Crashed, Modeled, then Rescued: AI Algorithms Reduce Rescue Time for Crash Survivors

Summary

This is an abstract from the "Material Aspects of Global Conflict" session, at the 90th annual meeting of the Society for American Archaeology.

A major topic of archaeological research includes the modeling of human movement across diverse landscapes, often in terms of how geography can facilitate or impede mobility. On an operational level, modeling human movement allows archaeologists to determine likely travel corridors that may aid in the identification of new sites and features, or assess the connectivity between regions. Standard modeling tools, such as Least Cost Path (LCP) analysis, require a full geographic knowledge of the interlaying region between start and end points. However, a crash survivor might not have knowledge of the local geography. An AI algorithm, such as A-Star Algorithm, can mimic human movement with its heuristic-based local decision-making process throughout the survivor’s journey. Combined with circuit analysis theory, calculated using Circuitscape, A-Star Algorithms can better model an individual's path compared to traditional LCP and other search algorithms, like gradient descent. We utilize World War II crash survivor datasets that consist of the crash point, path taken, and rescue location to compare travel models. These results have a broad impact within modeling human behavior in the past as well as contemporary cases, especially in consideration of the ongoing search for missing service members and crash survivors from past world conflicts.

This Resource is Part of the Following Collections

• Material Aspects of Global Conflict   •
• Society for American Archaeology 90th Annual Meeting, Denver, CO (2025)

Keywords

Record Identifiers

Abstract Id(s): 52059