FAA Unveils Drone Rules: Autonomy Is In, Drone Delivery Is Out
spectrum.ieee.org
spectrum.ieee.org
Overall, they are less strict on what you can do with a drone than I expected. The only really tough one is that you must be within line of site of the drone. I've had some ideas that would not fit in that box. Also, you can not fly over people not involved in the flight. I know that will be a point of consternation for a lot of people, but I'm happy about that one.
The operator licensing requirements seem a bit onerous. What the hell does the TSA have to do with it? "Licensed drone operator" is effectively now a thing. This is definitely written from the perspective that they expect drones to be remote controlled vehicles on steroids, not autonomous computer controlled machines. How do I get my algorithm FAA certified for flight?
This one is critical. Most drone operators using anything above a hobby grade UAV are using fairly sophisticated software to command GPS velocities. This is assuming even more complex GPS missions are not being carried out. How do we setup a certification regiment for that software and its extensions to sense and avoid?
Granted, I don't have the answers. The FAA will inevitably suggest the same certification regiment that large airplane autopilots go through, but that seems prohibitively cumbersome.
What's actually required by an autonomous system is closer to a large percentage of FAR 61 and FAR 91, which are pilot certification and general operating rules, respectively. Your very basic recreational or private pilot conforms to both of those sets of regulations on every visual flight. For instrument flights requiring positive separation from air traffic control because the pilot can't see and avoid themselves? That's another layer of regulations and conformance. For charter flights, that's FAR 135 a whole separate set of regs. For commercial passenger flights, that's FAR 121 yet another separate set of regs.
Track down a FAR/AIM manual and start reading FAR 61, 91, 135 and the AIM part. A bulk of this needs to be incorporated into an autonomous UAS controller. The details of course have to change, possibly by a lot in some cases, because we don't have things like primary or secondary radar or radio communications at the altitudes UAS's will operate. But there are analogues and they need to be invented, specs defined, infra created, and so on. And while some of this could be implemented in proprietary form, the portions that require communication or cooperation in shared airspace have to be publicly documented.
What do emergency procedures even look like for an autonomous aircraft? A bunch of failure vectors and likelihoods need to be defined, e.g. since a partial power failure can occur anytime. Where should it set down if it has that much control? If it lacks sufficient control to get to a safe setdown zone, how does it crash land? i.e. better to crash in the middle of the road than on top of a car or person. Progressive failure.
The certification and currency process for your algorithm should be more like a pilot because the autonomous UAS software is in effect a substitute pilot. Pilots are certified for life in the U.S. but they have 6, 12, and 24 month (depending on privileges) medical evaluations, and 30 day, 90 day, and 2 year currency requirements for various things. It's not legal for a pilot to fly (at certain times, environments, or at all) if they lack currency requirements, even though they're certified. Similar for UAS. You'd have to have a certain software update that ensures conformance with current version of the public use UAS operating environment behaviors. That way everyone is always on the same page.
Autonomous flight in a shared airspace is non-trivial.
TSA vetting is not even required for a pilot certificate that is a US citizen.
Regulation is not a bad thing. Sometimes long-term planning is required.
Why do you say this? When have I ever indicated this? If it is possible for a hotel to block RF that I want to transmit or receive, then obviously I have checked into the wrong hotel.
I don't think it's quite that simple. I like simple, pragmatic solutions. Generally, yes, I think that the internet(s) will replace government, and I think that this situation is desirable.
I'm concerned that the government will arm the skies and that the people on the ground will become targets. I think that's a reasonable concern.
It makes more sense to me to ban weapons in the skies altogether. Whatever happens, I think it's important that there's one set of rules for everyone, whether they happen to be acting on behalf of the 'state' or not.
Perhaps in the case of airplanes flying or building a subway system there is just some sort of easement or eminent domain ruling applied?
If anyone has some resources on this topic, I'd be interested in reading them.
But from a safety standpoint I don't see why they should be getting an exception. If the FAA judges that it's not possible for civilians to do this safely then it's certainly not easier for law enforcement. If law enforcement can do it safely, then there's no reason why civilians can't get the same certification.
It would seem that if speeding is dangerous for civilians it would also be dangerous for law enforcement. Perhaps law enforcement are able to take additional driving safety courses which enable them to have improved skill at higher speeds. If this is the case, civilians ought to also have the ability to take such courses to improve their skill as well.
Instead, I see the limitations on speeding meant to bring down the total number of fast moving vehicles. And in the same way, the total number of drones flying over crowds will be reduced by the FAA's rules, even if law enforcement are still able to fly them over crowds.
A better analogy would be that law enforcement is allowed to speed in order to reach the scene of a crime faster. One could argue that having a drone hover over a potential crime scene has similar benefits. I'm not sure about the costs and benefits in this case, but it's certainly a possible argument.
I imagine this would be extend to drone use for law enforcement - they will have certain exceptions to do things the general public cannot in special circumstances and will also likely have special training available to them which is also not available to the general public.
http://www.csmonitor.com/World/Global-News/2015/0204/Drones-...
You need to know a little about reading aeronautical charts but for an airport like Boston the inner thick blue circle labelled 70/SFC would probably be the most relevant usage restriction.
Mentioning safety features... let's think up a plethora of patents.
Obviously not written by a programmer. This causes a deadlock in the case that two drones encounter each other. :)
What this means, as far as I understand, is that in the event of an accident between a manned and unmanned aircraft the unmanned will be at fault. However, between two unmanned aircraft both will be at fault.