Interaction-Aware Merging in Mixed Traffic with Integrated Game-theoretic Predictive Control and Inverse Differential Game

This paper presents an interaction-aware motion planning and control framework for time-critical traffic scenarios in which interaction with vehicles driven by humans is required. For safe motion planning the proposed method considers interaction between the automated driving system and other vehicles using game theory. The framework includes a novel inverse differential game based on a LSTM to estimate the human driver’s objective function online. Then, a game-theoretic predictive controller utilizes these estimates for controlling the automated driving system and predicting the trajectory of the human-driven vehicle. The developed framework is validated in several safety-critical scenarios and testing conditions using CarSim high-fidelity simulations including human-in-the-loop case studies with six different test subjects.