Bitcoin Forum

Hmm..I am thinking of starting a robotic courier network...for real!

I think the first step is to start small, yes?

Delivery of USB drives is ideal! The first route between two bitcoiner that is the shortest distance will have to be determined.

That mean:

1. We can start with a small robot prototype..and can be used to conduct the first courier transaction!

2. Start in an area where bitcoiners are unusually concentrated such as New York and Washington, D.C. area!

3. Start with delivery of a cheap but still valuable good.

Use a blimp.

 a second hand predator drone.

or
http://www.suasnews.com/2011/03/4720/hexacopter-that-detects-motion-and-breathing/ (http://www.suasnews.com/2011/03/4720/hexacopter-that-detects-motion-and-breathing/)

A small autonomous robot capable of carrying a useful package (say ~1kg) and having sufficient sensors and machine vision capabilities to guide itself from point A to point B and avoid obstacles, vandals, unforeseen occurrences (bird hitting it or a kid running in front of it), and ward off the occasional angry house pet...

This is going to cost a LOT.

Since they have a limited lifetime and require fuel and maintenance would the total costs ultimately be cheaper than a minimum wage dude with a truck?

Yeah, I think dude wth a bycicle/truck/car/running shoes would be much more realistic starting off.

Apart from the reasons above, not many people in NewYork or Washington have landing strips.

Once bitcoin becomes self aware we don't want to give it a head start with drones now do we.

Having a look at the bitcoin map the biggest concentrations of bitcoin users is on the upper east, Boston, Mass., NY etc. And in L.A. and San Fran.

There are also quite a few around Florida, a lot in Dallas, Collarado for some reason, Utah, and near the coast in Washington state.

I am sure an informal local delivery service can be set up by those users, that then becomes a node, and every once in a while nodes can send large groups of packages to each other.

Who say the robot have to be...autonomous? It could start with a simple prototype.

Wait.... so we have to pay for a skilled-above-minimum-wage dude to drive it, AND pay for the robot itself?

A dude with a truck sounds cheaper.

Unless you're actually proposing to build a truck and just giving it a fancy name of "robot".... that *might* be cheaper.... but ultimately thats like FedEx making their own vehicles...

... i'm not saying its impossible, but complete "vertical integration" of a business to *make* the stuff they need to do their business I've only seen work when the stuff actually doesn't exist yet... (Like NASA...and that's not even a profit making business).

We are talking about robots that:

1. Drive short distance between two locations.

2. Pedestrian speed.

3. Poor performance characteristics.

That's what I have in mind. I don't have in mind a Serious Business with godly computer programmers cranking out algorithms.

Rather, it's composed of a network of robotic courier enthusiasts who sometime got paid to deliver a few goods there and there and the network will start extremely small. Eventually, we'll start learning how to create economical designs and the software will be of better quality.

(That being said, high quality sensors are getting cheaper and we're becoming more knowledgeable about machine vision)

I can see these non-flying, walking speed robots becoming victims of children and bored teens that populate their route.

You might have a look at how courier robots are used in a hospital. They have been minorly succesful but have turned out a little expensive and a little unreliable.

And their unreliability has nothing to do with the robot itself, but with people in the hospital just stopping them out of curiosity(or kids).

Seriously do this with people before you start using machines.

Because they sucks so we should never use machines?

Anyway, I am more interested in the robotic than the courier part.

I think then you should focus on probably with a robot can specifically do better than a human.

A courier service is one of those things that humans are naturally really good at. They have built-in sensors, obstacle and collision avoidance algorithms, their own propulsion system (takes up nearly half their body), energy consumption and distribution systems, AND they can deal with many unforeseen circumstances including angry house pets. Best of all, their initial base cost of around $110k-$150k is not paid by you, but by a parent human or even partially subsidized by the government. You just have to pay a rental fee that works out to approximately minimum wage.


Robots on the other hand are better for really high speed, excessively and rigidly repetitive, or extremely dangerous tasks. I think that's what you should go for.

Maybe design an ASIC that mines bitcoins even better than GPUs?

Google cars are changing all that.


BOO-ring. I am uninterested in mining.

I think we will one day see something like this. There are huge advances in robotics underway. This year I'm going to this event to check it out for myself.
http://www.robocup2011.org/en/

Or you could use one of these.
http://www.youtube.com/watch?v=ELr3U4l5HkQ

Wait.... So you're competing with google cars?


