PDA

View Full Version : Synthetic vision failure


john_tullamarine
7th Feb 2017, 07:22
A significant failure last year which should be of interest to most tech log readers.

There is a discussion starting on the flight test forum (http://www.pprune.org/flight-testing/590595-synthetic-vision-certification-judgement.html) and the relevant ATSB investigation report is here (http://www.atsb.gov.au/media/5772247/ao-2016-064-final.pdf).

Suggest that any local discussion be conducted here so we don't clutter the FT forum with non-FT commentary.

I'll sticky this one for a while (or until it appears that it will be a local discussion topic) so that it doesn't disappear before most have time to become aware of it.

GlobalNav
7th Feb 2017, 19:30
When SVS is displayed on the PFD, even if SVS is not the "primary" terrain and attitude indication, it is compelling and difficult for a pilot to ignore. The "pilot's perspective" (AKA "ego-centric") view in a color pictorial speaks loudly compared to other displays and information. So the effect of its display of misleading information can be dramatic and strong. I have seen proof of concept demonstrations where the SVS was known to the pilots to have database errors and they still followed it. Human nature or as we call it "human factors". The advantage of SVS is how it can enhance the pilot's "situation awareness" also carries with it a vulnerability when it's wrong - it presents the wrong situation which is hard to dismiss.

galaxy flyer
7th Feb 2017, 21:36
My experience in G6000 Vision tells me that it is compelling and hard to ignore, but the R-C system was, in understanding, pretty fail safe. Non-current data base, no SVS, not enough satellites, no SVS, not selected for display each flight, no SVS; prominent SVS FAIL EICAS message. That said, first trip with it, rain and low cloud going into Rio de Janiero. Very happy with the view, nearly VMC in IMC. Then, the system lost some satellites (6 is min, IIRC) and back to nothing. For a brief moment, it was like needle, ball and airspeed.

We had a similar argument with the regulator when displaying "own ship" in the charts as part of RC Proline 21. Chart with "own ship" is on MFD 2, between the pilots and FAA says, you can't use "own ship" as navigation. Reply was, "we're not, just another SA tool."

PEI_3721
9th Feb 2017, 14:10
This is a significant safety event.
No matter how much we think that we will be able to detect failure situations or ignore poor quality data, we may be unable to do so.
The quest for better awareness has just moved the potential for error elsewhere, and of great concern is that because 'it looks good' we tend to trust it. 'More' does not necessarily mean better awareness.

"Both pilots commented that they had previously experienced failure of primary flight instruments at low level and at night in different aircraft (without synthetic vision systems). They had been able to disregard the erroneous or failed instruments and reference the standby instruments to maintain control of the aircraft and situational awareness. However, the prominence of the synthetic vision display is such that it is difficult to ignore erroneous information and locate valid information. Additionally, the pilot flying reported feeling a level motion sickness, probably associated with the combined effects of the prominent synthetic vision display and conflicting vestibular sensory information.

The pilot flying reported that they realised something was wrong but could not initially figure out what it was.

The pilots commented that it was impossible to discern the valid attitude information on the PFD (overlaid on top of the synthetic vision) and revert to flying ‘power and attitude’ given the prominence of the erroneous synthetic vision information."

galaxy flyer
9th Feb 2017, 15:41
While not disagreeing with the premise of unintended consequences, I never found the SV presentation distracting. What was distracting in the Vision was the FD cue was "flight path" centered, not attitude oriented. Soon, one quickly lost a picture of pitch attitude unless one worked at watching the pitch bars.

GlobalNav
9th Feb 2017, 20:47
An interesting comment, which I accept at face value. I know that when industry and certification standards for SVS were developed, some argued hard for the benefits of a track centered ("track up") picture, while others argued that heading centered or nose-centered should be required on a PFD.

GF raises attitude awareness as a problem with the track-centered display, though many expected that where the drift angle was relatively small, this would not degrade pitch/roll awareness. On the other hand, track-centered typically put the primary objects of interest in the center of the display, thereby minimizing the effects of the limited field of view.

What was actually more anticipated as an issue was the loss or "drift" awareness, since the pilot found the foresight (AKA waterline) symbol rather inconspicuous.

There was an implicit notion, held by many, that since "situation awareness" is not a required parameter, it is less critical than an element of the Basic-T, and therefore has less hazardous effects when integrity is degraded. In my opinion, the purpose of the Basic-T was to give the trained pilot the necessary elements of the "situation". SVS hands this to the pilot all nicely wrapped up in a bow on a silver platter and in a strikingly compelling fashion. The idea that the pilot would not be misled by an erroneous SVS picture is outrageously optimistic.

galaxy flyer
9th Feb 2017, 22:04
I agree on the compelling nature of the SV picture, but system integrity and accuracy needs to be addressed in certification, so the SV, in service, can be relied on or it is removed and pilot alerted of its degradation. Yes, an erroneous SV picture would be very difficult to discern in any IMC case.

Your comment on putting the "primary objects of interest in the center of display" does make the flight path-centric cueing more sensible. To put an extreme example, with track-centric cue, one could fly an ILS inverted and the FD cue would be perfect. Or, more reasonably, the FD cue and FPV centered with some rather large pitch/roll excursions that wouldn't be as noticeable as when flying the attitude reference. For example, you are slightly off course, with flight path cueing, you roll to establish a new heading but the roll angle isn't seen by having only the cue in limited view and over bank. On raw data or attitude orientation, one might roll 5 degrees to make a 5 degree course change.

vickers vanguard
11th Feb 2017, 00:23
My experience in G6000 Vision tells me that it is compelling and hard to ignore, but the R-C system was, in understanding, pretty fail safe. Non-current data base, no SVS, not enough satellites, no SVS, not selected for display each flight, no SVS; prominent SVS FAIL EICAS message. That said, first trip with it, rain and low cloud going into Rio de Janiero. Very happy with the view, nearly VMC in IMC. Then, the system lost some satellites (6 is min, IIRC) and back to nothing. For a brief moment, it was like needle, ball and airspeed.

We had a similar argument with the regulator when displaying "own ship" in the charts as part of RC Proline 21. Chart with "own ship" is on MFD 2, between the pilots and FAA says, you can't use "own ship" as navigation. Reply was, "we're not, just another SA tool."

Database used by the SVS is not a periodic database and has no end date, and therefore can not expire.

galaxy flyer
11th Feb 2017, 01:17
Well, not on the installation I flew. We had a regular, but not AIRAC cycle, updates. It mist certainly could expire, but as it is not a required item for use, you coukd fly with it out of date. CHECK DATABASE STATUS was hard to miss on the MFD.