Was it a drone? British airways flight

[shipdriver]

What matters, from the perspective of assessing risk and determining how much it needs to be mitigated, is the worst case outcome.

I have never seen a risk assessment methodology which only cares about worst-case (that will result in utter paralysis every time and all forms of flight, manned and unmanned, would be unthinkable). Even the most basic forms of risk assessment include both probability and severity of identified risks. Until valid verified (and preferably peer reviewed) data on these risks involving UAVs is available, we have to rely on educated guesses as to the probability and severity of each risk. This applies to aviation as well, where all of the risk metrics involve probability.

And it doesn't change the fact that operating UAVs in aircraft flight paths is dangerous, so it is reasonable for the media to cover that issue too. At the apparent rate of increase of drone use by less cautious members of the public, collisions will likely occur unless prevented by robust no-fly zone implementation.

This seems to be based on the assumption that reports of such incursions are largely valid, which is probably not the case, highlighted of course by the subject of this very thread. This is really a huge part of the problem: the complete unreliability of what data we have (a reading of the actual reports on these incidents makes that abundantly clear) leads to wildly varying opinions on the probabilities involved with nothing more than educated and uneducated guesses on the severity.

 

[Shadow 81]

Just heard the good old BBC reporting that these 'Drones' are now being used to drop drugs onto prison roofs for the prisoners - no idea why prisoners would have access to the prison roof !!! - if they can't get a story one way they will get it another - thanks BBC


Sent from my iPhone using PhantomPilots mobile app

 

[sar104]

I have never seen a risk assessment methodology which only cares about worst-case (that will result in utter paralysis every time and all forms of flight, manned and unmanned, would be unthinkable). Even the most basic forms of risk assessment include both probability and severity of identified risks. Until valid verified (and preferably peer reviewed) data on these risks involving UAVs is available, we have to rely on educated guesses as to the probability and severity of each risk. This applies to aviation as well, where all of the risk metrics involve probability.

This seems to be based on the assumption that reports of such incursions are largely valid, which is probably not the case, highlighted of course by the subject of this very thread. This is really a huge part of the problem: the complete unreliability of what data we have (a reading of the actual reports on these incidents makes that abundantly clear) leads to wildly varying opinions on the probabilities involved with nothing more than educated and uneducated guesses on the severity.

You are correct that all hazard analysis includes consideration of both the probability of the event and the severity of the consequence arising from the event. Low consequence events are generally permitted higher probability of occurrence while still being regarded as acceptable. But then you miss the essential point of the methodology - because for each event under analysis, it is always the worst case consequence that is used to determine the hazard rating (risk).

In this case the event in question is a drone/aircraft midair collision. To determine the hazard rating, and thus whether the risk of this event is acceptable, one considers the probability of such a collision and the worst case outcome (consequence) IF that event does happen.

So - I was discussing severity of consequence, which is always worst case, using a combination of knowledge of terminal ballistics and the educated guesses that you mention. That led to my conclusions that drone/aircraft impacts cannot be ruled as low risk on the basis of low consequence. But that does not lead to paralysis of anything, since all that is required for the risk to be acceptable is to be able to reduce the probability of the event to an acceptably low level.

In the case of a midair collision with a slow-moving flying object, impact with the airframe, windshield or engines are all quite common, and so we take no credit for a significant reduction in probability of any of those scenarios as a subset of midair collisions, and so we have to reduce the probability of midair collision as a whole. Which, in turn, lead to my observation that the increased use of drones, based on the huge increase in ownership and numerous well-documented examples of drones being flown in the vicinity of aircraft flight paths, must, by definition, be increasing the probability of a collision. That trend is inconsistent with the requirement to keep the probability of a collision low to offset the high event consequence.

 

[GadgetGuy]

You are correct that all hazard analysis includes consideration of both the probability of the event and the severity of the consequence arising from the event. Low consequence events are generally permitted higher probability of occurrence while still being regarded as acceptable. But then you miss the essential point of the methodology - because for each event under analysis, it is always the worst case consequence that is used to determine the hazard rating (risk).

