|Donald A Sofge
The U.S. Naval Research Laboratory is interested in investigating the development of simulation frameworks for teams of autonomous underwater vehicles (AUVs) using game theoretic approaches for decision-making. The goals of the simulation frameworks are twofold. First, we seek to validate and quantify the effectiveness of theory extensions made in game-theoretic based algorithms adapted for underwater use and resource allocation. Second, the simulations will compute playbooks of best responses for AUVs to reference for autonomous decisions based on simulation results of naval scenarios. Extensions examined will include effectiveness of graceful and autonomous switching from one set of strategies to other sets under differing deployment events, or due to newly sensed information. Research will also include designing playbooks for multi-agent teams that are quickly searchable and selectable by autonomous systems for effective autonomous team decision-making.
1. Fang, F., Nguyen, T.H., Pickles, R., Lam, W.Y., Clements, G.R., An, B., Singh, A., Schwedock, B.C., Tambe, M., and Lemieux, A. (2017), Paws—a Deployed Game-Theoretic Application to Combat Poaching, in AI Magazine, edited, pp. 23-36, doi:10.1609/aimag.v38i1.2710.
2. Lou, J., Smith, A.M., and Vorobeychik, Y. (2017), Multidefender Security Games, IEEE Intelligent Systems, 32(1), 50-60, doi:10.1109/MIS.2017.11.