a thoughtful web.
Good ideas and conversation. No ads, no tracking.   Login or Take a Tour!
comment by user-inactivated
user-inactivated  ·  3809 days ago  ·  link  ·    ·  parent  ·  post: When drones fall from the sky

I've been on both sides of this. I once had to go on an emergency patrol because a Shadow crashed into a building in Baghdad and they only had a 4 digit grid as a locator. It was really disheartening because we loved doing night patrols with that thing. One of my favorite missions was following a car that was marked with an infrared beam emitted from the Shadow. It's just damn cool to flip down the NODs and see that green light coming down from the sky. Truly felt like I was living in the future.

I also went to Raven school and clocked some hours flying surveillance in our sector. The Raven is a pain in the ass but it gets the job done.

The Raven is one of smallest drones but even it has substantial "ease of use" technology behind it. If you use the laptop in tandem with it, you can plot GPS coordinates (much like marking something on Google maps) and instruct it to fly in generally two ways: circle a GPS point or circuit all GPS points. These, of course, were changeable in-flight and you could create a new GPS point, click a button, and instruct it to fly there and circle it. Flying freehand was only advised in rare and dynamic circumstances. However, you can still set a minimum altitude when flying freehand which alleviated a lot of the stress of flying it so.

The hand controls and interface were pretty crummy when I went to the school for it. During deployment I got to enjoy a week-long vacation at one of the bigger bases in Baghdad in the form of a Raven upgrade class (from RQ-11A to RQ-11B). It improved upon things substantially -- the interface, controller, laptop software, response, etc. was so so so much better. I imagine similar things, if only software, are rolled out often for the far more complex drones.

I don't want to question WashPo's integrity, but it seems like they are trying to raise the alarm on the physical/malfunctional danger of drones as a developing technology and I'm hoping Part 2 of this isn't about how we need to augment our building codes to account for drones that might hit our houses.

Fortunately, we've had the luxury of finding all of these issues overseas where the errors will only affect Abu Someone's house and life, which nobody really cares about. You can bet the FAA has been taking notes, and they will regulate the hell out of everything and anything when drones really hit the mainstream, and require various "oh shit" features built into them. More problems will be found via quantity but by then the risk might be worth the reward.

Like any other developing technology, all of this stuff is gonna be worked out and it will be fine. We won't need to put steel plates in our roofs, just foil on our windows.





kleinbl00  ·  3809 days ago  ·  link  ·  

    I'm hoping Part 2 of this isn't about how we need to augment our building codes to account for drones that might hit our houses.

I'd wager part 2 is "we need to license anything over 10 lbs and limit larger UAVs to law enforcement and life safety." 'member how in the '70s, the future was gonna be helicopters everywhere? And then one went down in Manhattan and all of a sudden there were a lot of empty helipads on skyscrapers? Licensed UAVs with hefty indemnity are no problem at all. 200 paparazzi sporting HD FPV around the Oscars? Death on a stick.

OftenBen  ·  3809 days ago  ·  link  ·  

Talk about firsthand experience, great insight white.

From your perspective, with the tech you used, does it seem like there would be good commercial applications?

People joke about Taco - Copter

But why can't that be real?(Legal issues aside) What about the theoretical Amazon Instant with same-hour drone delivery in certain areas?

user-inactivated  ·  3809 days ago  ·  link  ·  

    From your perspective, with the tech you used, does it seem like there would be good commercial applications?

Probably, but we won't know until their capabilities are more fleshed out and innovation takes hold. As it stands now going from Point A to Point B is their most obvious commercial application.

    But why can't that be real?(Legal issues aside) What about the theoretical Amazon Instant with same-hour drone delivery in certain areas?

kleinbl00 has a lot of good comments in this thread that I tend to generally agree with. There are many areas of the logistics involved that need to be figured out and drone technology still has a ways to go before they could be addressed. As it stands now I think they will remain heavily regulated in the same way that flying planes and using certain radio frequencies are. Drones are pretty safe to fly between 1,000 and 10,000 feet, but not if you have hundreds/thousands of them going this way and that way.

