Analytical Framework for Evaluating Traffic Capacity Impacts of Electric Vehicles' Regenerative Braking Dynamics

Abstract

Regenerative braking (RB) significantly influences electric vehicle (EV) car-following (CF) dynamics, yet traditional traffic-flow models rarely capture these effects. We introduce a comprehensive empirical dataset comprising 197.5 hours of driving data from 25 drivers across eight EV models to systematically investigate regen-induced CF behaviors. Two primary CF patterns emerge: (i) steady-state scenarios where EVs use regenerative braking and subsequently re-accelerate to equilibrium speeds with larger spacing, and (ii) dynamic scenarios involving lead oscillations, characterized by a distinctive three-phase CF process-regenerative deceleration, transitional plateau, and rapid re-acceleration. The paper's main contribution is the development of an analytical framework that models these EV-specific CF behaviors and quantifies their impacts on traffic capacity. We derive closed-form expressions for the established $\eta$ function from the literature, explicitly quantifying EV driving deviations from stable CF defined by Newell's CF model and assessing their implications for roadway capacity. Validation against empirical data and simulation confirm the model's accuracy ($R^2=0.96$) in replicating real-world $\eta$ trajectories. Sensitivity analyses demonstrate that increased RB intensity, prolonged transitions, and shorter reaction delays significantly raise values and cumulative capacity losses. These findings highlight a clear trade-off between enhanced energy recovery through RB and reduced traffic efficiency, providing critical insights for EV-aware traffic modeling, control strategies, and transportation policy.