Primarily for general aviation discussion, but other aviation topics are also welcome.
  • 1
  • 21
  • 22
  • 23
  • 24
  • 25
  • 29
User avatar
By Rob P
#1664813
Ian Melville wrote:
I was wondering if this should be discussed elsewhere by those with the skillset and tools to investigate this issue rather than a public forum? But on the other hand, you don't have to read it :D


I drop into the thread from time to time, search amongst the technical guff for something I can understand, generally fail and go back to thinking that the PAW in my aircraft is a useful addition to window gazing and that's the limit of my comprehension.

Still a useful thread to have public though I suspect.

Rob P
gaznav, skydriller, kanga and 2 others liked this
#1664819
GrahamB wrote:
In a Mode S ES installation the NCAp value is calculated by the transponder based on values passed by the position source..


That's interesting.
An alternate source of the error 'could be' a transponder firmware bug miscalculating the NCAp value.
It's only now that we are able to see the live NCAp value through the uncertainty circles, so it is likely that this is not a new phenomenon (do doo beedodo) .
Perhaps it will be possible to correlate those with poor NCAp outputs to a particular transponder type?
gaznav, Nick liked this
User avatar
By gaznav
#1664822
I’ve been doing a bit of reading on this and I’m not so sure the display of Navigation Accuracy Category - Position (NACp) and in particular the Estimated Position Uncertainty (EPU) with large red circles is a good idea. This is why:

NATS/CAA GA ADS-B trial: https://nats.aero/blog/wp-content/uploa ... report.pdf

The following are taken from the Executive Summary...

The Navigational Accuracy Code for Position (NACp) quality indicator that reports the expected accuracy of the position reported, was on the whole found to be very conservative in the non- certified fleet; however this is better than over-estimating the accuracy capability.


When assessed against the ESASSP requirements, the mean horizontal position error (HPE) was measured to be 43.94m well within the required 300m and recommended 210m, Overall 99.84% of ADS-B HPE’s recorded by the non-certified fleet were below 300m [which is NACp=7], which was a higher percentage than that of the certified comparison fleet which recorded 99.7% and a mean HPE of 40.35m.


The accuracy of the non-certified GPS position reports was found to be very similar to that of the certified aircraft; however the trial did record several extremely large horizontal position errors. Investigation of these errors determined they were broadcast by one particular airframe and were caused when the reported longitude position swapped from negative to positive for an update. It is not expected that a single large error would lead to significant issues to ATC as a single report would be considered as an outlier by a surveillance tracker and not lead to credible corruption.


Later in the document it states:

Navigation Accuracy Code for position was assigned with 297,435 (87.19%) of the ADS-B reports.
Of these, the most frequent accuracy provided was NACp=3 equating to a reported horizontal positional accuracy of 2NM which was reported in 193,462 reports or 56.7% of all messages. It should be noted that 77,473 (22.71%) of reports received indicated NACp=10 corresponding to an accuracy bound of less than 15m. A summary bar chart of these values is provided in Figure 20 below.


So this is probably why I am seeing lots of 2nm circles from various aircraft if SkyDemon is displaying the NACp ‘circle of uncertainty’. Certainly if it is “the most frequent accuracy provided” as stated in the trial report.

Now here comes the killer statement in my mind (bear with it as it is deep in techno-talk):

The quality indicator NACp=10 indicates that the position broadcast by the airframe should be within 15m of the aircrafts actual position. As this assessment uses radar track data as the baseline ‘truth track’ it is not possible to definitively assess to the reported accuracy to 15m (a DGPS fitted to the airframes would be required for this task), however a plot of the HPE of all airframes reporting NACp=10 would be anticipated all be within the 300m tolerance; given the expected accuracy of the radar track. Figure 21 below shows the cumulative plot of HPE’s for all GA airframes reporting NACp=10 (red) and NACp=3 (green).

