Air Traffic Control:
FAA Has Not Fully Assessed Its User Request Evaluation Tool
AIMD/RCED-98-59R: Published: Feb 23, 1998. Publicly Released: Feb 23, 1998.
- Full Report:
Pursuant to a congressional request, GAO reviewed the Federal Aviation Administration's (FAA) efforts to assess its User Request Evaluation Tool (URET), a prototype for detecting potential conflicts between aircraft paths, focusing on: (1) evaluations of air traffic controllers' experiences using URET and quantitative analyses of URET performance-assessments of conflict likelihood, conflict warning time, and trajectory accuracy and stability; and (2) documentation supporting URET's false and missed alert rates. GAO analyzed URET assessments by FAA and Mitre Corporation and did not independently verify the performance data.
GAO noted that: (1) FAA first deployed URET in January 1996 at its Indianapolis en route center, to gain operational experience with its new capabilities, determine operational and technical requirements for a full-scale production system, and demonstrate potential benefits; (2) because of its experience with URET, in April 1997 senior managers decided to acquire a deployable tool for detecting and resolving potential conflicts, called the Initial Conflict Probe (ICP); (3) while no clear industry or government standards exist for evaluating prototypes, experts in conflict probe technology agree that a comprehensive evaluation of a conflict probe capability should include both qualitative and quantitative analyses; (4) qualitative analyses have been conducted over the almost 2 years that URET has been installed at Indianapolis; (5) fewer data are available on a quantitative level and have not been independently validated; (6) in October 1997, FAA's William J. Hughes Technical Center began independently validating URET performance using a simulator; (7) this effort, however, had not been completed by the time GAO completed its audit work in January 1998; (8) FAA's Technical Center also evaluated URET algorithms and concluded that the assumptions and approximations used in URET were reasonable for prototype software, but it recommended that these assumptions and approximations be independently verified and validated; (9) quantitative analyses of URET are limited in two ways--they are based on traffic data in only two centers and they were not independently validated; (10) FAA's en route product team, responsible for developing and implementing ICP, acknowledges that the limited amount of information on URET performance constitutes a technical risk to developing ICP; (11) to mitigate the risk of developing ICP with limited URET performance information, FAA is drafting ICP performance requirements based on existing URET performance data; and (12) FAA has also drafted a conflict probe requirements verification plan, which officials expect to finalize in late February 1998.