PPRuNe Forums - View Single Post - So WestJet almost puts one of their 737 in the water while landing at St-Maarten...
Old 30th Mar 2017, 11:53
  #215 (permalink)  
underfire
 
Join Date: Aug 2013
Location: PA
Age: 59
Posts: 30
Likes: 0
Received 0 Likes on 0 Posts
Other than the 25' granularity, altitude reported by ADS-B should be as accurate (subject to adjustment for QNH) and timely as your altimeter, given that both are coming from the same source and subject to the same errors.
and subject to the ADSB reporting sequencing.

Where did you get the idea that there is a delay of anything like 8 seconds (not that we're particularly interested in absolute time values in this instance) ? Where is the data being buffered in the meantime ?
As we know, the system shows where the ac was, not where it is, the kalman filter estimates where the ac is.
8 seconds is what is typically used in the latency calcs, and the information is buffered in the INS. (the net has significant papers on latency issues with GPS and aircraft)
Each low band sat signal takes about 3 seconds to receive, (and in realty, it may be 5 seconds with correction) with multiple sats required, it takes a while to receive all of them, process them.
The kalman filter, tracks the data, determines the ac location where it was, and interpolates the data to where the ac actually is. That calculation has a latency of about 8 seconds. Within that 8 seconds, if there are changes to the horizontal flightpath, such as a turn, or even crosswinds, or vertically with descent, the data will not be accurate.


After all of that, the ac systems uses a lookup feature from the internal database for altitude. Depending on the manufacturer, that methodology is vastly different. That is then compared with the rad alt and baro data, and viola, that is your altitude as reported. (ohhh the algorithms) You want it to figure out the descent profile, in a turn, (a 3d spiral) using a lookup feature to an internal database with an 8 second delay? oh yea.

The descent on final approach, again based on 8 seconds latency, will not be accurate at all, even with continuous descent. Too much happens in 8 seconds for the ADSB data to be accurate, ie how much has the ac descended in 8 seconds based on an 8 second delay with 25 foot reporting....

The data transmitted through ADSB is far different that what you have on the ac, in both reporting and granularity.

Last edited by underfire; 30th Mar 2017 at 12:19.
underfire is offline