A connectionist approach to airliner safety

The present paper introduces the system SINCO-Flightsim, an intelligent hybrid symbolic connectionist approach for the treatment of emergency situations on commercial airliners, currently available as a computer simulation. The system´s main focus is on human failure, which accounts for a major part of accidents and incidents in airline traffic. The underlying architecture, using the biologically more plausible learning algorithm GeneRec and contrasting it to learning via back-propagation is presented. System modules are described as well as the learned data sets. In its first version, the system provides a series of typical sensors and means of interaction for treating emergency situations successfully. The respective results are outlined in this paper. We trust that the approach has the potential to contribute to airliner safety as it takes major stress factors off the pilots´ shoulders and helps treating emergency situations in a more objective manner.