In this case the event in question is a drone/aircraft midair collision. To determine the hazard rating, and thus whether the risk of this event is acceptable, one considers the probability of such a collision and the worst case outcome (consequence) IF that event does happen.

So - I was discussing severity of consequence, which is always worst case, using a combination of knowledge of terminal ballistics and the educated guesses that you mention. That led to my conclusions that drone/aircraft impacts cannot be ruled as low risk on the basis of low consequence. But that does not lead to paralysis of anything, since all that is required for the risk to be acceptable is to be able to reduce the probability of the event to an acceptably low level.

In the case of a midair collision with a slow-moving flying object, impact with the airframe, windshield or engines are all quite common, and so we take no credit for a significant reduction in probability of any of those scenarios as a subset of midair collisions, and so we have to reduce the probability of midair collision as a whole. Which, in turn, lead to my observation that the increased use of drones, based on the huge increase in ownership and numerous well-documented examples of drones being flown in the vicinity of aircraft flight paths, must, by definition, be increasing the probability of a collision. That trend is inconsistent with the requirement to keep the probability of a collision low to offset the high event consequence.

There's a very big jump from drones being flown "in the vicinity of aircraft flights," currently deemed to be anywhere within a 5 mile radius of any airport, to YOUR conclusion of an increased probability of a collision, unless aircraft flights are now consistently missing their runways for take off and landings by 5 miles, and those drones are also all engaged in flyaways up to 5 miles away, completely outside of user control.

 

[sar104]

There's a very big jump from drones being flown "in the vicinity of aircraft flights," currently deemed to be anywhere within a 5 mile radius of any airport, to YOUR conclusion of an increased probability of a collision, unless aircraft flights are now consistently missing their runways for take off and landings by 5 miles, and those drones are also all engaged in flyaways up to 5 miles away, completely outside of user control.

That's as bad a non sequitur as I've seen in a long time. I didn't deem anything about flying within a five mile radius. I simply pointed out that the increased prevalence of drones, the obvious disregard shown for flying in the vicinity of airports (evidenced by videos, drone pilot accounts, and observations), and the obvious disregard for flying in actual and designated flight paths (evidenced by videos here and elsewhere, and by pilot reports) is indicating an increasing probability of midair collisions.

Are you actually trying to make an argument that the probability of collisions between drones and aircraft is not increasing?

 

[GadgetGuy]

Are you actually trying to make an argument that the probability of collisions between drones and aircraft is not increasing?

It is grossly exaggerated, especially if you are giving serious credence to the pilot reports, where everything that was previously classified as a UFO, of embarrassingly dubious credibility, is now automatically classified as a drone sighting, of "indisputable" credibility! Seriously, plastic bags in the air are now being reported as drone sightings and near misses by airline pilots!

 

[sar104]

It is grossly exaggerated, especially if you are giving serious credence to the pilot reports, where everything that was previously classified as a UFO, of embarrassingly dubious credibility, is now automatically classified as a drone sighting, of "indisputable" credibility! Seriously, plastic bags in the air are now being reported as drone sightings and near misses by airline pilots!
View attachment 52420

Now you are talking about absolute probabilities, which I have never mentioned. I'm merely arguing that the probability is going up. As for your comments on pilot credibility I'm sure, from reading about them, that some are real and some are not. I have no idea what you are talking about regarding UFOs, and the spoof identification guide may be good juvenile comedy, but does nothing to further the argument or to give the impression that the UAV community is serious about safety.

And if you are arguing that one case of a plastic bag (unconfirmed) being mistaken for a drone means that all drone sightings are plastic bags, or even that all, or most, drone sightings are likely mistaken identity, then you really are going overboard with the logical fallacies.

 

[GadgetGuy]

Now you are talking about absolute probabilities, which I have never mentioned. I'm merely arguing that the probability is going up. As for your comments on pilot credibility I'm sure, from reading about them, that some are real and some are not. I have no idea what you are talking about regarding UFOs, and the spoof identification guide may be good juvenile comedy, but does nothing to further the argument or to give the impression that the UAV community is serious about safety.

