Application of Model Predictive Controller to Adaptive Cruise Control

Tianshu Wang
DOI: https://doi.org/10.1109/cvci56766.2022.9964914
2022-01-01
Abstract:In this paper, a model predictive control (MPC) controller is designed to address the adaptive cruise control (ACC) problem, in which fuel economy and tracking capability need optimizing at the same time. A quadratic cost function is developed that considers the trade-off between minimal tracking error and minimal fuel consumption while rear-end safety is formulated as linear constraints. The quadprog solver in MATLAB is used to solve this quadratic programming problem. Employing Bayesian Optimization, a proper number of prediction steps, sampling time and weighting matrices are selected. The proposed MPC controller is evaluated in a complex scenario with varying road slopes and frequently changed following speed. Under such simulation environment, the MPC controller can reduce tracking error index (TEI) to 0.18 while maintaining a relatively low fuel consumption per hundred kilometers (FE), 9.4.
What problem does this paper attempt to address?