10 ms·

The whole "drones will be used to deliver everything" shtick that's coming out of the USA really does not resonate in the UK or Europe, where most folks live in

by cstross 11y ago

The whole "drones will be used to deliver everything" shtick that's coming out of the USA really does not resonate in the UK or Europe, where most folks live in extremely dense (by American standards) cities -- US suburbia is equivalent in density to open countryside, and many if not most people live in apartments or houses that lack any kind of garden (yard) or garage. (And typically have convenience stores within a hundred metres' walk of the front door, more to the point -- which can already be used for deliveries.) I'm basically calling bullshit on the extreme formulation of drone delivery. There will be some niches that they colonize, and a few that they monopolize (they're going to be a boon for postal/small package deliveries to farms/exurban housing -- Post Office self-driving van goes past without stopping while drone shuttles packages to doorsteps a quarter of a kilometer away) but plenty of areas where they're going to be forbidden due to the unacceptable risk of them falling on someone's head. Here's maybe a better use: Put weatherproofed power sockets with fiber connections on rooftops every 50 meters in urban areas. Then hang your 5G picocells from drones, and dynamically re-structure your mesh of cellphone bases as demand shifts through the day (by physically lifting and transferring the base station to a new fiber/power feed as necessary). Goal being to feed a maximum number of radio channels where demand is high, while avoiding black spots.

