Identifying Instantaneous Anomalies in General Aviation Operations

Abstract

Quantification and improvement of safety is one of the most important objectives among the General Aviation community. In recent years, data mining techniques are emerging as an important enabler in the aviation safety domain with a number of techniques being applied to flight data to identify and isolate anomalous (and potentially unsafe) operations. There are two types of anomalies typically identified - flight-level (where the entire flight exhibits patterns deviating from nominal operations) and instantaneous (where a subset or few instants of the flight deviate significantly from nominal operations). Energy-based metrics provide measurable indications of the energy state of the aircraft and can be viewed as an objective currency to evaluate various safety-critical conditions across a heterogeneous fleet of aircraft. In this paper, a novel method of identifying instantaneous anomalies for retrospective safety analysis using energy-based metrics is proposed. Each data record is split by sliding a moving window across the multi-variate series of evaluated energy metrics. A mixture of gaussian models is then used to perform clustering using the values of energy metrics and their variability within each window. The trained models are then used to identify anomalies that may indicate increased levels of risk. The identified anomalies are compared with traditional methods of safety assessment (exceedance detection).