Stanford lab envisions delivery drones that save energy by taking the bus
venturebeat.com
venturebeat.com
(This of course is ignoring what happens if something fucks up and suddenly it has 1/8th of a second to decide to fly out of it or brake those giant swinging carbon fiber blades before they encounter human flesh. It's like a trolley problem on a trolley.).
No way. (IMHO of course)
(e.g
Video of a Matrice 600 with a payload of ~8lbs https://www.youtube.com/watch?v=jayVMmvLxOA
Even better, 25 lbs from 60' up lol https://www.youtube.com/watch?v=jULQrdUTngM
)
Another mitigation could be to specify allowed jump points (for example on big street crossings where the noise from cars would drown the drones' noise anyway).
If the difference is big enough it's like adding 0.01 to 10. You'd need a lot of drones to influence the total noise level noticeably.
From my experience (Fimi 8 SE ~ 300g drone) the drone is annoyingly loud at < 5meters, and becomes practicaly silent at >50 meters (at least I don't hear it anymore).
Though I reckon it's more likely that the constant noise of drones will drown out the traffic noise than the other way around. Not just because it's constant but also because it's LOUD.
It’s like when a car honks or someone whistles on the highway. There is a lot of noise, but the higher pitches are distinct.
That said I estimate a drone to be much louder than regular traffic, at least inside cities.
I'd imagine there's some overlap of "people who race cars" and "people with aftermarket exhausts."
Now imagine that instead of returning something back to a central warehouse, Amazon could have the next customer pick it up or ship it nearby.
Hopefully boxes would not be needed for most of this, too. And any online store(s) could do this.
But I would also love if they could drone the package over to my roof if I wanted to.
It's also something only the fitter half of the population can realistically do.
The critical issues of safety etc seem like a much bigger deal to me.
https://www.hollyspringsnc.us/DocumentCenter/View/23698/Soun...
So say 4m overhead 3m away (5m distance) you have ~80db
https://www.omnicalculator.com/physics/distance-attenuation
Add in the fact that you would feel the prop wash from it and it's going to be a fairly disturbing experience.
Oddly terrifying. Could be a trailer to a new horror film - The Drone Chainsaw Massacres.
Complex problem for sure.
See, drones traded moving parts for the annoying whiny motors(but fixed in a rigid frame). If we are going to use moving parts anyway, there's helicopters.
Perhaps the lead-lag hinges can be elastomeric and one could purchase props with different fixed hinge stiffnesses for different amounts of noise dampening or frequency spread but perhaps at the cost of prop-speed change response? There'd probably have to be some compensatory mechanism in the flight- or motor-control software that'd be set for a given prop's hinge elasticity.
Of course, I am just idly speculating here on whether there's any benefit at all to any of my blather.
There's someone in this thread saying that their F150 is less noisy than a drone. That is unlikely to be true, but the emitted frequencies are very different. A high pitch whine is, for whatever reason, very noticeable, compared to a lower pitch of similar amplitude. Think dentist drills.
Rotor guards exist. They reduce efficiency, but if this is a major concern, it can be done.
Not much can be done today about the noise, short of not using drones and switching to helicopters, with larger, but slower moving blades (more difficult to guard them too).
Joking aside, that does raise the idea of using a delivery truck as sort of a carrier for rechargable drones; truck drives around, packages are automatically distributed to a pool of drones who handle the last hundred yards of delivery and return to the truck to reload/recharge.
Technically anybody right now could just shoot a poor Amazon delivery drivers with a shotgun now to rob them but it would be murder as opposed to destruction of property. Very few would do so vs snaring or shooting a drone. It is one hell of a bizzare systemic perverse incentive - putting a human body to be at risk reduces the robbery incentive and frequency. Reminds me of one account of Italian medieval cities of merchants plastering their walls with the Virgin Mary so people wouldn't dare to blasphemously piss on her image.
Not to worry... Drones are corporate property and corporations are people, therefore destroying the drone is murder. Don't agree? Take it up with these heavily armed protection drones we built for the delivery drones.
