1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
|
Amazon and Google are planning to use drones for deliveries in the surprisingly near future. If that still sounds like science fiction to you, consider this -- aside from regulatory permission -- there's nothing stopping anyone else from doing the same.
Which is to say that the startlingly fast technological progression of consumer drones has already very nearly made it possible for your local coffee shop to send a drone to drop a latte off on your third floor windowsill.
As we [predicted](http://www.wired.com/2015/01/autopilot-tech-total-game-changer-drones/), the last year has seen an explosion of automation tools and software upgrades for all kinds of drones, most centered around automating various aspects of drone flight. We've seen waypoint and point-of-interest based navigation from DJI, which allow [Phantom]() owners to program flight paths with the flick of a finger across a map. Figure out how to attach a latte payload and you're not far off a drone deliver coffee shop.
Other drone makers like Yuneec have been busy creating both software and hardware aimed at making it easy to have a drone tag along behind you, which means we're probably not far off Justin.tv, the drone edition. I had the opportunity to play around with Yuneec's Typhoon Wizard -- basically a remote control for "follow me" flights that ditches the more complicated video interface for a control that's considerably slimmer and less conspicuous -- and found that it isn't that hard to wander around with a drone 350 feet above you recording your every move.
At this point there are two obstacles to both drone delivery services and ubiquitous clouds of drones in the skies above us -- battery life and the risk of collisions.
As good as some of the waypoints tools and following software is, from what I've tested none of it is capable of replacing the pilot just yet. That drone headed for your windowsill is far more likely to smash your window than gracefully set down a cup of hot coffee. Still, just a few years ago it was impossible to get your hands on a drone, let alone program one to fly where you want it.
While the dream of a delivery drone for every business will likely be crushed by government regulations, if the current pace of innovation continues, the technology will be there very soon.
That's great news for professionals working with drones. Frank Kivo, a videographer with [Concierge Auctions](http://www.conciergeauctions.com/), which frequently uses drones for real estate photography and video, tells Wired that "these automated and new system implementations are extremely helpful and useful, depending on the industry you're working in". Kivo's work sometimes involves filming large parcels of land, which can be automated by programing in waypoints ahead of time. Waypoints can also be useful for even larger scale land surveys such as scouting oil rig locations, data transfer (like retrieving images from a remote wildlife cam) and even search and rescue efforts.
Others are less sold on the automation. Mike Lord of [Barrelman Productions](http://www.barrelmanproductions.com/) tells Wired that, for photography at least, "nothing is ever going to be able to tell you what things looks like when you're up there, so you don't know what the interesting shot is going to be until you're actually flying."
That's something Kivo echoes, saying, "The one downside of automation is you cannot teach automation emotion... there is no way to teach a waypoint system to capture the beautiful sun flares that come into the camera for that split second."
"You need to know," he adds, "when to break out of the automated flight system."
Lord praised some of DJI's new automated systems for striking the right balance between automating and leaving the pilot in control, like the new point-of-interest feature and the course lock feature, which holds the current flight path. With the point-of-interest feature Lord says it's easy to "fly up and away while circling and shooting photos, which previously would have required a second person to do well."
Both Kivo and Lord were most enthusiastic about the various avoidance detection systems being developed. "I can't stress enough how much avoidance detection is a huge plus," says Kivo who had the opportunity to test the collision avoidance system in Monster-X's heavyweight drone in the Bahamas. "With a tropical storm forming, we were constantly getting blown around and with a common consumer drone, we would have lost it, I'm sure."
While lightweight Phantoms may never hold up in tropical storm winds, a collision avoidance system would at least keep them away from the palm trees. As it stands, most automation still requires the oversight of a skilled pilot.
For Lord, that's the way it should be. "the fun is flying ... and you're always going to be more adaptive and smarter than a series of waypoints."
Indeed for many, perhaps even most, the main appeal of drones is the actual flying. It does take practice to get good, and collision detection systems could help take some of the potential bank-breaking risk out of learning, but ultimately flying a drone is fun and trying to automate it all away is solving a problem that doesn't exist. As Lord told me, "I want things that give me more control and make the drone more responsive to my input, I don't want to just send a robot up in the sky."
|
