MFD Evolution and Nash Right-of-Way Game of Robotaxis in Mixed Traffic

Abstract

The transition toward fully autonomous transportation introduces a complex intermediate phase characterized by mixed traffic flows, where Robotaxis and human-driven vehicles must coexist. This paper investigates the macroscopic implications of microscopic interactions between these two distinct agents, specifically focusing on the negotiation of right-of-way at unsignalized intersections and merging points. By modeling these interactions as a non-cooperative Nash Right-of-Way Game, we analyze how local equilibrium strategies aggregate to influence the network-wide Macroscopic Fundamental Diagram (MFD). We employ an agent-based simulation framework to explore the evolution of the MFD under varying penetration rates of Robotaxis and different payoff configurations in the game-theoretic model. Our findings suggest that while conservative Nash strategies adopted by Robotaxis can initially degrade network capacity due to the hesitation effect, a critical mass of cooperative autonomous agents eventually linearizes the congested branch of the MFD, reducing hysteresis and improving flow stability. The study provides a theoretical bridge between micro-level game-theoretic decision-making and macro-level traffic flow theory, offering insights for policy formulation during the transition to autonomy.