Update Interval Performance and Outlier Exclusion Methods for Aircraft Surveillance Systems

Aireon’s space-based Automatic Dependent Surveillance-Broadcast (ADS-B) system provides global air traffic surveillance data that is delivered to various customers around the world for use within the aviation industry [1] [2]. The key performance indicators of Aireon’s service are Availability, Latency, and Update Interval (UI). Different methods have been explored and/or are used to evaluate the system performance. Focusing on update interval, the most contentious of the performance metrics, the following methods are available for evaluation: a per-aircraft evaluation, a tiled/gridbased evaluation, and a total airspace evaluation.The per aircraft evaluation uses a percentage and/or number of aircraft passing the requirements to determine performance for an airspace. The tiled evaluation separates the airspace into smaller regions, or tiles, which are each evaluated based on the requirements. A weighting approach can be applied to this method which is based on traffic density within the tile evaluated. The total airspace evaluation uses an approach where all update intervals from all aircraft are analyzed to determine airspace performance. This method is currently used by Aireon and is the standard for Aireon’s airspace UI performance evaluation. This paper explores each method using different airspaces to explore the pros and cons between each approach.Additionally, this paper will evaluate an in-depth aircraft analysis done to investigate why certain aircraft perform poorly compared to others and how these poor performers can affect update interval evaluations using the various methods described above. This investigation is necessary to accurately assess volume performance regardless of the method by defining appropriate exclusion categories and criteria. Mathematical models, and aircraft trends in data, are explored to determine a root cause hypothesis for avionics or environmental issues for the aircraft. The goal of this analysis is to highlight the frequency and impact of these aircraft-centric issues on a surveillance system and to explore ways to focus the performance measurements on the applicable aircraft population while separating out the population of outliers.