anabis 11y ago
Yes, in Japan we've had near misses from careless operators. https://youtu.be/uvA6dReF6-I?t=291 https://youtu.be/uvA6dReF6-I?t=291
golergka 11y ago
> unacceptable risk of them falling on someone's head Same could be said about airplanes 100 years ago, but now I have planes flying over my head about 20 times a day. Even in the post-9/11 world. The safety of drones will only increase, and with such a great financial incentive as the whole delivery industry, it will do so rapidly.
- cstross 11y ago
  The safety of airliners comes at a cost -- operating expenses (a lot of which is for maintenance) running so high that they're uneconomical except for bulk/public transport or the hyper-rich. Private flying cars they aren't. Small delivery drones, almost by definition, are your "flying car" paradigm, downsized to just the cargo compartment because they don't need to carry 70-300kg of shaved apes. Maintenance and QA is going to be a big issue, and will drive up prices.
- GFK_of_xmaspast 11y ago
  Small private planes do tend to crash into houses and golf courses and IRS buildings semi-regularly tho.
  - icebraining 11y ago
    Which due to pilot error, though, not technical problems. The kind of delivery drones we're discussing will most likely be auto-piloted.
- falcolas 11y ago
  > The safety of drones will only increase I hate to be a negative Nancy about this, but [citation needed]. I fly a multi-rotor coptor now, and when the battery hits a certain point it just drops out of the sky, wherever it might happen to be. This is just a toy model, so it doesn't have some of the fancier electronics a delivery drone might - but there's nothing inherently stable about a multirotor like there is with a plane. Even the best motor failure "recovery" I've seen so far is still a multirotor going into a barely-controlled death spiral - something which would still cause a lot of damage to whatever it hit. Imagine it for a moment: a 20-30 pound object falling in an unpredictable (to humans) spiral at speeds faster than you can run, with between three and seven 10 inch props moving at near supersonic speeds at their tips. I love 'em, but they're dangerous machines with lots of fast-moving parts, and need to be treated as such.
  - kenrikm 11y ago
    I fly and build a great many multi rotors and planes myself. And with current tech you can have it land itself when it gets to a certain voltage. No reason to have it falling out of the sky due to a dead battery.
  - icebraining 11y ago
    While not exactly inherent, there's plenty that can be done to further secure their falls. Firstly, is to have automatic auto-landing when the battery is getting low (with a good safety margin, of course). The other is what DJI is already exploring with their "DropSafe", ie., parachute. A simple software tweak (when accelerometer detects falling, shut down motors and deploy parachute) could make most falls reasonably safe (and less costly to the owner). With cameras, they could also try to avoid flying over humans. Finally, in cities, they could fly most of their way over the roofs of building, rendering a fall unlikely to hit anyone.
  - agumonkey 11y ago
    Even without motor failure, if control is jammed, your toy goes rogue. A frenchman showed a video were his drone would get lose contact, get confused, and start to seek for it's charging base where there was none, flying over a motorway full of cars. It could have landed there. It went further, almost crashed into a skycraper (wind dynamics pushed it away IIUC) then landed on the curb.
ghaff 11y ago
The thing is that I suspect the biggest limitation of rapid delivery to exurban and rural locations is insufficient density for distribution depots as opposed to last mile delivery. Where I live is considered exurban but, like the vast majority of the US, I have absolutely no issue with receiving daily deliveries from USPS, UPS, and any number of other delivery services. Indeed, I can get lots of things delivered the next day if I'm willing to pay for it. It's true that I can't get Instacart of some of the other same-day delivery services found in urban areas but then there's not enough density of the right kind of customers to support a Whole Foods store either. (I can get Peapod though I don't much care for it.) All of which is a long way of saying that I don't really see the non-niche problem that even an idealized drone delivery system aims to solve. >Post Office self-driving van goes past without stopping while drone shuttles packages to doorsteps a quarter of a kilometer away) Even this is largely a solved problem. Private roads in rural areas usually have some sort of lockbox at the end of the road.
- jessriedel 11y ago
  > Indeed, I can get lots of things delivered the next day if I'm willing to pay for it. >All of which is a long way of saying that I don't really see the non-niche problem that even an idealized drone delivery system aims to solve. Part of the idea is that you'll get same day delivery for a reasonable price. In fact, it could be cheaper than normal stores (although probably not walmart) since it'll just be a warehouse with a drone fleet.
  - ghaff 11y ago
    Here's the thing. There are very few items that I care (or that most people care) about next day delivery for. Groceries may be one thing--and drones won't help for that even in an optimistic scenario. I'm all for home delivery and am a big user of Amazon Prime. I'm just unconvinced that, even if one waves away the myriad issues, drones so drastically undercut the costs of human drivers as to create interesting new delivery models in the next decade or two.
    - jessriedel 11y ago
icebraining 11y ago
It's interesting comparing your post with the reaction to widespread cars in the start of the 20th century: http://99percentinvisible.org/episode/episode-76-the-modern-moloch/ http://99percentinvisible.org/episode/episode-76-the-modern-... The horrors of peace appear to be appalling than the horrors of war. The automobile looms up as a far more destructive piece of mechanism than the machine gun. The reckless motorist deals more death the artilleryman. The man in streets seems less safe than the man in the trench. The greatest single lethal factor is the automobile. It left shambles in its wake as it coursed through 1923. We know how that ended up.
- jules 11y ago
  Some countries have since recognized that mistake and have been successful in getting cars out of the city center.
- cstross 11y ago
  Actually, the carnage caused by primate-driven cars exceeds that of the machine gun; we lose around 2 million dead or seriously injured every year, a death rate rivalling that of the first world war, on an ongoing basis. (Self-driving cars can't come soon enough; 90% of RTAs are the result of human-driver error ...)
  - icebraining 11y ago
    I completely agree; my point was that cars are widespread despite the killings, so the potential for drones to fall on people's heads doesn't seem to be enough to prevent them from being adopted, if there's enough lobbying from the people running them.
    - robotresearcher 11y ago
      If they are so useful that the dangers are considered a worthwhile tradeoff.
  - kenrikm 11y ago
    I think that if cars were invented today they would most likely be made illegal due to the potential for death (40k/year in the US). Aside from a handful of freak accidents drones don't kill people yet there are a great many people trying to ban them out of the the fear that they might. Now compare that to the 40k number for cars. Drone deliveries have the potential to remove cars from the road thus saving lives.
kenrikm 11y ago
Imagine This: Drones could fly over medians of surface streets so if they do fall the only thing they are likely to hit is a car or a very, very unlucky pedestrian with some additional hardware and a connection to ATC (Think ADS-B 2.0) you would not need to worry about hitting other aircraft. And instead of landing at 1000 unknown landing points have them shuttle packages between a few known landing points. I.E go down to the corner Walgreens drone landing zone (rooftops or parking lots employees get the package from the drone drop it in a drop box for you to pick up and swap batteries for the next flight) Then you just head down to pick up whatever was just flown 10 miles to your area for you instead of fighting rush hour traffic, thus removing a car from said traffic and making the world a better place. The last mile becomes the last 100 yards. The point is there are solvable problems that can make drones viable. I've done a great many long distance autonomous drone flights (we're talking miles) to then have it land within 3ft of where it was supposed to. The tech is very close to being up to the task (3-5 years) the regulation however is not. Food for thought Cars kill 40k people per year in the US, even if a few are killed by drones the number of deaths avoided by having less car trips would likely well more then offset those killed by falling drones. One more thing, if cars were invented today they would most likely be made illegal due to the potential for death (40k/year). Aside from a handful of freak accidents drones don't kill people yet there are a great many people trying to ban them out the the fear that they might. Now compare that to the 40k number for cars.