Even while going through Raven school in the states we had to stay in constant contact with the nearest airport's air traffic control tower and report it's 8 digit grid. We designated a specific area of operations and altitude that we could not exceed. It was potential trouble if we left that designated airspace. The same should be expected of any private entity flying drones that can exceed x altitude. This creates a need for an infrastructure that is not yet created and does pose safety risks by the sheer quantity that could be encountered -- our lessons in drone safety so far mostly relate to the individual flying operations of the drones themselves without much other air traffic in the area.

kleinbl00  ·  3809 days ago  ·  link  ·  
OftenBen  ·  3809 days ago  ·  link  ·  

Ok, if we get over the 'Tech is scary' aspect, and remember we all drive cars constantly, which in my mind are far more dangerous, if you have good capacitors, how much would it add to shipping costs if you were within an operational range of a distribution facility?

kleinbl00  ·  3809 days ago  ·  link  ·  

This is not a "tech is scary" aspect, this is a "quadcopters are not bound by roads" aspect and a "cars can't fall from the sky" aspect.

OftenBen  ·  3809 days ago  ·  link  ·  

    "quadcopters are not bound by roads"

I would argue most cars aren't. Some are more bound than others, but on the whole I think roads (With the exception of cities, because concrete is hard) are more suggestions than rules.

    "cars can't fall from the sky"

No, but planes do, and we trust them with human lives, not iPad mini's. With any technology there's going to be some failure rate. We just have to weigh the possible risks against possible benefits, and I think long term it's going to balance out, to some degree or another, in the favor of commercialized drones.

As long as people drive cars, get on airplanes, and fly spaceships, you're going to have a hard time convincing me that unmanned drones are a bad idea.

kleinbl00  ·  3809 days ago  ·  link  ·  

    I would argue most cars aren't.

This is a very silly statement. Poll 100 people if they're afraid to walk on the sidewalk because cars jump the curbs all the time and 100 of them will tell you that "rogue cars" are not amongst their key worries.

    No, but planes do, and we trust them with human lives, not iPad mini's.

1) Not autonomous ones.

2) Not unlicensed pilots (except for gliders and ultralights, which face their own restrictions).

3) Not for commercial purposes.

4) Not without substantial limitations on the airspace they can inhabit.

5) Not without substantial insurance policies.

6) Not without an equal weight on file for every part on the aircraft.

I can buy an FPV quadcopter off Woot. For my dad to fly his Mooney without someone sitting next to him took 60 hours of flight training. There's no equivalence.

    As long as people drive cars, get on airplanes, and fly spaceships, you're going to have a hard time convincing me that unmanned drones are a bad idea.

You can lead a horse to water...

OftenBen  ·  3809 days ago  ·  link  ·  

    This is a very silly statement. Poll 100 people if they're afraid to walk on the sidewalk because cars jump the curbs all the time and 100 of them will tell you that "rogue cars" are not amongst their key worries.

Regardless of if they worry about it or not, it's still something that can happen.

I'm not saying there won't be, or shouldn't be legislation or licensing. I am saying that we accept all sorts of risks, and legislate around them.

When it comes to equivalence of skill, white was just saying that you can point and click on a map and it will circle at a pre-set altitude, completely on its own. A company doesn't need to have a fleet of elite drivers, they need good automation and a few guys to make sure shit doesn't get wonky. Skill has absolutely nothing to do with it.

    You can lead a horse to water...

In this particular regard kb, I just don't understand your argument. 150 people on an airplane that could malfunction and kill them all is ok, a 10lb drone maybe killing a few people over the course of years is not ok?

kleinbl00  ·  3809 days ago  ·  link  ·  

    n this particular regard kb, I just don't understand your argument.

'K. Walk with me:

1) Cars operate in a two-dimensional environment constrained by roads and traffic laws. Additionally, motor vehicle traffic is heavily constrained such that motion vectors of any adjacent vehicles are identical: like traffic moves at like velocity in identical direction.

2) Aircraft operate in a three-dimensional environment constrained by flight corridors and traffic laws. Additionally, air traffic is heavily constrained such that vehicles with disparate vectors are separated by nautical miles of distance. Finally, airspace over populated areas tends to be heavily restricted.

