I don't know which one of you out there still subscribes to print magazines (besides IEEE Spectrum's excellent and award-winning print magazine, of course), but there's this out there called Time that I guess some people read or something, and this month they've got a couple of very good stories on drones.
There's a lot in Lev Grossman's Time article about the military and privacy aspects of drones, which is totally understandable, but we talk about it a lot around here so I'm not going to rehash it. Instead, let's think about the future. From the article:
Five years ago the Parrot couldn't have existed; it's an anthology of fresh-off-the-vine technologies. Five years ago there weren't cameras as tiny and sharp or chips as tiny and fast. Batteries weren't as light and didn't last as long. Smart phones and tablets still had a long way to go, as did the hyperminiaturized sensors with which the Parrot is studded: an accelerometer, a gyroscope, a magnetometer and a pair of ultrasound altimeters. A few weeks ago, Parrot announced an add-on GPS widget that will be available later this year.
In a way, drones represent the much delayed coming of age of a field that has experienced a prolonged adolescence, namely robotics. For decades robots stumbled along on the ground, slowly and clumsily, rarely achieving even bipedal locomotion. Right now the apex of consumer robotics is that humble domestic trilobite, the Roomba. But it turns out that the earth's surface is simply not the robot's natural domain. When robots take to the air, they're faster and nimbler and more graceful than humans will ever be. All along, robots just wanted to be drones.
Grossman is totally right about the fact that robots have been held back by hardware, and it's no coincidence that smart phones and tablets have made a lot of recent progress just as robots have. It's an unfortunate fact that there isn't enough of a market for consumer robotics to really drive bespoke hardware innovation through mass production, but that's cool, because we can just take stuff like cell phones and tablets (and, for that matter, cars) and rip their guts out and repurpose them. It's not the fastest way to get things done, but it's been surprisingly successful, especially when companies get creative, like using tablets and smart phones for the brains of their robots directly.
I think, however, that robots aren't destined to be drones. They happen to be good at being drones right now because drones are, in a very relative sense, easy. Or at least, way easier than ground robots. Drones in the air can localize with GPS, and they can successfully avoid hitting obstacles with vision systems that are much simpler than what a ground robot would need to be equipped with. Those two facts mean that it's possible for a roboticist to build themselves an "autonomous" aerial platform, and it's just as possible for a consumer who's not a roboticist to buy one for a not outrageous amount of money.
We haven't quite yet hit that point with ground robots, however. We don't have a universal and robust and GPS-like way to localize indoors, and vision is a much bigger problem, because there are many more obstacles that are much harder to avoid. We'll get there, eventually, and when we do, I think it may have a much more profound effect on society than aerial drones do currently, because for the vast majority of people, drones don't (and won't) have much of an impact on their daily lives, while autonomous ground robots have a vast amount of potential. The "humble domestic trilobite" the Roomba is just the tip of the robotic iceberg, and incidentally, while I absolutely love that phrase, it's worth pointing out that there already IS a humble domestic trilobite in the form of the Electrolux Trilobite robot vacuum.
But I digress.
Aerial drones have definitely come into their own over the last half decade, but those technologies are rapidly approaching a state that we might be willing to call "mature." Smart autonomous ground robots are the next wave, thanks to innovations like Kinect/PrimeSense, other cheap 3D sensors, and a massive research and industry push to make it practical. Five years from now, we'll be due for another article in Time about household robots.
Until that happens, check out Lev Grossman's article in its entirety at the link below, and also don't miss a piece that Chris Anderson (previously of Wired and now with 3D Robotics contributed to the same issue here.
Via [ Time ]