by cstross 11y ago

anabis 11y ago

Yes, in Japan we've had near misses from careless operators. https://youtu.be/uvA6dReF6-I?t=291 https://youtu.be/uvA6dReF6-I?t=291

golergka 11y ago

> unacceptable risk of them falling on someone's head Same could be said about airplanes 100 years ago, but now I have planes flying over my head about 20 times a day. Even in the post-9/11 world. The safety of drones will only increase, and with such a great financial incentive as the whole delivery industry, it will do so rapidly.

cstross 11y ago
The safety of airliners comes at a cost -- operating expenses (a lot of which is for maintenance) running so high that they're uneconomical except for bulk/public transport or the hyper-rich. Private flying cars they aren't. Small delivery drones, almost by definition, are your "flying car" paradigm, downsized to just the cargo compartment because they don't need to carry 70-300kg of shaved apes. Maintenance and QA is going to be a big issue, and will drive up prices.
GFK_of_xmaspast 11y ago
Small private planes do tend to crash into houses and golf courses and IRS buildings semi-regularly tho.
- icebraining 11y ago
  Which due to pilot error, though, not technical problems. The kind of delivery drones we're discussing will most likely be auto-piloted.
falcolas 11y ago
> The safety of drones will only increase I hate to be a negative Nancy about this, but [citation needed]. I fly a multi-rotor coptor now, and when the battery hits a certain point it just drops out of the sky, wherever it might happen to be. This is just a toy model, so it doesn't have some of the fancier electronics a delivery drone might - but there's nothing inherently stable about a multirotor like there is with a plane. Even the best motor failure "recovery" I've seen so far is still a multirotor going into a barely-controlled death spiral - something which would still cause a lot of damage to whatever it hit. Imagine it for a moment: a 20-30 pound object falling in an unpredictable (to humans) spiral at speeds faster than you can run, with between three and seven 10 inch props moving at near supersonic speeds at their tips. I love 'em, but they're dangerous machines with lots of fast-moving parts, and need to be treated as such.
- kenrikm 11y ago
  I fly and build a great many multi rotors and planes myself. And with current tech you can have it land itself when it gets to a certain voltage. No reason to have it falling out of the sky due to a dead battery.
- icebraining 11y ago
  While not exactly inherent, there's plenty that can be done to further secure their falls. Firstly, is to have automatic auto-landing when the battery is getting low (with a good safety margin, of course). The other is what DJI is already exploring with their "DropSafe", ie., parachute. A simple software tweak (when accelerometer detects falling, shut down motors and deploy parachute) could make most falls reasonably safe (and less costly to the owner). With cameras, they could also try to avoid flying over humans. Finally, in cities, they could fly most of their way over the roofs of building, rendering a fall unlikely to hit anyone.
- agumonkey 11y ago
  Even without motor failure, if control is jammed, your toy goes rogue. A frenchman showed a video were his drone would get lose contact, get confused, and start to seek for it's charging base where there was none, flying over a motorway full of cars. It could have landed there. It went further, almost crashed into a skycraper (wind dynamics pushed it away IIUC) then landed on the curb.