Good luck!
 
(snicker)

What I am getting at is that advancement in technologies is making it practical.

I am irate about people shooting up baby step plans for getting a robotic courier network operational. You gottach learn how to crawl before you can run. Trying to make robots do things that it wasn't good at is the whole point.

Cyborgs are theoretically robots.

 :)

Decentralized delivery network?

Brought to you by Silk Road.

Baby steps? what baby steps? you want something that can compete with human beings for an objective that is complex enough even for a human, (kids get lost all the time after all).

that's like saying you want to run a marathon on baby steps. This is why we're shooting you up.

If you want baby steps, instead define baby distances for the objectives.

Instead of delivering stuff in a real city. Make a much smaller baby objective of delivering an item around the corner of a hall way. Much smaller and more realistic. Then increase complexity as needed.

Your research plan would be MUCH more sensible and reasonable if a series of baby objectives were included rather than the final objective.

There is nothing wrong with baby steps, but no one unleashes their babies on the roads, they first attempt those baby steps in the safety of baby's home.

What we're arguing about here, anyway? We agree on the baby steps idea, it's just that I didn't make it sufficiently baby steps.

To put it mildly, that your robotic system compares unfavorably to existing systems.

We're attacking your lack of "sufficiently baby steps".

So what if it is?

It's not going to be automatically superior from day one.

And remote control was not good enough for you.

nope

Whatever get us to a proof of concept. The worst performance characteristic, literally likely to fall apart, etc. I don't care if a human have to follow it.

That is the spirit of my plan.

I think robot couriers are a cool idea. You think it's impractical. Pfft. Who fucking care.

Cool Idea, and impractical are not mutually exclusive.

Do you not care about your project?

I said: whatever it takes to get a proof of concept going. If a human being have to follow it, let it be so. If it have to go around the neighborhood sidewalks ten time, let it be so. If it have to learn how to make it out of the door of my house, let it be so.

Is that so hard to communicate?

In any case, I have no knowledge, experience, or the resources to execute it.

I have nothing to say until I execute a prototype.

@Kiba

So... You want to do whatever it takes.... ok....

But... you have "no knowledge, experience, or the resources to execute it."


I rest my case.

It would be easy (and cheaper) to build a device that rides on rails.  Weight and range wouldn't be an issue.  It could be be autonomous very easily.

http://www.instructables.com/community/do-it-yourself-4-wheel-adult-sized-pedal-bike-car/

http://en.wikipedia.org/wiki/Standard_gauge

train detection (http://www.fra.dot.gov/downloads/research/ord0304.pdf)

I was looking at hexacopters some time ago. Didn't end up getting one, got a Makerbot instead, but anyway, from what I could tell, a working package would come to around 2000 € (whatever that is in USD at the moment), with GPS autopilot and such. Can carry maybe a pound of material, top speed was supposedly around 80 km/h. (!) Don't think it can land on its own at the moment, not sure. But the ability to take off, rise to the desired height and to fly through a series of waypoints is there.

Not a bad idea actually, a monorail, something as simple as a pole that runs along the side of a few buildings would provide a local area delivery system that could be reliable,quick and safe.

rail bike (http://www.youtube.com/watch?v=I2HLrpW0cLg)

$110k-$150k? oh man, where did you get that number from? sounds way too low to me.

Do it at night when no one is looking. But seriously, you're right. So pizza delivery boys/girls it needs to be then.

there is a solution: http://www.youtube.com/watch?v=xw8etotz3TI

Why not? Many pizza delivery people deliver other stuff, too. I'm sure you know the stuff I'm talking about.

Would probably easier to just get bike couriers in big cities to accept bitcoin.

http://en.wikipedia.org/wiki/Cost_of_raising_a_child

You can verify the data in the citations.

You don't even have to pay the human minimum wage necessarily.

also if you're going to build a robot and use it to make money, why not just sell the robot?

This thread made me lolol

You guys didn`t get Kiba's point: A ROBOT THAT DELIVERS THINGS!
That`s really cool!

