaper.

We already have less technical "drone delivery," it is called a courier, but outside of New York and a handful of other places they're niche. Nobody wants to pay to deliver just a single package, it is too expensive.

Even services like "next day delivery" work on the principle of economy. You have trucks which are on other deliveries (economical ones) just drop off the "next day" packages between them. That makes these "unnaturally cheap."

What do drones offer that a human courier doesn't? It cannot deliver inside most office buildings, you're still paying for fuel/maintenance/development/lost drones/etc. It likely has limited capacity.

Plus let's look at competing technology coming up: Automated vehicles. How long until a delivery van can drive itself to your home, and then send you a text to meet it outside? That technology is coming with or without drone development, and solves some of a drone's inherent issues (poor capacity, lower maintenance, better economy, etc).

I really think drones for deliveries won't go anywhere. It is just a fad thing because companies want to be cool. Too many problems/limitations/high costs, and few benefits. And as I said automated vehicles will kill any remaining shot drones may have had.

Except a lot of deliveries to residential areas happen when nobody is home. So getting a text to go outside and get my package from an automated vehicle won't help me at all. And this is where drones can come in handy. The drone could deliver the package from the truck to my doorstep. Sure, it may not work in urban areas with high rises but it will work just fine in single family house neighborhoods and that's a huge part of the country.

I think this highlights an important idea. The driverless vehicle and drone tech aren't mutually exclusive. You could potentially have a driverless box van that parks on a street corner and has drones pick packages off it and take them to your door.

Without doing any of the math, this actually seems like the right solution. An automated truck that can drive around while a swarm of drones are continuously dropping off packages - basically an aircraft carrier.

I deal with writing software for some shipping and logistics as part of my job.

This is already how a lot of this stuff works.

A manufacturer ships stuff to a DC, which ships stuff to a 3rd party pool provider, which ships stuff to a store, which then ships stuff to the customer (via "regular" mail).

Each step can be thought of as trading throughput for latency. Some manufacturers only ship things to the DC like once every few months (sometimes thousands of containers at a time), the DC to pools once a week (generally in tractor trailers), the pools to stores once a day (in either trailers, or delivery trucks), and the stores to customers multiple times a day (regular mail)

Adding another step on the end of that seems like the most likely "end goal" to me.

Get around the problem of the inefficient "last mile" of delivery by making the delivery truck a drone platform.

Or maybe the drone could drop it off in the back yard instead, or what not?

I'd assume that the drone would deliver from a truck to the front door. Combine that with a self-driving truck and you've got substantial cost and time savings:

- more than one drone delivering from the truck at a time

- truck not having to stop on the same street, just within fly distance of destination

- truck can move while drones are in-flight

- etc

I quite like this idea. But, in thinking about it -- why not a specialized ground robot from truck to doorstop? Kind of like the irobot packbot. I could buy the argument that drones are easier because they avoid the terrain hassles.

But for what reason do you need a truck. Why not self-driving "Packages". You need Range? Use self-driving "Batteries". I think range problems will be a thing of the past.

What do you do with the wheels & batteries after the package is delivered?

I think drone delivery makes a lot of sense in remote locations.

What does a drone offer that a human courier doesn't? The cost savings from not having to haul a human meat bag hundreds of miles into the bush to hand deliver a parcel of medicine that could've been air dropped from a drone.

Drones are good for extremely fast point-to-point and costs go way down if they're mass-produced. Far, far cheaper than human couriers.

Automated delivery vans are strong competitor but there's a lot more red tape involved with AI vans... drones might fill the gap until vans are on the road. By that point the cost of operating a drone might be so low that they still operate for expedite point-to-point shipping (think Amazon's Prime Now vs Prime Same-day shipping)

I don't think drones are going to fully replace anything we have now, but they are going to augment other methods.

The "last mile" is extremely inefficient to do in a delivery truck. Especially so in suburban areas.

Ships, trains, trucks, and even normal cars will probably always be cheaper than a drone in terms of raw price per lb of stuff moved, but there is a big "want" in the market for quick deliveries.

To use an already overused analogy, a truck full of hard drives isn't going to replace your home internet connection any time soon.

I agree. When we have self-driving cars, why would a drone be cheaper then having a self-driving truck drive around and drop stuff off? Also, I don't see a drone bridging the last step to your doorstep. What if it's a large package such as an air conditioner? There's so many odd package shapes and sizes. Why not just a robot picking up the package and walking it to the door? Then you could sign for packages as well.

Drone might be faster (ex: traffic)

The mistake, I think, is to see drone delivery as a competitor to ground shipping. Ground will always be cheaper, and self-driving will only further push down the cost.

What drones offer that no other shipment method can, in addition to rapid, point-to-point travel, is a completely different economy: As an example, it would be prohibitive to send a truck, even a self-driving one, out on a 40-minute drive from a warehouse for a single on-demand delivery. There’s no such overhead for a drone: You don’t have to wait until a bunch of other orders have piled up; the drone can just head out and make the delivery as soon as it’s loaded.

This is obviously overkill for your average Amazon order, but for on-demand services, it’s currently competing with an entire ground vehicle going out for a single delivery. I’d say it’s inevitable that drones will be both cheaper and faster than that.