And if you are arguing that one case of a plastic bag (unconfirmed) being mistaken for a drone means that all drone sightings are plastic bags, or even that all, or most, drone sightings are likely mistaken identity, then you really are going overboard with the logical fallacies.

Absolute probabilities are the only relevant ones in my book.

 

[sar104]

Absolute probabilities are the only relevant ones in my book.

Well happy reading in that case. In principle I agree, but those absolute values are often too difficult to quantify until a statistically significant number of undesirable outcomes have already occurred. It seems obvious to me that the current rapid proliferation of drones and drone users, combined with an assumed approximately constant fraction of those users willing to engage in risky flying, will inevitably lead to a commensurate (linear) increase in the probability of a collision. I think that it would be foolish not to try to get ahead of that with regulation and engineering controls, where possible. Note that only the major airports are currently protected by firmware NFZs, even with DJI aircraft.

 

[GadgetGuy]

Well happy reading in that case. In principle I agree, but those absolute values are often too difficult to quantify until a statistically significant number of undesirable outcomes have already occurred. It seems obvious to me that the current rapid proliferation of drones and drone users, combined with an assumed approximately constant fraction of those users willing to engage in risky flying, will inevitably lead to a commensurate (linear) increase in the probability of a collision. I think that it would be foolish not to try to get ahead of that with regulation and engineering controls, where possible. Note that only the major airports are currently protected by firmware NFZs, even with DJI aircraft.

No amount of regulation or NFZ's will stop the determined terrorist with evil intent, who is deliberately flying a drone into regulated airspace. That's the real threat, IMHO.

 

[shipdriver]

Note that only the major airports are currently protected by firmware NFZs, even with DJI aircraft.

Are you saying that ALL airports should be NFZs?

It seems obvious to me that the current rapid proliferation of drones and drone users, combined with an assumed approximately constant fraction of those users willing to engage in risky flying, will inevitably lead to a commensurate (linear) increase in the probability of a collision.

One thing that is not constant however is the engineering controls built in to the drones. As the large drone (I'm excluding tiny toy drones like the Hubsan X4) population increases, the proportion that are newer models with altitude limits and such increases rapidly. Also, the people that engage in risky flying probably have a significantly higher loss rate than responsible flyers (ask an insurance agent). The expense of something like a Phantom 3/4 means they probably give up on the "hobby". This turnover of machines has never been quantified (then again nothing regarding drones has really been quantified or studied, just guessed about) but surely that also has an impact on the flying population. This doesn't accord with increasing pilot "reports," a majority (yes, a majority, which greatly surpises me since drones are inherently low altitude machines) of which on the FAA's last compilation are above 2000 ft. All this and without a single validated collision, which as years go by without a collision, our understanding of the probabilities has to change as well.

 

[sar104]

No amount of regulation or NFZ's will stop the determined terrorist with evil intent, who is deliberately flying a drone into regulated airspace. That's the real threat, IMHO.

That's an interesting one. I would have expected that a determined terrorist would find better ways to bring down an aircraft - in fact I suspect that deliberately engineering a collision with an airliner is probably quite difficult, and it's not likely that they would get many chances to get it right. And payload would be an issue for them in terms of carrying anything particularly damaging. But I certainly wouldn't rule out someone trying.

 

[sar104]

Are you saying that ALL airports should be NFZs?


One thing that is not constant however is the engineering controls built in to the drones. As the large drone (I'm excluding tiny toy drones like the Hubsan X4) population increases, the proportion that are newer models with altitude limits and such increases rapidly. Also, the people that engage in risky flying probably have a significantly higher loss rate than responsible flyers (ask an insurance agent). The expense of something like a Phantom 3/4 means they probably give up on the "hobby". This turnover of machines has never been quantified (then again nothing regarding drones has really been quantified or studied, just guessed about) but surely that also has an impact on the flying population. This doesn't accord with increasing pilot "reports," a majority (yes, a majority, which greatly surpises me since drones are inherently low altitude machines) of which on the FAA's last compilation are above 2000 ft. All this and without a single validated collision, which as years go by without a collision, our understanding of the probabilities has to change as well.

