Hacker News Reader

Comment by programminggeek | Hacker News Reader

I think it's fair to say that between self driving cars and self guided quadracopters, it is entirely possible to have a fully or nearly autonomous end-to-end logistics system in the next 10-15 years. A lot of the problems with longer distance quadracopters could be overcome with a self driving delivery van, similar to an aircraft carrier with fighter jets.

View on HN

I was a bit distraught to see human hands touching my requested item at one point in the video. Surely that can be eliminated.

I have to go outside to retrieve the package? Deal-breaker.

In all seriousness, there's no reason that homes or apartment complexes couldn't have fairly standardized drone landing pads. Clever people could probably rig up some sort of simple elevator to bring the shipments inside. Incidentally, the landing pads could also be marked in a standard way to assist in computer vision controlled automated landing.

And if you provide a charging point and allow the drone to charge, then delivery is free.

Maybe this is what they need:

http://www.rethinkrobotics.com/products/baxter/

Hmm...I think Baxter will surrepent (http://prevaricate.me/?p=666) into my room though...

Pneumatic tubes, with a drop pad on the roof

Humorously, fetching a small package with a well-defined size and shelf location and putting it into a standardized box on a conveyor belt is by far the easiest step of this process to automate.

Amazon has warehouse robots already. I think it's just often cheaper to pay people.

http://www.bloomberg.com/video/amazon-s-robotic-workforce-no...

Don't forget the robotic mules.

[1] http://en.wikipedia.org/wiki/BigDog

I think you're right. Then the next huge $bn+ company will be the ones who are able to solve the mathematical challenge of optimizing the symphony of orders, autonomous vehicles, drones, and hindering variables such as traffic and weather to maximize system throughput (and thereby profit) in near real-time. If there are multiple competitors and one figures this out better than the others, they will ultimately win.

"How does the package get from the self driving car to your porch?"

Drone?

I don't really mind going out to the curb to pick up a package

Like the aircraft carrier and aircraft, you could have a delivery vehicle that acted as a quadracopter platform to take the item from van to house over shorter distances.

I'm imagining the code to plan the route of the autonomous truck as it accounts for travel time/range in and out for each delivery.

The truck never stops, optimizing its route for all the drones it has in the air.

Writing that code would be a pretty cool job.

Definitely. Time to start writing it now in time for the hardware to catch up!

It's probably been patented already.

Ugh. You say this as a joke, but I vote we come up with a list of features that could be cited as prior art and stick a hash of 'em into the BTC block chain.

I wasn't joking. This is how software patent creation works. Before something can be made in practice, make all the algorithms you might need to implement it.

I believe Intellectual Ventures has patented that business model. You'll be hearing from their drone soon.

Yeah, no matter how obvious something is, someone will probably be able to find a trivial variation of it to patent. I mean, if you have a Taco Copter, clearly a Burrito Copter is nonobvious. I wonder just how trivial they'd allow something like this. I mean it's obvious enough that you can adapt existing telemetry systems and collision avoidance algorithms to this activity, but I doubt the lawyers think it obvious.

If you think it's so obvious to implement the algorithm in discussion here, feel free to publish it openly and preempt future patent claims.

Oh, that's the thing. In a patent, you only have to describe the general idea. It's not like you put source code or anything useful.

But that's the problem: no matter how many variations of the obvious you list, they'll find something to patent.

I imagine much of the time the best optimization would end up being a single drone.

Yeah, with the van housing a system to swap battery packs between deliveries in densely populated areas.

Only in major urban areas where you see a UPS/Fedex truck literally go a block before stopping again would you likely see multiple drones.

But the bigger picture here is to expand what you order from Amazon. The eventual goal would be more along the lines of small singleton items you'd run to a convenience/drug/grocery store for.

Even after you account for a minimum price to make it worthwhile, that's many, many more deliveries than UPS/FedEx does.

Sorry, I was agreeing with you, I've seen pretty much that exact question asked when people talk about autonomous vehicles delivering packages and here we have a huge logistics company playing with the answer.

Amazon Ballistics. Coming in 2017.

Air cannon + http://www.youtube.com/watch?feature=player_detailpage&v=Kyk... ??? (bounce landing) :)

The garbage truck does something similar with a mechanical arm...

catapult.

Reply on news.ycombinator.com