So yes: accidents can happen with cars... but the odds of any given trip to the grocery store ending in a fireball are slim indeed. General aviation, on the other hand...

    Private-flight crashes were 12 times higher than the average rate for other types of general-aviation flying, Demko said.

    The rate of deadly wrecks in such private flying has grown faster than accidents as a whole, up 25 percent since 2000, Earl Weener, an NTSB board member, said in an interview before the forum. About 1,500 people have died on general-aviation flights since the crash by Pinnacle Airlines Corp. (PNCLQ)’s Colgan, Weener said.

So it comes down to statistics and manageable risk. UPS, for example, delivers 16 million packages a day with 93,000 motor vehicles. For those sixteen million packages, UPS averages less than one accident per million miles driven.

Now - let's assume that a UPS drone is 100% as safe as a UPS truck. Not likely in my opinion (we'll get into that in a minute), but let's presume. The problem being a UPS drone won't be able to deliver as many packages as a UPS truck for the simple fact that UPS drones won't look like this:

So let's ignore for a minute the fact that there will be many packages a drone can't deliver. Let's ignore for a minute the fact that anything reliant on GPS is relying on a minimum accuracy of 6 feet and a median accuracy of 30 to 40 feet. Let's ignore the fact that you can put a UPS truck out in the morning and bring it in in the evening, all packages delivered. Presume you've got ten drones replacing one UPS truck.

Your accident rate just went up by a factor of ten.

The other immediate, glaring problem is that "things that fly" experience different traffic incidents than "things that drive." A UPS truck can have a fender-bender. A hexacopter with an iPad isn't going to "bump into" something - it's going to collide and convert its rather sharp rotating mass into kinetic impact. And granted - it won't collide with the same kinetic energy as a UH-60 slamming into a housing development... but it will collide with enough kinetic energy to get the editorials churning.

Then there's the "delivery" aspect of it. Consider: getting that iPad onto someone's porch is the technological equivalent of getting an AGM-65 into a Toyota. And in order for the CIA to get an AGM-65 into a Toyota, we had to generate LIDAR for all of the Hindu Kush, triangulate via special forces and coordinate with airborne and satellite-borne reconnaissance. GPS for military is unlocked; they have half-inch accuracy. Meanwhile dGPS for civilian use will get you within six feet if everything's perfect; most municipality GIS databases are off by 50 feet or more.

Amazon also cares a bit more about budget than the CIA does. And in the end, it's still a person driving the thing.

From a pragmatic standpoint, it makes more sense for Amazon (or Taco Bell, or anybody) to deliver your shit via bike messenger than it does by drone. Bikes have brakes. They can loiter without consuming resources. They don't need a power-to-weight ratio greater than one. And if you're tying up an employee to deliver your package anyway, might as well tie him up with the package, rather than driving a flying moon rover across Manhattan to bring you an iPad while dealing with everyone else's UAVs in a mad attempt to not clip a pigeon and spiral down on a bus stop with 30 lbs of lithium-ion powered carbon fiber.

Here's my argument: I know quad copters and even the ones capable of 10g payloads hurt like a mutherfucker when they hit you in the face. And they hit you in the face a lot. And while the technology may well advance in a direction to keep them from hitting you in the face to the point where it makes sense to throw an iPad on a Hexacopter to get it to your apartment ten minutes faster, the technology is likely to advance even further in far more pragmatic, far less dangerous, far less catastrophe-prone directions.

kleinbl00  ·  3809 days ago  ·  link  ·  

    n this particular regard kb, I just don't understand your argument.

'K. Walk with me:

1) Cars operate in a two-dimensional environment constrained by roads and traffic laws. Additionally, motor vehicle traffic is heavily constrained such that motion vectors of any adjacent vehicles are identical: like traffic moves at like velocity in identical direction.

2) Aircraft operate in a three-dimensional environment constrained by flight corridors and traffic laws. Additionally, air traffic is heavily constrained such that vehicles with disparate vectors are separated by nautical miles of distance. Finally, airspace over populated areas tends to be heavily restricted.