Of course he doesn't care about the "that delivers things" part as much as he cares about the "a robot"

Are you just pretending not to be geek enough to get how cool that is?
Just make the robot dance around the block and it`ll be way cooler than a pizza-kid delivering stuff. =D

Of course I woudn`t trust the thing to bring me my weed and the robot wouldn`t be able to cross a street without being poked, stoped or stolen... but that`s not the point =D

Perhaps for now we shouldn't worry about the robot providing its own locomotion. Why not have it hitch rides? Perhaps just a box that would only open for the intended recipients and destroy the contents if tampered with.

Otherwise, I like blimp-bots or pipe laying tunnel bots. The latter would allow for pneumatic delivery later on.

How about just a rocket with a robotic paraglider payload that deploys at apogee? The payload would glide to its destination. Seems like a good way to cover a large area.

It can be from time to time. I think the novelty of a robot courier service would overcome the weather imposed delays of airborne ideas. I suppose gliders would be better suited than blimps.

How about just robot carrier pigeons (http://www.festo.com/cms/en_corp/11369.htm)?

On a related note, this robot (http://www.youtube.com/watch?v=tPg1ZMiC9pA) is something many people will find useful.

Quadricopters are probably best for urban environments where you'd need precise maneuverability to avoid collisions, but I was thinking of something a little longer distance, lower precision. I think we already have the technology (http://diydrones.com/) at low enough prices to make drone mail possible in either case.

The drone would land on the receiver's lawn where he would remove his package, load another one if he wanted, and relaunch the drone to claim his deposit less any new sending fees. Folks could also sign up to be nodes to increase the range of the service.

If you nanobots idea works, every garbage landfill will become a gold mine.

Also if peak oil theory turns out to be true, this may be the best option.

If the bots are going to propel themselves through the air in an urban or suburban environment, a quadricopter platform would work best. But for the most utility, I'd recommend something similar to a Harrier or Osprey like setup, but using at least four ducted propellers, thereby allowing the bots as much range and landing sites as possible.

I think this story is relevant: http://arstechnica.com/gaming/news/2010/12/how-a-rc-airplane-buzzed-the-statue-of-liberty-with-no-arrests.ars (http://arstechnica.com/gaming/news/2010/12/how-a-rc-airplane-buzzed-the-statue-of-liberty-with-no-arrests.ars)

me? nah. oxygen, hydrogen, carbon, cheap stuff.

It would behoove the service's operators to have a few pilots ready to take control of bots that find themselves unable to negotiate a given circumstance.

Latest News: Drug carrying drone crashes into Brooklyn Bridge, spreading $1.5millionUSD of Cocaine on the road. Several motorists were seen to stop their vehicles and began snuffing the white powder before being taken by the previaling wind(the poweder) and later the police(the motorists). The developer of the drone believes that it was the use of a float instead of a precision decimal that caused the crash.

Helicopters (including multirotors) are not efficient at all. Same thing with real helis versus airplanes. Helicopters' range is fairly limited.

Although its probably still good enough for, say, manhattan.

Other than that, building good robots/rc stuff generally takes a _lot_ of time and a _lot_ of money. I don't see this idea happening for a couple of years at least.

Personally, I love the idea of flying, autonomous delivery services as much as the next guy, but I think this conversation needs some re-alignment.  The issue more presently at hand is that of a lack of decentralized distribution services.  I had an idea for such a system.

BitDrop.

Bitdrop would be a system for human "runners" to pass goods from person to person in the most cryptographically secure way possible. 

Person A negotiates a sale of an item to person B online.
Upon deciding on BitDrop as means of distribution, distance would be calculated and current gas prices would be tabulated.  This is the base shipping cost. 

Person A types in their "willing to travel" radius, and, the machine would chew on the data and randomly/securely spit back out an exact lat-lon point within that radius.

A nearby runner gets notified that a package is ready for distribution.  He's signed up with his own "willing to travel" radius as well.  Ideally, if there are enough warm bodies in the chain, the venn diagrams all start overlapping, and form a human supply chain.

Upon the completed transaction, all runners in the chain would receive good feedback (ideas for delivery confirmation? sms? scanned qr codes?), and get higher GPG rankings, which would enable them to deliver goods that had been marked with higher trust requirements...and likewise, the more you're trusted in in the chain, the more the runners can charge.  Seller wants to sell something of importance/high value? they can raise the trust threshold (and subsequently, the shipping rate)

As far as black market goods go, I've thought of this scenario.  The TOS for bitdrop would state specifically that transfer of illegal goods is strictly prohibited; however, you may anonymously mark your package "shaded" if you feel the contents pose risk for any member of the chain, including senders, receivers and most importantly, runners.  Shaded goods may be distributed exclusively via dead drops, and runners can opt into notifications of shaded packages, but only after a GPG trust threshold is passed, so a significant time expenditure is required before you go "googling for drugs".  You basically knowingly waive your rights when you deal with shaded transacations. That being said, I'm sure it would turn into a profitable venture for risk-takers, while keeping clean runners away from things that could land them in jail for trafficking.

