The article ‘Beyond Simulated Drivers: Evaluating the Impact of Real-World Car-Following in Mixed Traffic Control’ studies the effect of human driving behaviors on traffic congestion. It suggests that Human-driven Vehicles (HVs) can intensify naturally occurring traffic disruptions, leading to congestion. The study incorporates real-world human driving trajectories in a simulation with Robot Vehicles (RVs) to mitigate these issues. The authors also introduce a reinforcement learning-based RV that uses a neural network to optimize safety, stability, or efficiency in the presence of diverse human driving behaviors. The proposed RVs are evaluated in mixed traffic control environments under various configurations and densities.
Publication date: 22 Nov 2023
Project Page: Not Provided
Paper: https://arxiv.org/pdf/2311.12261
