3 ms·

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

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.

munk-a 5y ago
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 https://youtu.be/2efhrCxI4J0?t=111
- meheleventyone 5y ago
  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.
  - munk-a 5y ago
    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).

munk-a 5y ago

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

meheleventyone 5y ago
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.
- munk-a 5y ago
  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).