ghaff 11y ago

The thing is that I suspect the biggest limitation of rapid delivery to exurban and rural locations is insufficient density for distribution depots as opposed to last mile delivery. Where I live is considered exurban but, like the vast majority of the US, I have absolutely no issue with receiving daily deliveries from USPS, UPS, and any number of other delivery services. Indeed, I can get lots of things delivered the next day if I'm willing to pay for it. It's true that I can't get Instacart of some of the other same-day delivery services found in urban areas but then there's not enough density of the right kind of customers to support a Whole Foods store either. (I can get Peapod though I don't much care for it.) All of which is a long way of saying that I don't really see the non-niche problem that even an idealized drone delivery system aims to solve. >Post Office self-driving van goes past without stopping while drone shuttles packages to doorsteps a quarter of a kilometer away) Even this is largely a solved problem. Private roads in rural areas usually have some sort of lockbox at the end of the road.

jessriedel 11y ago
> Indeed, I can get lots of things delivered the next day if I'm willing to pay for it. >All of which is a long way of saying that I don't really see the non-niche problem that even an idealized drone delivery system aims to solve. Part of the idea is that you'll get same day delivery for a reasonable price. In fact, it could be cheaper than normal stores (although probably not walmart) since it'll just be a warehouse with a drone fleet.
- ghaff 11y ago
  Here's the thing. There are very few items that I care (or that most people care) about next day delivery for. Groceries may be one thing--and drones won't help for that even in an optimistic scenario. I'm all for home delivery and am a big user of Amazon Prime. I'm just unconvinced that, even if one waves away the myriad issues, drones so drastically undercut the costs of human drivers as to create interesting new delivery models in the next decade or two.
  - jessriedel 11y ago

icebraining 11y ago

It's interesting comparing your post with the reaction to widespread cars in the start of the 20th century: http://99percentinvisible.org/episode/episode-76-the-modern-moloch/ http://99percentinvisible.org/episode/episode-76-the-modern-... The horrors of peace appear to be appalling than the horrors of war. The automobile looms up as a far more destructive piece of mechanism than the machine gun. The reckless motorist deals more death the artilleryman. The man in streets seems less safe than the man in the trench. The greatest single lethal factor is the automobile. It left shambles in its wake as it coursed through 1923. We know how that ended up.

jules 11y ago
Some countries have since recognized that mistake and have been successful in getting cars out of the city center.
cstross 11y ago
Actually, the carnage caused by primate-driven cars exceeds that of the machine gun; we lose around 2 million dead or seriously injured every year, a death rate rivalling that of the first world war, on an ongoing basis. (Self-driving cars can't come soon enough; 90% of RTAs are the result of human-driver error ...)
- icebraining 11y ago
  I completely agree; my point was that cars are widespread despite the killings, so the potential for drones to fall on people's heads doesn't seem to be enough to prevent them from being adopted, if there's enough lobbying from the people running them.
  - robotresearcher 11y ago
    If they are so useful that the dangers are considered a worthwhile tradeoff.
- kenrikm 11y ago
  I think that if cars were invented today they would most likely be made illegal due to the potential for death (40k/year in the US). Aside from a handful of freak accidents drones don't kill people yet there are a great many people trying to ban them out of the the fear that they might. Now compare that to the 40k number for cars. Drone deliveries have the potential to remove cars from the road thus saving lives.

kenrikm 11y ago

Imagine This: Drones could fly over medians of surface streets so if they do fall the only thing they are likely to hit is a car or a very, very unlucky pedestrian with some additional hardware and a connection to ATC (Think ADS-B 2.0) you would not need to worry about hitting other aircraft. And instead of landing at 1000 unknown landing points have them shuttle packages between a few known landing points. I.E go down to the corner Walgreens drone landing zone (rooftops or parking lots employees get the package from the drone drop it in a drop box for you to pick up and swap batteries for the next flight) Then you just head down to pick up whatever was just flown 10 miles to your area for you instead of fighting rush hour traffic, thus removing a car from said traffic and making the world a better place. The last mile becomes the last 100 yards. The point is there are solvable problems that can make drones viable. I've done a great many long distance autonomous drone flights (we're talking miles) to then have it land within 3ft of where it was supposed to. The tech is very close to being up to the task (3-5 years) the regulation however is not. Food for thought Cars kill 40k people per year in the US, even if a few are killed by drones the number of deaths avoided by having less car trips would likely well more then offset those killed by falling drones. One more thing, if cars were invented today they would most likely be made illegal due to the potential for death (40k/year). Aside from a handful of freak accidents drones don't kill people yet there are a great many people trying to ban them out the the fear that they might. Now compare that to the 40k number for cars.