This paper describes an effort to create adaptive opponents for simulation-based air combat, where the opponents behave realistically while at the same time fulfilling instructional objectives. Three different models are developed to control the behavior of red pilots against (simulated) blue trainees in a set of 2v2 scenarios. These models are then evaluated on their tactical and instructional performance, with the machine-learning model performing on par with the two hand-constructed models. The contribution of this paper is to investigate technology and infrastructure enhancements that could be made to existing systems used for simulation-based air combat training.
CITATION STYLE
Ludwig, J., & Presnell, B. (2019). Developing an Adaptive Opponent for Tactical Training. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 11597 LNCS, pp. 532–541). Springer Verlag. https://doi.org/10.1007/978-3-030-22341-0_42
Mendeley helps you to discover research relevant for your work.