There are other details I've thought of, including the distinction between dead drops and trusted drops.  Dead drops would be entry-points into the runner market.  Let's say I want to make some extra money, and I want to sign up as a runner.  I go to the bitdrop site and sign up for an entry-level position.  This requires that I pledge a deposit for the first item I'll deliver. This means that new runners will mainly focus on delivering items of little value.  It's a Low-risk, low-return scenario for everybody involved with the transaction.  Once I, the runner, complete my transaction, the deposit is returned to my account, as well as the base shipping cost + small delivery fee.  Part of becoming a runner is setting up GPG-Authentication. The buyer and seller then add ratings to my transaction, and after a threshold is reached, I'm allowed to forego placing a deposit on items of certain values or lower.  Eventually, dead drops would give way of chains with "trusted drops", where people who've established their trust arrange meetups based on encrypted lat-lon values and anonymous instructions. The specifics should be debated about at length to promote the safety and longevity of the chain :)

I don't expect my first ramblings to be the end-all-be-all of this system, but unless I'm way off base, I think the concept has potential.  I actually wrote up an extensive post detailing this idea yesterday, and stupidly hit f5 before I sent it, erasing my whole post instantly...  I'd love to hear feedback and concerns.

Drone operators could participate in BitDrop too, no?

sure!  They could be easily integrated into the system :D

Also, in addition to tracking trust, I think the system should also track reliability separately. A perfectly trustworthy person might be late for reasons beyond his control.

when you say reliability, are you talking about timeliness specifically?  If so, perhaps the runner can designate a delivery schedule (available weekends, tuesdays after 8pm) kinda thing so everybody knows what sort of time-frame all the links in the chain are running on.  If the package needs to be delivered to the next runner or the destination before this runner can deliver it based on their schedule, they're opted out of the "delivery available" message all together.  

Likewise, if they agree to run a package and are late circumstantially, a negative mark should be considered.  If a theft occurs by a runner, the account will essentially self-destruct, and their deposit (if applicable) should be either given to seller, buyer, or donated to a fund to be voted on by the two.   At this point, criminal proceedings may occur.

I guess it's a semantic thing. Labeling people as untrustworthy when that isn't really the case doesn't sit well with me.

The same way, except the runners would be robots managed by an operator.

What we need is a strong leader to make the courier network run on time.

Who else but Megatron?

I think it's a great idea, especially the shaded package concept, and is doable. As a delivery network it won't be fast but I couldn't imagine it being more expensive than say USPO for national delivery.

Wouldn't shaded packages just attract attention though? As a result, contraband senders would just not label their packages as shaded. Why not make all packages shaded?

If you're going to go the airborne route - and I do think that's the most feasible currently for short-distance - then I think you'd be better off not landing at all. mount a smartphone with 3g connection on the drone, facing down. Set flight altitude to 50 metres or whatever is high enough to avoid any buildings in the flight path. When the drone reaches the dropoff point, check video from the phone to see that the area below is clear enough and drop down to 5 metres or so, maybe less. Then drop the package. Obviously you'll need a way to drop it, but that shouldn't be too difficult. Set the drone to automatically return after a few minutes of inactivity, in case something goes wrong - 3g fails, whatever.

Packages could have accelerometers, GPS, and cell phones. When it lands it can send its location via text to the recipient.

Another drone could come and retrieve it. Perhaps using something like this: http://en.wikipedia.org/wiki/Fulton_surface-to-air_recovery_system (http://en.wikipedia.org/wiki/Fulton_surface-to-air_recovery_system)

exactly.  Start small.  We can think about all sorts of new technology in time...who knows, we might vast arrays of solar powered personal helicopter drones at some point in this crazy future....for now though, I say we reel it back in and try to come up with a solution using our primitive cryptography and blood/sweat

o_o *looks around*

...

my mistake!

