Hacker News Reader

Comment by Analemma_ | Hacker News Reader

I don't remember where, but I read a lengthy comment from someone in the industry which said there were two main things preventing delivery drones from being a viable market:

1. FAA regulations - delivery drones can't operate within X miles of an airport (technically they can, but it requires a much stricter degree of certification and compliance nobody wants to bother with)

2. Drones need a landing space, so people without yards (like apartment and townhouse dwellers, who make up a lot of the population in exactly the densely populated areas where you'd want to use drones to begin with) can't be served

And it turns out that once you exclude "houses within X miles of an airport" and "houses without an LZ", there aren't enough customers left to make delivery drones worth it.

View on HN

2 is surprising as that’s the opposite of where I’d expect delivery drones to be useful. You don’t need them in dense urban environments where a stocked truck offloads inventory with low mileage. You want them in rural areas where as the crow flies makes a difference and it’s expensive to have a truck driving about sparsely populated areas.

I saw a demo recently where they had a drone flying high and it zip-lined a package down to the destination directly below it, so the drone never had to get close to the ground. I’m not sure if that’s current strategy but it seems like a cool idea to cut down on noise and proximity issues.

I think this is the popular route they’re going, to avoid drone theft/vandalism, problem is still the liability of dropping things on people’s heads regardless of CV or other sensors.

That’s actually an idea: create a lightweight Ariel tram way just for package delivery. Sort of like the old vacuum tube networks some cities used to have.

This is happening in select parts of Australia.

https://wing.com/en_au/press/australia/

This is how Manna Aero (1) are doing it in Ireland (2) and soon parts of Texas (3).

1. https://www.manna.aero/

2. https://www.youtube.com/watch?v=JCnpxmhWRuw and https://www.siliconrepublic.com/machines/drone-regulation-ir...

3. https://dronelife.com/2023/05/08/manna-drone-delivery-flies-...

The issue there is that the battery tech to support long distance delivery drones does not exist yet.

Interesting, I am woefully ignorant on typical drone ranges.

Winner. Battery capacity needs to double and charge twice as fast to make EV adoption possible.

Swappable batteries makes the charge time a nonissue.

Swappable 1000lb batteries? Gasoline weighs 80lbs.

Now I’m imagining a larger, fixed-wing loitering drone carrying the package(s) and a small last-mile drone that detaches, drops its payload at the destination and then returns to the mothership.

There should be even larger drones that are always in the air that are used as supply hubs for these smaller drones. Actually those should be whole warehouses where people work.

Why not a zeppelin?

(The prospect of delivery drones is silly, but whatever it is fun to think about).

I'd like to see someone try integrating them into the delivery trucks. Roof mount it, make it easy to load from the back storage area, and have it auto charge. That way drivers can choose to hand off smaller packages to it and launch when they are delivering close by.

This is a wonderful idea! I’d love to see this implemented somewhere.

The old "why not both" solution :)

(I'm not being sarcastic. "Why not both?" thinking is powerful!)

Where have I seen this before?

https://wiki.factorio.com/Personal_roboport

And the 'drone spitter' USPS vans in the Ready Player One film.

That's how I've always imagined it. The top of the FedEx truck is a mini-helipad, and there are two people on the truck. As the truck makes its normal route stopping to deliver heavy packages on foot, there are two drones darting back and forth delivering light ones, with someone constantly feeding them packages and batteries.

Somewhere years ago I have a chat log where I describe the whole thing as imagined, including the phrase "print out helipad_qr.pdf and tape it to your driveway" as part of the process. ;)

I would bet money that the cost of

- a second person (pay + benefits + training)

- drones

- helipad

- vehicle idle time for loading the drones

is more expensive than one person just driving the boxes around.

I don't think so.

You're describing the total cost and ignoring the larger amount of packages being delivered.

Cost per package drops.

What delivery men need isn't messing around with slow and complicated drone launches. They need the equivalent of a mailbox, i.e. package lockers in front of every apartment building and in front of every house. Then all they have to do is get off the delivery vehicle, retrieve the package, open the locker, insert the package, close the locker and then go back to the vehicle.

At that point you have reached human peak efficiency and then the only thing you can do is build a self driving delivery van with an integrated crane/robot arm to operate the locker.

Self driving is hard. Building a robot arm that inserts packages sideways into a locker doesn't appear as difficult. The post office can dictate the package formats and add visual markers to make it easier for the robot to grab the right side.

Sounds lethal for cyclists and children on toys to pull out a metal rod over sidewalks and what not.

