Game Theoretical Motion Planning: Tutorial ICRA 2021

Abstract

The scope of this manuscript is to ease the initial steps for a reader interested in modeling the strategic nature of motion planning problems in a multi-agent environment. Historically, game theory has been devoted to studying rational decision-making in numerous fields: social science, economics, system science, logic, computer science, and many more. Today, as robots leave the factory floors for a more complex world, we do believe that many of the game theoretical concepts are well suited to capture the dynamic and interactive nature of multi-agent motion planning. The promise (and hope) is that explicitly taking into account the others' decision making in its own, endows standard techniques with a richer descriptive power. If this promise holds true, a better decision making for our robots will facilitate a seamless integration in our society.