关键词: |
Intelligence collection, Unmanned aerial vehicles, Lessons learned, Algorithms, Humanitarian assistance, Mathematical models, Machine learning, Disasters, Missions, Predictions, Military planning, Cognition, Human performance tests, Mission plan performance prediction, Conceptual spaces, Feature engineering, Scope(separating cognition performance from execution environment) |
摘要: |
The goal of SCoPE was to predict mission plan performance, taking into consideration mission plan characteristics as well as cognitive and environmental factors influencing the plan. Given the large number of features that go into any one particular mission, the team proposed to use a conceptual spaces based model to identify features spaces that correspond to successful missions. To test proposed algorithms, the project initially aimed to use datasets related to HADR missions. An HADR simulation model was created using AnyLogic for the OtK program, and this model was used to generate datasets for the SCoPE program. Developing a validated simulation model, capable of generating large sets of simulation runs, proved challenging and largely unsuccessful. Highlights of the work with the HADR simulation model are presented first, including lessons learned from working with complex simulation models. |