TY - GEN
T1 - Model validation and exploration of new concepts in aviation safety
AU - Bareither, Chad A.
AU - Luxhøj, James T.
PY - 2005
Y1 - 2005
N2 - The world of commercial aviation is extremely important to the infrastructure of countless industries within and outside of the United States. Beginning with former Vice President Al Gore's "Safer Skies Program" in 1997, the safety of domestic air travel has become an important area of research and investigation. As a result, NASA has been developing numerous new technological products to aid in the improvement of airline safety. Under contract from NASA's Aviation Safety and Security Program (AvSSP), Rutgers University has been modeling aircraft accident cases with Bayesian Belief Networks (BBNs) to identify accident precursors and to assess the potential risk mitigating effects of new technology products in these types of accidents. BBNs develop mathematical models through graphical representation of the inter-dependency of events or system components. This methodology has become increasingly more established for the representation of knowledge and reasoning under uncertainty. Systems in which data cannot be collected or do not exist (e.g. human factors) provide a natural fit for the Bayesian approach. In addition, calculations, both forward and backward, of how events influence each other and combinatorial effects of components are made efficient with this methodology. This paper begins with validation of Rutgers' existing model structure through external audit and then presents advanced BBN analysis with both sensitivity and importance aspects to identify the most influential accident precursors. The preliminary findings are used to propose and explore new technology and system concepts to improve aviation safety.
AB - The world of commercial aviation is extremely important to the infrastructure of countless industries within and outside of the United States. Beginning with former Vice President Al Gore's "Safer Skies Program" in 1997, the safety of domestic air travel has become an important area of research and investigation. As a result, NASA has been developing numerous new technological products to aid in the improvement of airline safety. Under contract from NASA's Aviation Safety and Security Program (AvSSP), Rutgers University has been modeling aircraft accident cases with Bayesian Belief Networks (BBNs) to identify accident precursors and to assess the potential risk mitigating effects of new technology products in these types of accidents. BBNs develop mathematical models through graphical representation of the inter-dependency of events or system components. This methodology has become increasingly more established for the representation of knowledge and reasoning under uncertainty. Systems in which data cannot be collected or do not exist (e.g. human factors) provide a natural fit for the Bayesian approach. In addition, calculations, both forward and backward, of how events influence each other and combinatorial effects of components are made efficient with this methodology. This paper begins with validation of Rutgers' existing model structure through external audit and then presents advanced BBN analysis with both sensitivity and importance aspects to identify the most influential accident precursors. The preliminary findings are used to propose and explore new technology and system concepts to improve aviation safety.
UR - http://www.scopus.com/inward/record.url?scp=33644526921&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=33644526921&partnerID=8YFLogxK
M3 - Conference contribution
SN - 1563477408
SN - 9781563477409
T3 - Collection of Technical Papers - AIAA 5th ATIO and the AIAA 16th Lighter-than-Air Systems Technology Conference and Balloon Systems Conference
SP - 1254
EP - 1265
BT - Collection of Technical Papers - AIAA 5th ATIO and the AIAA 16th Lighter-than-Air Systems Technology Conference and Balloon Systems Conference
T2 - AIAA 5th Aviation Technology, Integration, and Operations Conference, ATIO and the AIAA 16th Lighter-than-Air Systems Technology Conference and Balloon Systems Conference
Y2 - 26 September 2005 through 28 September 2005
ER -