Just put the mailbox so the postman reaches it from the driver seat window.

No SaaS, apps, drones or VC money needed.

I really doubt that cost per package drops more than adding multiple extra drivers.

The truth is that the wave of “imminent drone delivery” mania was always just marketing, and wasn’t going to make sense except in extremely limited scenarios.

Also the drones take up more capacity in the vehicle, as does the charging station, the batteries, the loading mechanisms. And so does the second person. You can carry far less packages now.

https://www.youtube.com/watch?v=TYBu2glKHTA

I don't remember seeing this demo, but considering it is exactly what I imagined and it's a demonstration from 6 years ago, maybe I did.

It would probably be a lot cheaper and safe to make an autonomous or semi-autonomous wheeled platform that unloads and delivers the packages, sort of a motorized pallet cart, which would be usable also for bigger loads.

The thing that makes rural areas inefficient for trucks (distance between customers) is far worse for drones. Drones have to be based out of somewhere close to the customer due to their short battery life, which means you have to have a lot of bases, each of which only serves a handful of customers. A truck will burn a lot of fuel per customer, but a single one can still serve a large area on one tank and can be based out of a nearby city.

A truck will burn a lot less fuel per kg of cargo. That's because it is carrying many packages at once and fuel consumption is split across all packages. A drone only carries one package. So in reality a truck is a way more efficient way of delivering good.

Seems like the best compromise is something like the "drone spitter" delivery trucks from Ready Player One.

The truck wouldn't need to slow down (let alone stop), multiple drones would give high utilization, and you could further optimize delivery routes by "crow flying" to houses that aren't even on the same street.

https://www.inverse.com/article/43043-ready-player-one-drone...

https://www.imcdb.org/v001094117.html

Anytime you think, "ah, the solution is this thing from Ready Player One", step back and remember the following:

- that book was set in a shitty gig economy dystopia, even if the main character managed to personally get rich in the end

- that book was fucking terrible and it's frankly embarrassing that grown adults make reference to it in a positive manner

You know what’s even worse? The audiobook version, narrated by Will Wheaton.

I had to watch an ad before I could watch that company's commercial.

I suspect this still only works if you can utilize the capacity and have simultaneous deliveries, which would required a cluster of deliveries, thus likely not good for rural areas.

>I suspect this still only works if you can utilize the capacity and have simultaneous deliveries

You can still eliminate inefficient stop-and-go driving for some fraction of stops (ie not U-turns), and (most important) the time-wasting "last 100 feet" walk. Sure it's not as superior in rural areas vs suburban, but with low enough overhead you could still beat a traditional truck.

Naturally if the drone gear cost millions of dollars and took up half the truck, then I'd agree it wouldn't be worth it. In reality I don't think that's the trade.

UPS showed off an interesting "roof mounted" hardware approach (ignoring how they depict of the rest of the ops model). https://www.youtube.com/watch?v=TYBu2glKHTA

The numbers I've seen don't pencil out costwise even without assuming the more fuel efficient hybrid and electric trucks becoming popular.

Can you share why you think it makes sense cost wise?

Who is gonna take the box off the shelf in the truck and attach it to the drone? Boxes are stacked on the truck and move all around, not set neatly spaced on a shelf that doesn't move. You'll need another robot to do that for you.

You're probably right, it's the equivalent of public transportation for cargo. And you still need a truck to deliver the packages to all those drone bases anyway.

One of them also don't need to generate upwards lift. A helicopter is also less efficient than a truck.

A drone must supply its own lift against the pull of gravity.

A truck gets that for free via tyres and tarmac.

Air travel makes sense for high-value and time-sensitive payloads, most especially passengers, but also communications (airmail delivery was, and remains, a significant revenue source for airlines).

Fixed-wing drones (as Zipline uses in Rwanda) are more efficient than rotary craft, as forward velocity generates lift, though at a cost of being unable to hover, which makes point delivery to locations lacking a landing strip (or other capture device) more challenging. Zipine utilises parachutes (and lightweight cargo) for its deliveries. Pivot-rotor craft (which Zipline seems to be experimenting with for urban deliveries) make hovering more viable whilst preserving the efficiency of fixed-wing flight.

I still see this as an option with very limited applicability.

One "hybrid" solution is to base the drones on a truck that drives around the city.

An aircraft carrier aircraft carrier!

that sounds like a fun travelling salesman problem with more moving parts. What is the best route for this drone platform truck, taking into account how efficient it is for each drone to make its delivery, and how the truck’s position will have changed in the time it takes for the drone to return.

