Peter wrote:
Many interesting snippets there. One is that the eye fails to spot aircraft if scanning continuously. One has to move about 10 degrees at a time and stay on each bearing for a few seconds.
I recall that from IR theory…
What really caused an “aha” moment for me was the 1928 (!!!) idea to send everyone at the same altitude in the same direction, completely avoiding head-on risk and reducing collision possibility to overtaking where the conflicting aircraft is straight ahead, the scanning area is very small, and except for a fighter jet overtaking a J-3 the closing rate is much slower with more time to react. The greater the intercept angle, the larger the altitude delta.
Another insight was that the current HCAR was more to satisfy ATC needs than for safety.
chflyer wrote:
What really caused an “aha” moment for me was the 1928 (!!!) idea to send everyone at the same altitude in the same direction, completely avoiding head-on risk and reducing collision possibility to overtaking where the conflicting aircraft is straight ahead, the scanning area is very small, and except for a fighter jet overtaking a J-3 the closing rate is much slower with more time to react. The greater the intercept angle, the larger the altitude delta.Another insight was that the current HCAR was more to satisfy ATC needs than for safety.
Assuming that we can all hold a given altitude, the 1928 solution would have pretty much ensured that collisions such as the Grand Canyon mid air between a DC7 and a Super Connie would have probably never happened – with a 25° angle of closing, they could have missed each other by 75 feet.
For me the a-ha moment of that webinar is the fact that you are more likely to avoid a mid air by pushing down or pulling up on the yoke than trying to divert right, because of the area created by the wings. My strategy, when I had only a PCAS for traffic recognition, was to ensure sufficient separation by increasing the difference in altitude between myself and conflicting traffic. The webinar underscores the need to climb or descend in order to avoid a MAC.
Steve6443 wrote:
For me the a-ha moment of that webinar is the fact that you are more likely to avoid a mid air by pushing down or pulling up on the yoke than trying to divert right, because of the area created by the wings.Not only that, but also the time it takes to roll into a turn and start changing course is much longer than the time it takes to go into a climb or dive. There is a reason why TCAS only gives vertical resolution advisories.
My near-miss was a motor glider on a perpendicular course which passed directly underneath maybe less than 20 feet below. We made eye contact. Wide-eyed eye contact. Judging by the other pilot’s expression (funny afterwards, not at the time), he hadn’t seen me until too late either.
Despite thinking I keep a good lookout, I didn’t see anything, so a good lesson learnt, but does looking outside a little bit more often make much difference? We were both doing 100-ish knots, so this wasn’t even at any great speed. The scary thing is that I didn’t do anything, just stared in disbelief as the other aircraft slid by underneath. Rolling away would have been more dangerous, with one wing down.
I was right at the bottom of Lyon’s class D, and gave the controller quite a mouthful; I do feel guilty about doing this because I assume the motor glider was non-transponder, otherwise he would have told me about it, and hope he didn’t have problems afterwards.
The probability of two objects colliding in 3D space is a function of position (separation) and relative velocity, position error and the size of the objects. The underlying assumption is the objects follow known trajectories. If the trajectories are unknown and changing, a collision is of course a pure random event, modulated by the number of objects per cubic meter of space, their size and relative velocity. In any case, if the relative velocity goes toward zero, the probability of collision also go towards zero.
Peter wrote:
I believe there is a 3-letter place in the known universe where this is the case, or will be soon, but one is not allowed to mention it in Europe
Whew, must not be the USA! US only has an airspace requirement for Class A, B, C and E above 10,000 MSL If you are not equipped, you can still operate in most of the airspace used by piston GA aircraft. IFR and VFR is still supported, even without a transponder, much less ADS-B Out. At the current equipage rate of about 2000 new installs per month, that will get the US fleet to 95,000 equipped by 2020 or just over half of the 180000 fleet that has an electrical system. I doubt this will get above 70% anytime in the near future.
This quite brilliant study from Australia is worth a read.
Have to read it later. But the probability of actually hitting someone is almost zero in 3D space without lookout. The probability of seeing someone is much larger than that.
Any such study without looking at the probabilities is worthless IMO. What could be interesting is the ratio of the probability of seeing someone over the probability of hitting someone without seeing. I bet it is a million to one, at least.
The Australians may not be the right group to understand high density GA operations, given the near total absence of such in their country.
ADS-B IN (not so much OUT given that ground stations provide all traffic with a Mode C transponder) has potential to help spot traffic, but is currently limited by software like Foreflight making it difficult to use the data effectively. Foreflight allows only screening to within 3000 ft vertically and 15 miles, or something like that, which in my regular experience (every flight from base) makes the display so packed with traffic that a pilot cannot easily see a threat and figure out what to do while also flying the plane. This is a similar situation to having one ATC person on the ground attempting to provide separation service for everybody – too much traffic, too many potential conflicts most of which don’t materialize (plus with ATC a low data rate associated with voice communication). ADS-B IN does have the advantage of avoiding horribly inefficient voice communication but for now I think it’s just another tool in the toolbox and that when airspace is densely flown, collisions are rare mainly because of intelligent airspace design, basic VFR procedures and probability. It also has a downside: every bit of attention directed to traffic conflicts that may not in probability materialize distracts from everything else the pilot has to do.
Obviously a lot of the concerns above do not apply when there are only a few planes around, but then again the collision risk probably doesn’t apply either in that circumstance.
I agree with the posts above that creating separation is fastest done vertically but in my plane with relatively quick roll rate I tend to do it by rolling 60+ degrees while pulling power, then rolling upright before recovering. That sends you down very quickly. I suppose if it were very close indeed you might see a meaningful delay due to the initial roll and instead just push over with negative G. Less space eventually but the first part comes quicker. Something to think about.