It can be seen that 99% of airframes reporting NACp=10 are within a HPE of 150m. However, the largest reported HPE for the NACp=10 sample was 1,126m. The mean HPE for NACP=10 reports was 34.43m, with a standard deviation of 28.94m.

For the airframes reporting NACp=3, which indicates that the aircraft are reporting a position within 2NM of their actual position (see Table 3), the mean HPE was 38.54m with a standard deviation of 29.4m and a maximum HPE of 1,080.76m.

The results and Figure 21 below illustrates that although the error bound NACp=10 is much less than for NACp=3, the overall the magnitude of the HPE’s are very similar suggesting that the for the majority of the time, NACp=3 airframes are extremely conservative at reporting the accuracy capability of the GPS source under fault free conditions.


So a trial that used RADAR to verify position versus the reported “position of uncertainty” found the accuracy reported to be “extremely conservative”. Indeed the uncertified systems reporting NACp = 3 (up to 2nm) were found to be nearly as good as NACp = 10 (up to 10 metres). Therefore, by displaying a circle around the aircraft we may well be misleading what is happening in real life - that is my experience recently, where the aircraft are pretty much pegged to the position they are reporting. Considering that these devices are being used for “Collision Awareness” rather than “Avoidance” (there is a significant difference), then is the displaying of these circles likely to lead to confusion and also missing other things displayed (like a 2nm ATZ or not)?

@leemoore1966 Provided a useful reference to the GDL90 format on “Page 27 in the PDF”. Where it states as a footnote for Multi-Function Displays (MFDs) like SkyDemon:

MFD Recommendation: Targets with either a NIC or NACp value that is 4 or lower (HPL >= 1.0 NM, or HFOM >= 0.5 NM) should be depicted using an icon that denotes a degraded target.


So @Tim Dawson can we please have a different icon, and not a circle, to depict NACp<=4? Otherwise, when I go to the LAA Rally this year my SkyDemon screen is going to be obliterated up to 10nm out from Sywell with red overlapping circles! If nothing else, could you give us a user selectable option?

I hope this helps?
GrahamB, dewidaniels, T67M and 1 others liked this
User avatar
By Cub
FLYER Club Member  FLYER Club Member
#1664823
I have now had the opportunity to take a look at fairly emphatic source of surveillance data for the time and place of Gazzer’s screenshot of G-EDGA with a red circle. While I can see a fairly solid primary return, I cannot see any displayable Mode S data perhaps indicating that my big boys surveillance system is doing exactly what it should do and not depicting data without the appropriate integrity/resolution values? This being the case, we could debate the depiction (colours, shading) or not of such a target in SkyDemon but it would appear that the philosophy behind handling such traffic is exactly as requested by Lee and implemented by Tim.

I should also point out that I have now observed hundreds of aircraft both big and small via SkyDemon and SkyEcho and still not observed any red circles.
gaznav liked this
User avatar
By GrahamB
FLYER Club Member  FLYER Club Member
#1664828
Cub wrote:I have now had the opportunity to take a look at fairly emphatic source of surveillance data for the time and place of Gazzer’s screenshot of G-EDGA with a red circle. While I can see a fairly solid primary return, I cannot see any displayable Mode S data perhaps indicating that my big boys surveillance system is doing exactly what it should do and not depicting data without the appropriate integrity/resolution values?

Surely pure Mode S doesn't carry position data, and therefore also integrity data? Or did you mean Mode S ES/ADS-B data?
User avatar
By kanga
#1664829
skydriller wrote:..All this tech-speak from some of you is making my head hurt, can we tone it down a little and explain a little more in general terms?


other aircraft symbol:

a. without circle: "here be other aircraft"
b. with circle: "there be dragons" ? :)
gaznav, skydriller, patowalker and 3 others liked this
User avatar
By Tim Dawson
SkyDemon developer
#1664830
Gaz, I was sent that report overnight by another reader here and it made really interesting reading. It’s great to have the actual figures from a trial to hand.

It’s obvious that most transponders are really conservative in what they broadcast for NACp. Also, some appear to get it wrong completely, broadcasting absurd values.

