Hacker News Reader

Noise pollution may be a reason drone delivery _shouldn't_ happen, but don't confuse that with an actual practical obstacle. In the US, at least, the airspace (including drones) is regulated at a federal level by the FAA, and their mandate doesn't really include residential noise levels. For example, I've never seen noise levels discussed when it comes to private aviation, even though small planes can _definitely_ be loud enough to affect quality of life (they don't even have mufflers...) Drone regulations seem to be following a similar trajectory, with categories and rules based on mass and prop guards, and no formal measurements of perceptual noise at all.

For stationary noise concerns like a highway being built, it's possible for local communities to rally and be heard, but it's entirely possible that drones will be regulated purely on safety grounds with minimal conversation about noise levels.

This is the point where I should end with a call to action, but I'm honestly not sure what the path is to resolving this sort of structural blindness. If anyone with experience working in government or advocacy wants to chime in, I'd be curious to hear.

noise pollution laws already exist. why wouldn't drones be subject to them?

The FAA's immunities do not apply to noise sources less than 400 feet above non-airport property.

But the FAA does not have rules against it, nor do many localities.

> " For example, I've never seen noise levels discussed when it comes to private aviation,..."

Then you must not have looked very much. Noise was the primary reason cited why the Concorde was not allowed to travel supersonic above the continent. But if you want something fresher then you only have to look at the Airport Noise and Capacity Act (ANCA) of 1990. Or the many airports where they have nighttime curfews because of noise concerns.

It is also an everyday concern while you are learning to fly. I was doing my PPL in Europe, so my anecdote is not directly related to FAA, but I remember a ton of "local rules and customs" around noise. Think about things like "at this airport we only do engine run-up at the far end to not disturb those people", or "at this (other) airport we only use the left pattern to avoid spooking the horses on that field there"

> Every 6 decibels sound aptitude doubles so 70-80db is a big difference.

It’s every 3db the sound intensity doubles. At 6db, the sound pressure level doubles. Loudness (aka the perceived volume to a human) however only doubles every 10db, so the difference in this context is 80db is 2x louder than 70db.

Not the best link, but skimming it it seems to do an ok job explaining the difference:

http://www.sengpielaudio.com/calculator-levelchange.htm

It correctly explains that loudness is an arguable concept, that depends as much on the subject as it does on other qualities of the sound (for example the frequency makeup)

Thanks for the link and the explanation.

Yup! Happy to help. To further break it down:

sound intensity increase of 10db: 10x increase

sound pressure level increase of 10db: ~3.5x increase

loudness increase of 10db: ~2x increase

It sucks that there’s so many variations that all use decibels as their unit of measurement with different rates of change. I had to research it years ago because I’d always been told 3db = 2x, yet found it odd that 3db didn’t seem 2x louder. Now I understand.

Yeah, I don't see why drones and vans would be mutually exclusve. You drive a van down the main road and have it spin off a fleet of drones for individual residential deliveries. They do their thing in 1-2 minutes - far less time than it would take for the driver to get out of the van and walk the package to the door - and then return to wherever the van is currently driving to recharge and pick up the next delivery. The drones never need more than a 1-2 mile range, they recharge while the van is going to the next neighborhood. You avoid all the congested residential streets, the need to find parking or double-park, the risk of hitting kids that run out into the street to catch their ball, the difficulty finding houses, plus all the hard work drivers need to do to walk up driveways and carry heavy packages. Also is very amenable to self-driving trucks, since the delivery van only needs to navigate the main arteries and they have drones acting as the capillaries.

Truck + driver means you have a real person who can kick boxes around and find stuff in the back.

Add drones and you have to create an automated drone loading system which works on multiple drones simultaneously.

Otherwise you have truck + 2 (3? 4?) drones + person, and the person loads one drone while the other is/others are delivering.

That will save some time, but not enough to be a game changer. There will still be kicking of boxes.

Meanwhile drones will have to autodock to a charging cradle, land, be loaded, deliver, and then either repeat or autodock again.

The system breaks down badly if there's any kind of issue.