That is actually a problem under research; there are a few papers on it and luckily they use the obvious name — TSP with drones or TSP-D.

I gave that problem a very shallow go for a Metaheuristics assignment, and while I didn't really mess with the domain specific heuristics for it, it did seem pretty fun indeed.

2 is wrong and you are exactly right. Amazon explains drones are exactly for meant for rural areas here: https://youtu.be/yMqbj4Kj-z0?t=688

A lone driver might get paid X $/hour but in a rural area, only reach a single house with his truck, whereas the same driver might make 10x deliveries in the same time period in an urban area. And optimally, you'd have one drone pilot controlling multiple drones at the same time with each one taking less time than a car.

I'm guessing the biggest cost here isn't fuel or energy, it's the staffing.

In his point #2, he is saying you can deliver to rural areas, but not to densely populated areas like cities, where there are not enough landing sites. I have no idea whether this is correct or not.

3. Drones, autonomous or not, aren't anywhere near reliable enough to do autonomous remote delivery. Final delivery is fraught with a myriad of things that can go wrong. GPS is not 100% reliable or accurate in urban areas. Failure in flight means both drone and package are lost, with the added bonus of potentially hurting or killing someone, liability for which the insurance company is not going to underwrite because drones aren't reliable. Backup flight modes add to weight and cost, and reduce deliverable payload. And even after all of that, aviation authorities at least in my jurisdiction are extremely conservative and won't approve them for autonomous delivery and flight near anywhere people live.

4. Payloads on flying things have severe weight and volume limitations. This limits what you can deliver to things that are small and light.

5. An autonomous drone represents an opportunity for theft or interference. See the drone coming, then: steal the goods when it delivers, interfere with its GPS and force it into an error mode that lands the drone, etc. You don't even need to be at the destination.

6. Multicopters, the devices most people think of when they think drones, have terrible range and payload capacity. The flight mode is akin to helicopters, "beat the air into submission". On the upside, they can hover on the spot. In contrast, plane-style drones which fly aerodynamically or "on the wing", have more range (being more efficient), but have difficulty landing on specific spots and then returning back to base. There are solutions around this, but not 100% viable in all cases. VTOL planes exist for instance, but they're finicky. Parachutes as a delivery mechanism also exist, but not everything can be delivered by dropping a payload.

And much more...

Air delivery exists in places where these aren't concerns and where the cost benefit analysis skews the other way. Zipline [0] has been doing a great job delivering blood to remote areas in Rwanda for a while now. Their product, payload, delivery method, geographical location, regulatory environment, all align to make it worth it. Watch some youtube videos of their operation, it's pretty neat.

[0] https://www.flyzipline.com/

This. The tech was never there, the hype came from people who didn't understand the fundamental problem, and the secondary stuff about regs and people not having landing zones is cope to disguise yet another failure on the part of the hype zombies.

> 4. Payloads on flying things have severe weight and volume limitations. This limits what you can deliver to things that are small and light.

This point to me seems like such an obvious hard limit on how well this could ever work without some serious advancements in battery tech, and it's why the whole thing seemed like mostly snake oil to me.

The x-copters have very low weight limits (and low flight times when loaded) and realistically this wouldn't be possible for weights over ~10 pounds without very short range limitations and massive downtime for recharging (and/or truly massive drones, which would pose too much of a safety risk in populated areas).

Not to mention the physical shape of even many small / light objects would make stable flight difficult or impossible. The second the wind picks up, they are in serious trouble.

Maybe some small and very specific niches can utilize it, but I suspect in almost all cases (even when it is possible to use these things optimally) driving is actually going to be cheaper overall.

This (as well as range) depends on what size and cost you can accept. You can absolutely get drones that can carry useful loads and/or that have long ranges, but they're large and very expensive.

As someone who used to be in the industry, and living in one of the only US locations with a current UAS delivery program I can help clarify.

> 1. FAA regulations - delivery drones can't operate within X miles of an airport (technically they can, but it requires a much stricter degree of certification and compliance nobody wants to bother with)

Nope. This isn't really the issue at all. We have an airport near by (two if you count the hospital and the FAA does). UAS delivery isn't actually under UAS rules, it's under commercial freight delivery rules. This is how Project Wing/Google operate because the FAA still doesn't allow beyond line of sight operations for UAS.