I think that all runway approach and departure routes should be NFZs - it is overkill to surround all airports with 5 mile exclusion zones. Finals, base on short final, and climb out are the risky areas. Upwind and downwind legs should be safe if we assume reasonable altitude restrictions.

Interesting comments on the loss rate with risky flyers. Could be significant, unless their risk taking only threatens their aircraft by flying into regulated airspace - such as the guy on here who flew over London and the Thames.

If years do go by with no collisions then that will be great, and will lead to refinement of risk estimates. But right now it is an assumption that no collisions will happen and, when an activity is ramping up this fast, I would argue an unreasonable one.

 

[GadgetGuy]

That's an interesting one. I would have expected that a determined terrorist would find better ways to bring down an aircraft - in fact I suspect that deliberately engineering a collision with an airliner is probably quite difficult, and it's not likely that they would get many chances to get it right. And payload would be an issue for them in terms of carrying anything particularly damaging. But I certainly wouldn't rule out someone trying.

It's more likely the deliberate mischief maker than the terrorist or accidental flyer that will fly over an airport. Same principle as the registration of Small UAV pilots doing nothing to locate the pilot of a rogue drone, as they will simply never register, nor put any identifying info on their aircraft, and will be flying from 5 miles away!

 

[sar104]

It's more likely the deliberate mischief maker than the terrorist or accidental flyer that will fly over an airport. Same principle as the registration of Small UAV pilots doing nothing to locate the pilot of a rogue drone, as they will simply never register, nor put any identifying info on their aircraft, and will be flying from 5 miles away!

That may be true. Although there seems to be a considerable allure to mixing it up there with the real airplanes, either just for the thrill of it or to get cool photos/video. And I suspect that you are right about the registration issue.

Trained eagles it is then.

 

[GadgetGuy]

That may be true. Although there seems to be a considerable allure to mixing it up there with the real airplanes, either just for the thrill of it or to get cool photos/video. And I suspect that you are right about the registration issue.

Trained eagles it is then.

Or just drone jammers, which won't get sucked into the engines like the trained eagles!

 

[Dusty]

Oops: It Turns Out a Drone Didn’t Hit a Plane Over London, After All..... The Daily Signal, May 3, 2016, Oops: It Turns Out a Drone Didn’t Hit a Plane Over London, After All

 

[Tenly]

"The Mercatus analysts estimate that a collision between this class of drone and a commercial aircraft resulting in an injury or fatality to a passenger would likely occur no more than once per 187 million years of drone flight time. "

"the FAA recently opined to the media that it is a federal felony, punishable by up to 20 years’ imprisonment, to shoot down or otherwise interfere with the flight of a drone, even one hovering mere inches above your own back yard."

 

[sar104]

"The Mercatus analysts estimate that a collision between this class of drone and a commercial aircraft resulting in an injury or fatality to a passenger would likely occur no more than once per 187 million years of drone flight time. "

"the FAA recently opined to the media that it is a federal felony, punishable by up to 20 years’ imprisonment, to shoot down or otherwise interfere with the flight of a drone, even one hovering mere inches above your own back yard."

Interesting analysis from Mercatus, but it is entirely predicated on the the assumption that drones and birds represent equivalent threats on impact - in other words it is only mass that matters. That seems to indicate that they are lacking in certain scientific disciplines on their team, supported by the fact that the article in question was authored by an economics student and a Research Fellow in the Technology Policy Program at George Mason.

 

[GadgetGuy]

Interesting analysis from Mercatus, but it is entirely predicated on the the assumption that drones and birds represent equivalent threats on impact - in other words it is only mass that matters. That seems to indicate that they are lacking in certain scientific disciplines on their team, supported by the fact that the article in question was authored by an economics student and a Research Fellow in the Technology Policy Program at George Mason.

Sounds like a drone flyer!
At least he's on our side!