A Data-driven Optimization-based Control Model for Cooperative Adaptive Cruise Control Under Uncertainty
Motivated by connected and automated vehicle (CAV) technologies, this talk presented a data-driven optimization-based Model Predictive Control (MPC) model for the cooperative adaptive cruise control (CACC) of a string of CAVs under traffic uncertainty. The proposed data-driven MPC modeling framework aimed to improve the stability, robustness, and safety of longitudinal cooperative automated driving under uncertainty using Vehicle-to-Vehicle (V2V) data. Based on an online learning-based driving dynamics prediction model, we predicted a set of uncertain driving states of the vehicles preceding the controlled CAVs. With the predicted driving states of the preceding vehicles, we solved a constrained Finite-Horizon Optimal Control problem to predict a set of driving states of the controlled CAVs. To obtain the optimal acceleration or deceleration commands for the CAVs under uncertainty, we formulated a Distributionally Robust Stochastic Optimization (DRSO) model (i.e., a special case of data-driven optimization models under moment bounds) with Distributionally Robust Chance Constraints (DRCC). To solve the DRSO-DRCC model, we reformulated its relaxed dual problem as a Semidefinite Program (SDP) based on the strong duality theory and the Semidefinite Relaxation (SDR) technique. We used the Next Generation Simulation (NGSIM) data to demonstrate the proposed model by numerical experiments. The experimental results and analyses show that the proposed model can obtain a string stable, robust, and safe longitudinal cooperative automated driving control of CAVs within proper settings.
Kuilin Zhang is an Associate Professor in the Department of Civil and Environmental Engineering and an Affiliated Associate Professor in the Department of Computer Science at Michigan Tech, Houghton, MI. Dr. Zhang received his Ph.D. degree in Transportation Systems Analysis and Planning from the Department of Civil and Environmental Engineering at Northwestern University in December 2009. After working as a Postdoctoral Fellow in the Transportation Center at Northwestern, he joined the Energy Systems Division at Argonne National Laboratory as a Postdoctoral Appointee in November 2010. He joined Michigan Tech in August 2013 through the university-wide Strategic Faculty Hiring Initiatives in Multimodal Transportation Systems. Dr. Zhang is a member of the Editorial Advisory Board of Transportation Research Part E, as well as Transportation Research Board (TRB) standing committees of Transportation Network Modeling (AEP40) and Freight Transportation Planning and Logistics (AT015). Dr. Zhang’s research interests focus on applying optimization, control, simulation, machine learning, and on-road vehicle testing to address safety, congestion, environmental, energy, and resilience issues of civil infrastructure systems in smart cities such as transportation and logistics systems, connected and automated vehicles, cyber-physical systems, vehicular ad-hoc networks (VANETs), on-road vehicle testing of CAV applications, interdependent large-scale networked infrastructure systems, and open-source transportation simulation platform development. Dr. Zhang’s research has been supported by NSF, DOE, DOT, and MDOT. Dr. Zhang was a recipient of the NSF CAREER award in 2019.