Application of Isolation Forest for Detection of Energy Anomalies in ADS-B Trajectory Data
Satvik G. Kumar, Samantha J. Corrado, Tejas G. Puranik, Dimitri N. Mavris · AIAA SCITECH 2022 Forum · 2022
View Video Presentation: https://doi.org/10.2514/6.2022-2441.vid In recent years, the aviation industry has seen a large increase in the volume of operations. The maintenance and improvement of safety at acceptable levels is one of the most important concerns in civil aviation operations. Reactive methods to aviation safety improvement are being augmented with proactive and predictive approaches that leverage large amounts of routinely collected aviation data. Due to the increased availability of airborne sensor data and improvements in computing power, application of machine learning methods to various aviation safety problems for identifying, isolating, and reducing risk has gained momentum. Work in this domain focuses on identifying anomalies or abnormal operations as a first step towards identification of potentially risky situations using aircraft sensor data, such as ADS-B trajectory data. Specifically, detecting anomalies in energy metrics derived from ADS-B trajectory data has been an active area of research. While various machine learning techniques have been leveraged to detect anomalies in the energy dimension, many, such as those that are clustering-based, rely on a distance function to be defined. However, definition of a distance function is not straightforward and may be considered to be a limitation. Recently, use of the Isolation Forest (iForest) algorithm has become prevalent in other domains. The iForest algorithm does not rely on the definition of a distance function. Rather, the iForest algorithm recursively generates partitions of the data set to isolate anomalies. In this paper, a novel application of the iForest algorithm is proposed to identify anomalies in arriving aircraft operations. The application is demonstrated on 120 days of data collected for San Francisco International Airport as a case study. Energy states between nominal and anomalous flights are compared and an analysis between anomalous flights and unstable approach is conducted. The developed method has the potential to aid air traffic controllers in identifying abnormal trajectories such that these trajectories may be further investigated to determine a level of risk.