> 2. Drones need a landing space, so people without yards (like apartment and townhouse dwellers, who make up a lot of the population in exactly the densely populated areas where you'd want to use drones to begin with) can't be served

Sorta. They don't actually land, they drop the payload while hovering. So they do need some uninstructed space to drop the package, but not much. Works well in the suburbs and rural areas but awful in the city.

> And it turns out that once you exclude "houses within X miles of an airport" and "houses without an LZ", there aren't enough customers left to make delivery drones worth it.

Again, not an airport problem.

It's really economics. It's v expensive per gram to deliver by drone. That's why it works well for light but high value things (like medicine) but sucks for stuff like food.

> It's really economics. It's v expensive per gram to deliver by drone. That's why it works well for light but high value things (like medicine) but sucks for stuff like food.

I always suspected the reason was purely profit-driven so it's good to hear it confirmed.

I know I'm getting cynical as I age, but the idea that regulation was the only thing in the way of it seemed laughable given recent history.

It's certainly not the only thing, but it's also not helping. It's an area where we need some more research into better power efficiency systems for cost to improve. But regulations are not encouraging investments in the space as much as you would think.

Also there is a strong education and hobby to industry pipeline. However with the way the UAS space is being regulated that pipeline has been drying up.

> Sorta. They don't actually land, they drop the payload while hovering. So they do need some uninstructed space to drop the package, but not much. Works well in the suburbs and rural areas but awful in the city.

What's the plan there? Leave my package out in the middle of the yard to get rained on or stolen?

So the way project Wing works is you order through an app and track the delivery. Delivery is within a few minutes so the lag time is fairly low.

I don't see this system being used to deliver stuff not being specifically requested (like say an Rx). Strong winds, fog, and rain would ground delivery also.

> And it turns out that once you exclude "houses within X miles of an airport" and "houses without an LZ", there aren't enough customers left to make delivery drones worth it.

That would make sense if this was some incredibly capital-intense thing to develop and enjoyed immense economies of scale. But I don't see why that would be the case. Surely there are many thousands of towns each with tens of thousands of single-family homes that are sufficiently far from an airport and currently served by daily package deliveries via automobile. Unless we're considering every tiny community airport.

Also in the LZ column: trees and power lines. From above, looking down, power lines are almost impossible to spot for human pilots. A CV model steering down from 400 feet would need to avoid such.

Forget Drones, even the land robots [the ones that look like small moving pillars] haven't been put to actual use - the companies that made them or built software for them ended up raising good cash and exiting by getting sold to other big companies in the logistics space, but no one knows for sure what their future is, lol.

https://starship.co/ is in actual use in a few cities, but they have not exactly taken over the logistics space yet.

1 is a bit surprising. It’s not like housing around airports are particularly popular for noise reasons. Here in Seattle, if you don’t count float planes, we don’t even have general aviation airports to worry about until you hit Boeing field to the south and Paine field in the north (again Boeing related).

Most new town homes have rooftop decks (because Seattle doesn’t get much snow), why can’t they land on those?

Renton Airport, Boeing Field, Seatac, Paine Field Harvey Field, Arlington Field, plus scores of smaller poorly marked airstrips if you check OpenStreetMap. It turns out very few places in the US are sufficiently far away from an "airport" to allow drones to operate. Even out in the boonies.

There is literally nothing in Seattle between Paine field and Boeing field in the main I5 corridor if you don’t count sea planes (all the airport symbols along the lake are float planes, Boeing field, SeaTac, and Renton are all basically right next to each other). It is a pretty air strip dead zone, you had to go all the way out to Arlington to mention another one (they have a nice playground next to it that the kid likes to watch planes from)! And ya, the boonies have landing strips everywhere (like the one out near Snohomish used for sky diving).

You're clear with no restrictions from about Downtown to North Lynnwood. Even then, as you approach Paine and SeaTac, you can fly with altitude restrictions (300, 200ft, etc). On the eastside along 405, permissible areas are even greater.

https://b4ufly.aloft.ai/?lat=47.124171412997555&long=-122.39...

I always thought of it as a luxury, instant delivery thing. Like if you want your package on 15 minutes, pay $50 extra and someone will pilot it for you. And call your phone to run outside after finding the nearest LZ, if they have to. You know, like do real work. I guess that's not what they had in mind?

Also I imagine there's still a huge risk of the drones malfunctioning, being shot down, messed with, disrupted, breaking due to weather, etc.

Currently work at company that was working on delivery drones.

3. Drones kept getting shot down during testing

Not kidding

Reply on news.ycombinator.com