I am going to remove the uncertainty circles from SkyDemon. Given the source data, they are not helpful.
gaznav, kanga, G-BLEW and 8 others liked this
User avatar
By gaznav
#1664831
Thanks Cub, as ever.

Obviously it’s not every aircraft that is showing a ‘circle of uncertainty’, but a few occasional ones. That’s why I am convinced it is not a device reception issue - also my flight where I first saw it at the end of the last year and the ones in the last few days were with a different device (same size device but different revision). The reason for switching out was that the earlier pre-production device I had was not configured for FLARM reception and the one I now have has been (and works really nicely, I hasten to add).

So that still leads me to believe that some aircraft have been reporting NACp values less than or equal to 4. Whether that is a correct value or not, I don’t know. However, if it is, then I still want to know their position because 95% of the time they will be roughly where they are reporting they are. However, I believe that displaying large, up to 10nm, red circles around various ADS-B aircraft with NACp<=4 will potentially see quite a few aircraft, probably those with non-certified GPS/GNSS, starting to obscure the map in SkyDemon. That could lead to misread airspace boundaries and height/altitude details that in turn could lead to an unwanted side effect of infringement of airspace/ATZs/avoids.

That is why I think a more subtle symbol change - colour, shape or even occulting would be my preference. Does anyone know what ForeFlight or EasyVFR do for ADS-B reports with a NACp<=4 in accordance with the GDL90 protocol document that Lee shared?
User avatar
By gaznav
#1664833
Tim Dawson wrote:Gaz, I was sent that report overnight by another reader here and it made really interesting reading. It’s great to have the actual figures from a trial to hand.

It’s obvious that most transponders are really conservative in what they broadcast for NACp. Also, some appear to get it wrong completely, broadcasting absurd values.

I am going to remove the uncertainty circles from SkyDemon. Given the source data, they are not helpful.


Thanks Tim, that would allay my concerns just replied to Cub. Could you have a think about how you might more subtly display NACp<=4 as per the GDL90 document? If that is too tricky, then I’m not that concerned.

By the way, Tim, my FLARM reception issue that I PM’d you about is all sorted. It seems I had an early device that had not been configured for FLARM. So now my SkyEcho 2 and SkyDemon combo is working really well with the £30 FLARM subscription that you set up. Thank you :thumleft:
User avatar
By skydriller
FLYER Club Member  FLYER Club Member
#1664834
Ian Melville wrote:I was wondering if this should be discussed elsewhere by those with the skillset and tools to investigate this issue rather than a public forum? But on the other hand, you don't have to read it :D


I kinda skim this thread when all the acronyms start buzzing about and concentrate on the pictures... :mrgreen:

Id like to thank those of you that occasionally put the meaning of something longhand then explain it...

Regards, SD..
gaznav liked this
User avatar
By neilmurg
#1664836
gaznav wrote:.... displaying large, up to 10nm, red circles around various ADS-B aircraft with NACp<=4 will potentially see quite a few aircraft, probably those with non-certified GPS/GNSS, starting to obscure the map in SkyDemon...
AIUI
1) NACp is an assessment of GPS accuracy due to number of satellites 'seen', atmospheric effects (& SBAS input?). Certified/non-certified (or badly installed) can have an effect but there doesn't seem to be direct evidence for that?
2) Red circles partially obscure the map, and don't have to be drawn with a radius directly related to the NACp translation, especially as the aircraft are empirically close to their reported position.

We don't always (or even much) agree on EC solutions Gaz, but this has been an excellent report on what you've observed aloft. I've learnt a lot from it. Thanks and 'good job!' as our cousins would say.
#1664838
I have been told if the ADSB is broadcasting a SIL=0. Then the NACp should not be used
In which case it is indeed a receiver problem, as the receiver is passing NACp over the GDL90 interface for an ADSB source providing SIL=0
gaznav liked this
  • 1
  • 21
  • 22
  • 23
  • 24
  • 25
  • 29