>and then return to wherever the van is currently driving to recharge and pick up the next delivery

Are battery charging times fast enough to support this? I thought we're still at the "5 minutes flight = a couple of hours charging" stage, but I haven't used anything more advanced than one of those miniature toy quadcopters so I'm not sure how more serious drones (DJI etc.) handle it.

They simply have easily replaceable batteries. For a major event, they’d have lots of batteries charging.

Simply worse for the environment?

Replaceable doesn't mean one-time-use necessarily - it might just mean that the recharge time is longer than the downtime expected between deliveries so you'd need to swap out a new battery after each run while you recharge the used one. There are a lot of assumptions here but it seems sane.

That just raises more questions - do you have some complicated automatic battery swap system, or have an extra human on-board to manage the drones? At least if it was a person they could also load packages onto the drones as you drive, else you'd need to automate that as well.

Do you charge batteries on the van, or just swap trays of them each time you return to the depot?

How much of the van's internal volume are you trading for drone-related subsystems rather than cargo storage?

It seems impractical.

if you have developed drones with the ability to return to a van and autonomously pick up the next package to deliver, picking up a battery at the same time shouldn't be a difficult step on top of that.

there's lots of things about drone delivery that seems impractical, but not sure that battery swaps are one of those things.

You can't build the battery-changing system into the drone due to size, power, and weight constraints, so it has to be part of the van. So you need something on the van that can

1) Secure the drone in place while the battery is being changed, and ensure drone can successfully land with sufficient precision for the battery changer to work, or to manoeuvre the drone into place.

Probably you would also want to keep supplying auxiliary power to the drone so you don't have to wait for it to reboot and reconnect to the van's swarm management when the battery is removed.

2) Physically remove and replace the battery, including manipulation of any connectors and fasteners required.

3) Convey batteries between the drone and the storage inside the van. I imagine this would work something like an LTO tape library, where the van tracks which storage slots are occupied, if they've been used or not, and have some internal mechanism which moves batteries between the storage slots and the battery changing mechanism.

I can see this being quite bulky, as with that many batteries you would need compartmentalisation and maybe even a miniature fire-suppression system in case one of the batteries "cooks off".

4) Must not interfere with whatever cargo-loading system is in place - you could have a nice flat battery tray in the roof, but you might need to punch a hole through it to fit a cargo elevator or so on.

Also, must not take up too much internal volume of the van that could be better used for profitable activities (storing cargo for delivery)

5) It would still have to comply with normal road requirements such as maximum heights, environmental resistance (rain, dust, etc.) and so on.

That's just off the top of my head, I'm not an engineer and I'm sure there are plenty more challenges that I haven't thought of.

I think it's easy to do #1 by having a dock station with the van actively reaching out to grasp and control the movements of the drone. For #2 I think we've already solved that by just having some sort of cartridge battery with exposed contacts - think something closer to an old flip-phone battery that is forced to slide into place correctly. For #3 if we assume the van is equipped to manipulate drones into properly docking the drone this really isn't a stretch - and inventory retrieval systems have been realistic long enough that they're actually featured in Hackers (1995)[1]. Regarding #4 I guess this might be a problem somehow? - but let's just assume the battery storage bank is on an external wall of the van - if security is an issue we're going to lose like a foot or maybe a foot and a half of the internal space to the retrieval system - again I don't think this is a problem. Lastly onto #5 I could see weight being possibly the most difficult road req to comply with here - batteries are heavy and generators are heavy - that's a lot of weight and ideally these vans would be able to drive on any road out there - including those with old wooden bridges or "seasonal pavement".

I think of your points above - #1 is the most difficult in that step you're still needing to contend with the environment which may involve strong winds - once you dock to the van (assuming you can limit yourself to a single docking station or a small number) everything else seems pretty trivial... That all said I don't program autonomous drones for a living so it'd be ridiculously hard for me - but I assume if we can get drones to reliably do pre-programmed flight paths we can probably tackle this.

