Eunjeong Hyeon

Current Institution: University of Michigan


Bio: Eunjeong Hyeon is a postdoctoral appointee at Argonne National Laboratory. She received her B.S. and M.S. in Aerospace Engineering from Korea Advanced Institute of Science and Technology in 2015 and 2017, respectively. She received her Ph.D. degree in Mechanical Engineering from the University of Michigan, Ann Arbor. Her current research focuses on trajectory forecasting and optimal control of connected automated vehicles to improve vehicle energy efficiency. She has participated in various research projects sponsored by the U.S. Department of Energy and National Science Foundation. In addition, she received the Automotive and Transportation Systems (ATS) Best Paper Award at American Control Conference in 2021.

Abstract: Speed Forecasting Strategies for the Energy-Optimal Car-Following of Connected and Automated Vehicles

Accurate previews of the preceding vehicle’s future trajectories are essential for automated driving in car-following scenarios to minimize energy consumption. This work proposes strategies for forecasting the preceding vehicle’s speed for the eco-driving control system of connected and automated vehicles. Accurate and economical predictors are developed based on linear-regression-based algorithms targeting short and midlength horizons. The proposed predictors avoid the need for an exhaustive training process by enabling instant learning, and can provide high accuracy by leveraging information obtained by vehicle communication. Then, the novel methodologies for designing data-driven predictors are then developed. First, the input weighting strategies are developed for prediction algorithms without dedicated tools tuning input weights. Moreover, a new loss function is designed to increase the eco-driving control performance. Energy benefits from using the developed predictors and strategies are evaluated by applying them to powertrain-agnostic eco-driving controllers minimizing longitudinal acceleration across various powertrain configurations.