Data-Driven Adaptive Cruise Control for Trucks With Lane-Keeping Model

This paper proposes a data-driven adaptive cruise control (ACC) strategy for trucks, incorporating the lane-keeping assistance. It addresses both the maintainace of a safe longitudinal distance and lateral tracking of the road centerline. To handle system constraints, model predictive control (MPC) is employed. The nonlinear characteristics of trucks are approximated using Koopman operator theory and deep neural network (DNN) techniques, yielding a global linear predictive model. Furthermore, a comparative analysis between first-principle modeling and data-driven modeling methods is conducted using TruckSim and Simulink. The results demonstrate the proposed ACC strategy's effectiveness in reducing computational time while ensuring longitudinal safety and lateral lane-keeping capabilities.