Crash site prediction paper out in the Journal of Aeronautics, Astronautics and Aviation
October 30, 2026 · 2 min readOur paper on probabilistic aircraft crash site prediction is now published, after a spell as to appear:
The method starts from the last ADS-B message an aircraft broadcast — position, velocity, altitude, and the quality indicators that bound how much each of those can be trusted — and propagates a ballistic trajectory through a wind profile by Monte Carlo simulation.Simanjuntak, T., Nasution, M. R. E., & Favian, E. (2026). Probabilistic crash site prediction model using ADS-B data and Monte Carlo simulation. Journal of Aeronautics, Astronautics and Aviation, 58(10S), 2045–2061. https://doi.org/10.6125/JoAAA.202610_58(10S).26
The result is a probability map of the impact point, summarised as circular error probable (CEP) regions that a search-and-rescue team can use to define a search area.
We tested it against three Indonesian accidents: Sriwijaya Air SJ182, Lion Air JT610, and AirAsia QZ8501. The agreement ranges from strong for the rapid-descent cases to poor for QZ8501, where the aircraft flew an extended non-ballistic path after signal loss — a limitation the paper discusses at some length.
The code behind the paper is released as probimpact, listed on the Codes page. The full list of my papers is on the Research page; the citation is also available as BibTeX.