The thing is, you'd have an autonomous vehicle that could park in an area and use a drone(s) to deliver to your door step. Remember, this doesn't have to apply to everywhere; urban and rural delivery likely makes up the majority of companies like Amazon's B2C delivery.

You make reference to underlying assumptions about the economics of drone delivery, but I don't think you've shown anything.

What is the average cost per UPS delivery, and what would the cost be with a drone?

Assume a drone has an avg 1 year lifespan, can make 20 deliveries a day, costs $9k, and $1k in annual power and maintenance. Thats $1.40 per delivery.

If a UPS driver + truck cost $100k / year and can do 200 deliveries a day, that's also $1.40 per delivery.

I have no idea about the accuracy of my numbers above, but neither seems out of the realm of possibility. I can definitely see drones being economical.

The mistake you make here is thinking that drones are dominated by existing shipping methods.

They're not -- while they're currently more expensive than existing shipping methods, those costs will drop.

Furthermore, drone delivery is going to be much faster than any method of delivery. And what's really the difference between a self-driving bike vs a self-driving small-air-vehicle? One is going to be easier to implement.

There's no doubt for me that drone delivery will happen. It probably won't make sense for all packages in NYC, but it will make sense in smaller towns and for out-of-town deliveries.

Drones will provide speed, reliability, and at some point cost over the alternative. [Remember, they can be charged in an arbitrary location, so their electrical bill will be the cheapest place in their flight range]

I am honestly curious where the economics are going to come in. Drones aren't exactly a power efficient way to do something, and they're also heavily range limited in many cases.

There are certain cases I could see it being really useful and effective, prescription drug deliveries make a lot of sense to do by drone, especially because of the possible urgency and mobility issues for many people needing them. And the fact that pill bottles are light.

$0.50 buys an awful lot of electricity. I would expect other costs to be a bigger deal than energy.

These drones have minutes of flight time carrying kgs of weight with Watt-hours of energy consumption:

http://www.dronesglobe.com/guide/heavy-lift-drones/

As a sanity check, Googling on drone electricity costs has other people saying pennies per delivery.

> As a sanity check, Googling on drone electricity costs has other people saying pennies per delivery.

This makes sense, even assuming half an hour in the air for each delivery, the cost of residential electricity in the UK would mean ~5p of electricity for a 1KW drone making a delivery.

This is pretty sloppy reporting. The single-wing tail-sitter drone design depicted in both the picture and video accompanying this article and discussed in the text had been abandoned as of last year: http://blogs.wsj.com/digits/2015/03/17/google-working-on-new...

They do still have a program for drone delivery, but we know nothing about the drones involved except that these aren't them.

It looks like they show different prototype designs on their webpage: http://www.solveforx.com/wing/our-approach/

[including the old design. But in the image, they show something with helicopter-style rotors]

Every time I read Vernor Vinge's _Rainbows End_ I find myself wishing that someone would invent Up/Ex --- big pneumatic cannons which fire disposable cardboard gliders into the air; they glide down autonomously to the delivery site and then you tear then open and remove the package. Last mile delivery.

I don't care that it probably wouldn't be practical, or feasible. They'd still be awesome.

Package delivery is one of the highest profile uses for drones, but I don't think it's a lasting one. Drones are unlikely to become cheap and efficient enough to eclipse current delivery techniques on a mass scale. Especially because traditional techniques will become much more effective as autonomous cars become feasible.

However, lots of the technology being developed (robust autonomous flight, improved endurance & range, lower costs) will speed up developments for other applications that make better use of their strengths. Even more importantly, companies like Google are finally forcing the FAA to craft realistic regulations that don't completely cripple commercial applications. The FAA's lethargic pace has already severely hampered domestic development to the degree that Google had to do most of their development in Australia.

Google's Project Wing and Amazon's Prime Air are great projects and their advancements should be encouraged not discouraged. Now as far as the feasibility goes, "delivery by drone" should be limited to remote areas only where traditional delivery services do not have good infrastructure and network. Think a ranch in eastern Texas or Colorado vs a condo in Austin or New York.

No one need these guys swarming the skyline of Manhattan when the city has one of the best delivery networks in the nation as many other major metropolitan cities would enjoy.

I'm amazed with fixed wing vertical takeoff. And landing. There don't seem to be any moving parts other than the rotor blades. There ought to be some clever software in there.

Well, it looks like some Narcos already did (and without the Feds approval): http://mexiconewsdaily.com/news/smugglers-used-drones-since-... Wonder how this new delivery method affects street prices for illegal substances?

My guess is that this article is just indirectly referring to the six UAS test sites that most companies can already test at for interesting UAS projects.

https://www.faa.gov/news/fact_sheets/news_story.cfm?newsid=1...

I still can't figure out how these things actually deliver the package. Do they just drop it on my front lawn? That doesn't seem desirable. Does it drop it at a centralized location because my street has thick tree coverage? Don't most streets have thick tree coverage?

Let us hope that this doesn't turn into a protectionist regulatory nightmare.

first question in my mind; "in exchange for what?"

Google partnered with the government for data sharing (PRISM), so the Government could probably easily get any data that Google will be collecting...

Reply on news.ycombinator.com