Most of the problems I can think of are on the other end where stupid users will do all sorts of non-compliant things with the reception pads - like placing them in impossible to reach areas due to locked doors or too little clearance to actually get a package on the target - also just like... keeping those things in a functional condition.

1. https://youtu.be/2efhrCxI4J0?t=111

The answer to a lot of these problems is to flip the script and have the drones homebase be the delivery location or a nearby substation rather than the van itself.

That's also a really interesting idea - though the substation one probably. Right now drones capable of lifting a box of cans are well over 1k which is an uneconomical one time expense for slightly easier package delivery... and extremely uneconomical if occasionally some neighbor is going to hit your drone with a frisbee or if drone theft becomes more profitable than package theft (which it almost certainly would be).

While it's not meant for delivery, Airobotics has/had a very slick battery swapping mechanism for their multi-rotors.

https://www.youtube.com/watch?v=r3QWtXjWJTA

Battery seems eventually solvable with a swappable system.

The thing I don't see is "getting to the front door". As a machine problem, it seems so insanely ambiguous.

From a technical standpoint, a workable system might be to have an official "drone pad" of standard size and appearance, similar to a helipad. You might put one in your backyard if you have space, or if you live in a city block there's probably room for a shared one on the roof.

Of course, the social logistics of that would need a bit of work. It's one thing to make sure everyone's within walking distance of a drone pad, but it's another to make sure nobody gets their parcel stolen.

Not bad, definitely have to start somewhere. Rain would be a big concern though.

Sadly, people in apartments, those that lack a back yard (my parents bought a house w/o one so there's no maintenance, just a ravine) are possibly worse off since theres no "hiding" the package. Its definitely possible to say it's not for everyone though.

While I don't believe particularly drone deliveries as imagined today is a great idea, the drop off should be solvable with an physical token that you place where the goods should be dropped off. Something like an RFID-equipped mailbox.

The location seems solvable, the routing to get to that location not such a simple problem in many urban layouts, especially if people aren't very sensible about where on their property they locate their delivery box, or optimise for having their parcel delivered to locations which are sheltered from the rain

It's only today I read a story suggesting that delivery bikes (electric) would be better for the environment and throughput than delivery vans:

https://www.theguardian.com/world/2021/aug/05/cargo-bikes-de...

This only holds, if it holds at all, for the delivery of low-weight packages in the city centre of major metropolises like London. And everyone in London already knows it's often faster to cycle than drive short distances.

Glad to see this. I've been making the sound pollution argument since the original Amazon commercial.

My lab neighbors a drone research lab. The noise from a single drone is surprisingly distracting even with the door shut. The HOAs would be in the streets the first week.

Volume aside, a lot of the annoyance comes from the unnatural sound of four motors running at slightly different speeds, going in and out of tune with one another.

Unfortunately, noise pollution is at the very bottom of public agencies' priority list.

Even in Europe, with stricter regulators and less tolerance towards loud and obnoxious behavior, it's way too common to hear tuned motorcycles blasting 110dB engine screeches at 2AM, keeping everyone in the neighborhood awake. In theory it's not allowed, in reality the bikers can do it with impunity any time they please. Drones would at least provide some beneficial service.

It's a lot easier to get a small handful of delivery companies to comply with noise pollution regulations than it is to get a large number of motorcyclists to comply. It's also easier to write very strict noise pollution standards for new technology like drones than it is to write very strict standards for existing popular technology like cars and motorcycles.

Workhorse has full on plans for drone delivery that use the van as a base:

https://workhorse.com/horsefly.html

A drone's only in earshot for 10-15 seconds if it's doing a drop off, and will usually be flying high enough to be inaudible. At 200m you can barely even see them.

200m high object you can't hear or see loses power....

> At 200m you can barely even see them.

But you can certainly still hear them.

Do you actually know how high 200m is? Hint: If you've seen a consumer drone it's probably not at 200m because in most regions their ceiling is 120m.

Yes, I do know how high 200m is.

The major expense for deliveries is the man-hours needed, NOT the truck (or drone). You also get time savings from traveling in a straight line between point A and point B.