Modeling and utilization of synthetic data for improved automation and human-machine interface continuity

Abstract

System degradation due to sensor failures has been a contributing factor in many aircraft accidents. For example, an autopilot (A/P) disconnect due to incorrect airspeed data was a primary factor in the events leading to the crash of Air France Flight 447 (AF447) into the Atlantic Ocean. Similarly, on Turkish Airlines Flight 1951 (TK1951) a faulty radio altimeter lead to an unexpected retard of the autothrottle (A/T), and consequent lack of energy state and flight mode awareness by the flight crew resulted in a crash on approach to Amsterdam. This paper addresses a method to increase the continuity of the sensor information that drives the aircraft automation modes and human-machine interfaces by utilizing dissimilar sensors and systems to synthesize critical data in case of sensor failure. The two main topics in this paper are data synthesis methods in the event of sensor failures and the modelling of the automation mode logic based on the available literature. The latter topic provides the information requirements for continued operation of the automation and, therefore, identifies what information would require a synthesized replacement in the event of failure. All topics will form an important automation logic component of a simulation tool that is being built to evaluate predictive altering methods that are developed to improve flight crew's aircraft state awareness.