FIELD EVALUATION OF DESCENT ADVISOR TRAJECTORY PREDICTION ACCURACY

Steven M. Green, and Robert A. Vivona


Abstract

The Descent Advisor (DA) automation tool has undergone a series of field
tests at the Denver Air Route Traffic Control Center to study the feasibility
of DA-based clearances and procedures.  The latest evaluation, conducted in
the fall of 1995, expanded the operational nature of DA testing to include a
wider variety of test conditions.  A total of 197 commercial flights from 
three airlines participated in the study over twenty-three days of testing.
Aircraft included large and heavy jet transports, both conventional and 
flight-management-system-equipped, and turboprop commuter types.  The primary
objective was to measure DA trajectory prediction accuracy for use in 
validating DA metering advisories and developing conflict-probe error models.
Previous evaluations, involving large jet type only, demonstrated an arrival 
time prediction accuracy within 20 sec.  The 1995 test results indicate a 
mean error of 0.5 sec late with a standard deviation of 14.3 sec. The least
variation was found for flight-management-system-equipped jets with a standard
deviation of 11.9 sec compared to 15.2 and 15.4 sec for conventional-equipped 
jets and turboprop types respectively.  This paper describes the test and
presents an analysis of the descent trajectory prediction accuracy in terms 
of error in the horizontal profile, altitude profile, and arrival time.