ABSTRACT
This paper proposes a dynamic game-theoretic framework that is used as an analytical tool and uni fying perspective for a wide class of problems in mo tion planning. This approach is inspired by the founda tion laid by configuration-space concepts for basic path planning. In the same manner that configuration-space concepts led to substantial progress in path planning, game-theoretic concepts provide a more general founda tion which can incorporate any of the essential features of path planning, sensing uncertainty, decision theory, bounded-unceriainty analysis, stochastic optimal con trol, and traditional multiplayer games. By following this perspective, new modeling, analysis, algorithms, and computational results have been obtained for a vari ety of motion planning problems including those involv ing uncertainty in sensing and control, environment uncertainties, and the coordination of multiple robots.