*backs out slowly*

Same as amateur high altitude balloonists. I have a Motorola 425i that I broke and installed some java that sends GPS (lat,lon,alt) to instamapper. That should get you close enough to see it or switch to other telemetry. Paid $35 for the phone $10 for the prepaid card.

It seems to me that most people here in this thread are trying to model the robotic courier network (RCN? RoCoN? RoCoNe? RoCoNet?) after something that already exists - humans. So why not figure out how humans do it most efficiently and justly first?

Robot see, robot do. (So do it right the first time)

I'm going to start a new thread for those who want to explore the human courier network (HCN? HuCoN? HuCoNe? HuCoNet?).

I've already done this with BitDropor Shady Delivery Network.

http://bitcointalk.org/index.php?topic=6279.msg92782

What about this.

A robotic ocean-going sailboat.

No fuel required.  No traffic/obstacles.  You can start out pretty small.  It would have instant value, as opposed to a more localized courier.

eg: http://thenextweb.com/2008/12/06/roboat-a-robotic-sailboat-powered-by-linux/

No fuel, ok but what powers the electronics and servos?

Wind/water turbines and/or solar.

I'd like to see a craft that mimics a shipping container, mostly submerged, slowly propelled with the juice from some roof mounted solar cells. Security through obscurity?

Unfortunately, cargo ships use vast amounts of energy.  A decent-sized (20ft) vessel wouldn't generate enough electricity to overcome even a modest ocean current.  A collector area of 18ft x 6ft, for instance, could (even at 100% efficiency) produce only about 14 horsepower -- barely enough to troll around a quiet lake.

What about floating container with an anchor? The seller would leave it somewhere in the ocean and then give its coordinates to a buyer. It would would like a meatspace Loom. Although, that idea isn't very robotic. I think the container would have to remain totally submerged, only to release its anchor and rise to the surface when the buyer came along and emitted a signal of some sort. There: kind of robotic.

...but not much of a courier.   :-\

The Colombians are doing some amazing things with submarines these days.

This whole thread reminds me of that book by Daniel Suarez.

Actually the whole Bitdrop project (that is the implementation of this network) stalled a couple of months ago (due to me not being able to give it the attention it needs) and then has come back to life.

I've got several people on board with me about this (developers etc.), a business plan, and soon we'll be IPO'ing on the bitcoin stock market (GLBSE.com) for funding.

http://tacocopter.com/

So awesome.

Yeah I was reading about this, really cool.

Dammit someone already registered pizzacopter.com  :D

what? are you kidding. I am so going to want to try this if I goto sf.

An anonymous network for shipping physical items is theoretically possible, and could be organized like Tor.

1) When you buy something you give the seller an encrypted version of your address that s/he cannot read.
2) The seller ships the package to a reshipper who has the ability to decrypt your address.
3) The reshipper then ships your package through normal channels like UPS.

The main benefit of this system is if the seller is busted for some reason, your name/address are not compromised.

An obvious weakness of this system is if the reshipper is compromised.  A way around this is to have multiple reshippers with layers of encryption and each reshipper peels off only one layer to see which reshipper to send it next (or to you).

You the buyer could even encrypt the entire reshipping route to follow.  The seller would then ship the package to the reshipper you specify when s/he decrypts the address.

The main weakness of this whole deal is the "exit node" i.e. reshippers that can see your real physical address so it can be shipped conventionally to you.

Interesting concept. In theory, the "exit node" could simply be a drop location, although this could still be subject to surveillance.

http://www.google.com/search?rlz=1C1SKPM_enAR434AR442&sourceid=chrome&ie=UTF-8&q=carrier+pigeon+drug+mule

"carrier pigeon drug mule" google search request

That is actually quite brilliant.

http://news.yahoo.com/blogs/lookout/four-tons-marijuana-found-floating-off-southern-california-170826556.html

Every great idea is destined to be copied, and implemented poorly.

You don't anchor at the ocean you have to keep moving. Only cases are sea anchors used during storms or you could anchor for a while if the water was very shallow and calm but someone would still have to be put on watch.

http://www.reddit.com/r/Automate/comments/1fo44j/dominos_tests_pizza_drone/

http://www.theguardian.com/world/2013/oct/15/drone-book-delivery-service-students

If this existed, drugs would be trafficked like this all over the place