A Game Theoretic Method for Two-Team Multi-Player Autonomous Racing

This letter explores an autonomous driving competition between two teams, where the number of members in one team is greater than or equal to the other team's, but their maximum speed is lower. The letter proposes a hierarchical decision-making approach to address how the slower team can compete against their opponents through collaboration. Initially, in the first layer, the teams are paired using weighted bipartite graph matching, followed by the task reassignment to address opponents who pose a significant threat. In the second layer, each player computes its optimal path through the matching-based iterated best response, taking into account the opponents determined by the first-layer matching. Through this hierarchical decision-making module, each player assumes specific roles and tasks, enabling cooperative blocking, aiding lagging teammates to catch up or contributing to the team's leading member to amplify their advantage. This method aims to increase the team's chances of winning competitions at a higher rate.