I love dogs, but think their time in LE are fading. Both as an affront to the rights of the people and of the dogs. The 'beast of burden' ethic is quickly becoming outdated.
Make a blimp mothership and keep it under 400 ft (FAA class G).
You save energy by dropping drone packages down the gravity well rather than flying it up from ground-based transport.
However, this is likely another one of those "the exceptions are the norm" problems where the exceptional cases (long packages, unusual elevators, etc.) happen so often that it's just better to keep a human in the loop.
Of course the size/weight of some packages delivered via amazon would be an issue in terms of using a drone at all. Not sure about the energy use for flying drones... Land based drones could also work for some deliveries.
not having to stop the truck would be a huge carbon saver, would it not?
Drones are a visual nuisance and produce noise. Do you want to disturb bus riders and residential areas.
The Silicon Valley episode where a Stanford lab produced a robotic deer that is completely useless but caused a road accident is reality.
I hope people will react like Erlich Bachman when these drones become a thing.
2. I didn't say there would be micro-farms and micro-ranches. I don't believe in "village-level" autarchy. But in my view there should be much less ferrying of food from halfway across the world because people work there for a couple of USD per day and ship fuel is (still) cheap.
3. My point was that it's a huge waste of social resources, not to mention fuel, to send out vehicles across town to deliver food. It's also part of a process of social atomization where more and more of our interactions are mediated and modulated by impersonal commercial processes.
Again, I'm speechless at the level of indoctrination. It is for my own good!
I'll be working on my next business idea. A shower head with an invisible camera that streams to the manufacturer for the protection of the shower user. After all, he might slip and the device could call an ambulance.
The shower head could monitor if the user is overweight and send some weight-loss ads. It is for the user's own good!
Also, I'll make it opt-out, and the opt-out process will be in a shrink wrap license.
It is all for our own good!
I do not understand this point.
IoT devices are often evil, and I would no longer be surprised if anyone took the above business "idea" as a manual and went ahead with it.
It would have been unthinkable 20 years ago, and those were better times in that respect (and many others).
There are a lot of existing laws about aircraft noise that cars, trucks and scooters don't need to worry about. And once you reach a certain number of takeoffs and landings your "airport" falls under additional regulations that require things like permanent noise monitoring and paying money for noise abatement for nearby property owners.
In my state, any citizen can file a public nuisance suit for unreasonable noise that interferes with quality of life. With trucks, cars and scooters it is perhaps a little difficult to track and attribute noise sources. One company making thousands of drone deliveries makes for a very easy target.
* California limits exhaust noise to 95 dBa: http://www.californiacarlaws.com/exhaust-noise/
* Colorado has a subjective standard "prevent any excessive or unusual noise" https://law.justia.com/codes/colorado/2016/title-42/regulati...
* Connecticut has a pretty complicated scheme depending on the vehicle's weight and age, as well as the road surface: https://www.cga.ct.gov/2009/rpt/2009-R-0205.htm
* Delaware has a subjective standard: "muffler in good working order and in accordance with manufacturer's specifications and in constant operation to prevent excessive or unusual noise." https://law.justia.com/codes/delaware/2015/title-21/chapter-...
There are, of course, laws about other sources of noise too. Ask any college kid.
The level of enforcement for this has to be something like zero....
As for cars and trucks, the noise they produce is regulated in about half of the US: https://www.semasan.com/resources/exhaust-noise-laws-state
It's too early to make the prediction with certainty, but I won't be at all surprised if 20 years from now "autonomous delivery drone" is right up there with flying cars and nuclear fusion on the "20 more years" shelf.
If the standard is that new inventions must be invisible and inaudible, I don't see a lot of innovation in the future.
But I don't think my housing situation is that relevant for the issue.
I'm not smart enough to know what makes me "reductionist".
