A few months ago, Amazon surprised everyone by announcing a drone-based package delivery system which, in theory, would use robotic flying drones to carry packages from one of Amazon’s many distribution centers directly to your door.
The announcement certainly got everyone talking, though once the initial hype died down, cooler heads began to call it essentially a publicity stunt, citing many reasons why such a service won’t be possible, at least, not until some significant tech hurdles are solved. Better battery technology, for example, would be a major factor in enabling bots to carry heavy loads over significant distances. Also, drone navigation and situational awareness. As many skeptics pointed out, it would only take reports of one drone crash accident killing someone’s pet (or worse, a child) and the whole project would be set back years, if not regulated into extinction.
Anyway, fast forward to this morning. As I was skimming over the list of acquisitions made by Google, I realized that Google, not Amazon, already holds the right pieces to put together the ultimate package delivery program. (Knowing Google, they’re surely way ahead of me on this; I’m not claiming I’m the first person to think of it, but merely pointing out what I assume they’re already working on.)
So, what would this ultimate automated delivery system look like? Well, first of all, let’s go back to one of the hurdles of a drone-based system: batteries. Specifically, the need of drones to have a strong and reliable battery that can operate under heavy load for sustained flight. But, wait a minute. That hurdle is partially just us making a pretty big assumption: that the drones would be carrying packages from (and returning to) the Amazon distribution centers themselves. When you think about it, wouldn’t it make more sense to launch instead from a mobile platform? Enter: the ubiquitous delivery truck.
So now we’re part of the way there. Let’s say we’ve got three or four flying drones zipping back and forth, delivering packages from the truck to the door. I think it’s safe to assume that would enable much faster delivery than the traditional method of having a human being making multiple trips back and forth to the truck. (No disrespect to the delivery folks who already do this; I know the ones in our neighborhood are very fast and efficient at what they do.) But with drones, the truck wouldn’t even necessarily need to come to a full stop at all. Smart folks have already got drones that are able to land on a moving platform. So, you’ve now got your delivery truck rolling slowly down the street with drones flying in and out, finishing a street’s worth of deliveries in a fraction of the time. After delivering one or more packages, the members of the drone mini-fleet would “dock” with the truck, where they can recharge their batteries while in transit to the next stop, or easily take turns based on who’s got the most battery life at any given moment.
So, if we stop right here, I’d say you’ve got the most likely scenario for how such a service would realistically appear, at least initially. That’s all well and good, you might say, and Amazon could do that without Google. Yeah, of course they could. But what fun is it to talk about the most realistic way to start the drone service, when we could try to look further into the future and consider ways Google’s technologies could take things to the next level?
For starters: the driver-less vehicle. I assume you’re probably at least semi-familiar with Google’s self-driving cars. If not, you can, well… Google it. Personally, I find them fascinating, and I could probably write a full post on the topic. (Maybe I will.) Also, I know that, at this point in time, driver-less cars are still required to have at least one person in them, but that will change. So yeah, self-driving delivery trucks, when used in concert with package-carrying drones, will eventually eliminate the need for a human delivery worker at all. (While we’re talking about it, taxi drivers and truckers, consider yourselves “on notice” as well.)
Aside from delivery trucks that drive themselves, what else does Google have to offer a drone-based delivery service? Drones, of course! But perhaps… these aren’t the drones you’re looking for… You probably assumed (as I did) that Amazon intends to use the increasingly-popular multi-rotor ‘copters that are currently doing everything from serving as camera platforms to monitoring crop yields. And, judging from all the promo photos, that seems like a safe assumption. But what about other kinds of robots?
For example, robots like those developed by Boston Dynamics. Even if you don’t recognize the name, you probably recognize their handiwork. You’ve no doubt seen the fascinating/creepy videos of BigDog, the quadrupedal robot dog/packmule built to carry heavy loads over uneven terrain for the US military. It’s not hard to imagine a version of BigDog, or his closely-related (and faster) robot cousin Wildcat, shuttling packages back and forth between your doorstep and a moving vehicle. Or, if that is too terrifying a mental image for you, let us consider the scenario I prefer: packages delivered by a humanoid robot based on Boston Dynamics’ “Atlas” platform. After all, the environments a delivery bot needs to traverse all share one common trait: they were designed with humans in mind. Sure, a flying drone could fly over a gate, and BigDog could probably jump it, but why not send a humanoid robot who can just turn the latch? What about gates where visitors have to press a button and get buzzed in? Or complexes where apartments are actually inside a larger structure. Are we going to have to duck as Amazon quad-rotors are buzzing loudly down hallways? Seems doubtful, but a robot could walk right to your door.
Interestingly enough, for those of you keeping score at home, Boston Dynamics was acquired by Google barely more than a month ago.
Clearly, the biggest obstacle for my “ultimate drone delivery system” would be the huge cost of putting the fleet together. Would there ever be a tipping point beyond which the time and wages saved would outweigh the costs of this rolling robo-fleet? Maybe, maybe not. Either way, it is fun to think about.