Deep Fictitious Play-Based Potential Differential Games for Learning Human-Like Interaction at Unsignalized Intersections

TL;DR

This paper introduces a novel framework using Deep Fictitious Play to learn human-like driving behaviors at unsignalized intersections by modeling vehicle interactions as a Potential Differential Game, trained on real driving data.

Contribution

It is the first to apply Deep Fictitious Play to train interactive driving policies using a Differential Game reformulated as a Potential Differential Game, with theoretical convergence guarantees.

Findings

Successfully learned human-like driving policies from real data

Captured variations in driver aggressiveness and preferences

Achieved satisfactory performance in realistic intersection scenarios

Abstract

Modeling vehicle interactions at unsignalized intersections is a challenging task due to the complexity of the underlying game-theoretic processes. Although prior studies have attempted to capture interactive driving behaviors, most approaches relied solely on game-theoretic formulations and did not leverage naturalistic driving datasets. In this study, we learn human-like interactive driving policies at unsignalized intersections using Deep Fictitious Play. Specifically, we first model vehicle interactions as a Differential Game, which is then reformulated as a Potential Differential Game. The weights in the cost function are learned from the dataset and capture diverse driving styles. We also demonstrate that our framework provides a theoretical guarantee of convergence to a Nash equilibrium. To the best of our knowledge, this is the first study to train interactive driving policies using Deep Fictitious Play. We validate the effectiveness of our Deep Fictitious Play-Based Potential Differential Game (DFP-PDG) framework using the INTERACTION dataset. The results demonstrate that the proposed framework achieves satisfactory performance in learning human-like driving policies. The learned individual weights effectively capture variations in driver aggressiveness and preferences. Furthermore, the ablation study highlights the importance of each component within our model.