[1] https://en.wikipedia.org/wiki/Health_effects_from_noise
(scnr)
Some techs suck because they just need refinement/people aren't used to them, and some are just bad ideas. It's hard to tell in advance so people bet on all sorts of things.
Every technology goes through a period of societal decision whether to tolerate the downsides (even very broadly distributed, vague downsides) compared to the benefits created.
Super capacitor based drones would probably be the best outcome (if we could figure out how to harness the power)
There are still several issues with this scheme. The drone needs to be autonomous enough to know when to take off, which partially involves knowing the trucks current and future route information. It also needs to be able to navigate safely in low level flight to the destination, and be able to safely drop the package off. The original prototype delivery required homeowners to build a small helipad in their back yard to get packages, which is impractical. The drone then needs to be able to fly back to the truck, knowing exactly where the truck is and where it will be in a few minutes and possibly land on a moving vehicle with enough precision to hook up the charging system and somehow select and attach the next package for delivery.
When you look into the details this is a horrendously complex system. Some of the tech isn't really there yet and the rest requires a whole elephant load of integration and testing. There are loads of things that can go wrong. The truck might get diverted around roadwork, or be in an accident, or there could be kids playing in the yard when the drone arrives with the package and nowhere safe to land, territorial birds may attack the drone, or the battery pack on the drone might fail and catch fire, a no-fly NOTAM may appear suddenly over your delivery area, or so on and so forth.
Joking aside when Amazon's drone program first made news I thought they were going to make a deal with USPS to use the gazillion post office offices as drone bases. Guess it never went anywhere.
The reason USPS is having financial troubles is because conservative extremists in the American government have put huge financial burdens on them like fully funding their medical plan 10 years in advance in a blatant attempt at bankrupting one of the few Federal agencies specifically enumerated in the Constitution so they can then sell this services to FedEx and UPS and make huge gains for shareholders.
And it was passed with broad bipartisan support. I wish people would stop repeating the canard about "conservative extremists" with respect to this law.
[0] https://www.washingtonpost.com/business/2020/04/15/postal-se...
Davis being the congressman who introduced the bill.
I personally think this could work particularly well, maybe not with blimps but with trucks doing optimized routes and deploying drones for the last mile.
That amount of packages is not extraordinary, and in my car was more full today. I’m on lunch now, 166 stops down, 42 stops to go.
Besides, seeing as our roads are exclusively maintained from fuel taxes, I cannot see this going on forever - the government will have to find a way to tax the electricity going into electric cars somehow.
I don't see why something similar couldn't be done for taxes - the car keeps track of kWh used and you just pay tax based on how much you used. Obviously, that runs into certain immediate problems - like, what if you drive abroad? Normally you'd simply buy fuel while abroad, so therefore contribute to taxes there - but with this system this wouldn't work anymore.
Nothing is utopia. In fact any utopia is probably a dystopia.
That is not to say it doesn't have benefits but to suggest the entire population can haul their arse on a bike doing a 20mi round trip every day for 40 years and be hopping along nicely is a dream.
Not to mention the incidental injuries of which my colleagues suffer from, and myself going back a few years (hospital wasn't much fun)
It is easy to avoid incidental injuries, and everyone would much rather the trauma of an incidental cycling accident than the trauma of a highspeed car crash. Mentioning them in just scaremongering. Ride at a safe speed on a proper path and look where your going. If you can't look where you're going, you're going too fast; slow down. (You can't say "proper paths don't exist", because we're talking about the dreamland in which they do, therefore, by definition, they do.)
20 mi round trips are unlikely. You get long trips in cities developed for cars.
This is already true when you augment public transport with bike lanes, cheap or public bicycles, and slow narrow vehicles for the movement-impaired which can go on bike lanes.
... that is, in many cities in the world, especially in Europe. The Netherlands is particularly good in this respect.
This was from a few years ago. https://techcrunch.com/2017/02/21/ups-tests-show-delivery-dr...
However, I think practically speaking, I don't want the roof of my car damaged, so there would have to be some capital cost to create some kind of roof protector landing pad thing. So that might limit adoption.