So yes: accidents can happen with cars... but the odds of any given trip to the grocery store ending in a fireball are slim indeed. General aviation, on the other hand...

    Private-flight crashes were 12 times higher than the average rate for other types of general-aviation flying, Demko said.

    The rate of deadly wrecks in such private flying has grown faster than accidents as a whole, up 25 percent since 2000, Earl Weener, an NTSB board member, said in an interview before the forum. About 1,500 people have died on general-aviation flights since the crash by Pinnacle Airlines Corp. (PNCLQ)’s Colgan, Weener said.

So it comes down to statistics and manageable risk. UPS, for example, delivers 16 million packages a day with 93,000 motor vehicles. For those sixteen million packages, UPS averages less than one accident per million miles driven.

Now - let's assume that a UPS drone is 100% as safe as a UPS truck. Not likely in my opinion (we'll get into that in a minute), but let's presume. The problem being a UPS drone won't be able to deliver as many packages as a UPS truck for the simple fact that UPS drones won't look like this:

So let's ignore for a minute the fact that there will be many packages a drone can't deliver. Let's ignore for a minute the fact that anything reliant on GPS is relying on a minimum accuracy of 6 feet and a median accuracy of 30 to 40 feet. Let's ignore the fact that you can put a UPS truck out in the morning and bring it in in the evening, all packages delivered. Presume you've got ten drones replacing one UPS truck.

Your accident rate just went up by a factor of ten.

The other immediate, glaring problem is that "things that fly" experience different traffic incidents than "things that drive." A UPS truck can have a fender-bender. A hexacopter with an iPad isn't going to "bump into" something - it's going to collide and convert its rather sharp rotating mass into kinetic impact. And granted - it won't collide with the same kinetic energy as a UH-60 slamming into a housing development... but it will collide with enough kinetic energy to get the editorials churning.

Then there's the "delivery" aspect of it. Consider: getting that iPad onto someone's porch is the technological equivalent of getting an AGM-65 into a Toyota. And in order for the CIA to get an AGM-65 into a Toyota, we had to generate LIDAR for all of the Hindu Kush, triangulate via special forces and coordinate with airborne and satellite-borne reconnaissance. GPS for military is unlocked; they have half-inch accuracy. Meanwhile dGPS for civilian use will get you within six feet if everything's perfect; most municipality GIS databases are off by 50 feet or more.

Amazon also cares a bit more about budget than the CIA does. And in the end, it's still a person driving the thing.

From a pragmatic standpoint, it makes more sense for Amazon (or Taco Bell, or anybody) to deliver your shit via bike messenger than it does by drone. Bikes have brakes. They can loiter without consuming resources. They don't need a power-to-weight ratio greater than one. And if you're tying up an employee to deliver your package anyway, might as well tie him up with the package, rather than driving a flying moon rover across Manhattan to bring you an iPad while dealing with everyone else's UAVs in a mad attempt to not clip a pigeon and spiral down on a bus stop with 30 lbs of lithium-ion powered carbon fiber.

Here's my argument: I know quad copters and even the ones capable of 10g payloads hurt like a mutherfucker when they hit you in the face. And they hit you in the face a lot. And while the technology may well advance in a direction to keep them from hitting you in the face to the point where it makes sense to throw an iPad on a Hexacopter to get it to your apartment ten minutes faster, the technology is likely to advance even further in far more pragmatic, far less dangerous, far less catastrophe-prone directions.

OftenBen  ·  3809 days ago  ·  link  ·  

    "quadcopters are not bound by roads"

I would argue most cars aren't. Some are more bound than others, but on the whole I think roads (With the exception of cities, because concrete is hard) are more suggestions than rules.

    "cars can't fall from the sky"

No, but planes do, and we trust them with human lives, not iPad mini's. With any technology there's going to be some failure rate. We just have to weigh the possible risks against possible benefits, and I think long term it's going to balance out, to some degree or another, in the favor of commercialized drones.

As long as people drive cars, get on airplanes, and fly spaceships, you're going to have a hard time convincing me that unmanned drones are a bad idea.