Discovering Anomalous Aviation Safety Events Using Scalable Data Mining Algorithms

The world-wide civilian aviation system is one of the most complex dynamical systems ever created. Most modern commercial aircraft have onboard flight data recorders (FDR) that record several hundred discrete and continuous parameters at approximately 1 Hz for the entire duration of the flight. This data contains information about the flight control systems, actuators, engines, landing gear, avionics, and pilot commands. In this paper we discuss recent advances in the development of a novel knowledge discovery process consisting of a suite of data mining techniques for identifying precursors to aviation safety incidents. The data mining techniques include scalable multiple kernel learning for largescale distributed anomaly detection. A novel multivariate time series search algorithm is used to search for signatures of discovered anomalies on massive data sets. The process can identify operationally significant events due to environmental, mechanical, and human factors issues in the high dimensional Flight Operations Quality Assurance (FOQA) data. All discovered anomalies are validated by a team of independent domain experts. This novel automated knowledge discovery process is aimed at complimenting the state-of-theart human-generated exceedance-based analysis that fails to discover previously unknown aviation safety incidents. In this paper we discuss the discovery pipeline, the methods used, and some of the significant anomalies detected on real-world commercial aviation data.