On the other hand, if I drove a pickup truck, I wouldn't care. I would opt in without any changes to my vehicle. I would just need to make sure the bed is empty.
No need to go that high at least initially or even use parachutes. Just mount your system that lobs packages on a truck. It just drives a route and some software lobs everything where/when needed. Use a special sticker that's reflective in a particular spectrum for the truck to target and give the stickers out for free. Have the truck scan for the stickers, record gps coordinates (so you know which addresses are deliverable with your new system) and scan for obstacles so you can compute potential firing angle (too low and you're in the shrubs, too high and you're in the power lines or tree branches) as it drives around. Once you know the potential firing arcs for a supported address you can calculate what packages can be delivered without going too high/low or over/under shooting. Even if the truck has to stop to shoot it's still a labor saving improvement over traditional delivery and the sound of a low pressure pneumatic cannon isn't any louder than a truck door. You could use existing warehouse infrastructure to make sure that packages that are not of the right density (or are market in your system as too fragile)to safely toss don't wind up getting tossed.
You're basically just using software and data input to replace the judgement of the paper boy (hard) and some powered rollers and actuators to replace his arm hard (easy).
Disclaimer: I spend 3min thinking about this, this is by no means a serious pitch.
I can definitely agree with the theory behind it. Pair autonomous systems with delivery trucks and let them deliver easy packages (light weight, non-fragile) and let them deliver from the truck to harder addresses (long driveway, up a flight up steps in a breezeway) or deliver ahead.
To me, it seems like a more reasonable assumption is that the drone operators would pay the transit agencies money. Since the buses are already going on those routes, the marginal cost for the transit agency should be nearly zero, so it should be easy to make a profit.
So this seems like a good way for transit agencies to increase their funding with little downside.
Okay, "nodrones123," if the companies paid for this, do you now approve? Or is this just the first in a litany of complaints that will be rolled out whack-a-mole style?
Would need to make a package lockbox similar to a mailbox that is accessible by the dronecars, but I think in this day and age of frequent package theft, many would go along with it.
at this point, you can almost get rid of the whole last-mile problem that you're trying to solve with the drones, and the delivery truck can just drop off pre-filled lockboxes roadside.
There's a point at which things are "good enough", and small improvements aren't really worth the costs. How much time would using drones for the last couple kilometers save? A few hours? Local delivery trucks tend to make rounds every day, so it really does seem like the maximum amount of time drone delivery could save is one day.
Will the cost of operating this (not only monetary, but also organizational, and in terms of noise pollution, drone repair, the occasional package/drone lost, etc) be worth the whole one day faster delivery to customers?
This, especially considering drones can't deliver everything, so trucks would have to keep operating. Truck's are large— economically what makes sense is to have them full of packages and with a whole-day schedule. Would drone usage offset enough deliveries to cut a few truck routes? Would they be enough to make it all worth?
A lot of questions in the air (heh).
Here's the code: https://github.com/sisl/MultiAgentAllocationTransit.jl
Much more interesting that some sensationalist "Wow, AI drones are taking the bus" piece
What are people even thinking, mixing spaces for humans with our fragile little egg heads and corporate swarms of falling death blocks?
hmm...
yes, humans can attack drones. but humans can also attack postal carriers, or steal packages off porches. neither of those facts seem to be disqualifiers for continuing with the current system. Some level of theft will always happen, and the cost of that theft needs to be factored into the overall cost of the system. if that cost is too high, it can make the system untenable, but eliminating the possibility of theft is not a realistic goal for any delivery system.
That's what I mean. If postal service workers and their trucks get abducted / stolen once every 10-20 deliveries then that would make the postal delivery system untenable as well. :D
It's a little bit apples to oranges here. While in the classic systems you can only get the good stolen, here we're talking stealing the good and the delivery system with it.
Never mind the noise; we'll design the cars such they output the inverse frequencies being output by the drones, canceling out their sound.
Even better, the drones can be designed to look like pies.