Model Predictive Multi-Objective Vehicular Adaptive Cruise Control

TLDR

The paper introduces a novel vehicular adaptive cruise control system designed to simultaneously improve tracking performance, fuel economy, and driver‑desired response. The system employs a hierarchical architecture: a lower controller compensates for nonlinear vehicle dynamics, while an upper model‑predictive controller optimizes a quadratic cost that balances tracking error, fuel consumption, comfort, and safety through constraint softening and quadratic programming. Simulations on a heavy‑duty truck demonstrate that the ACC achieves significant fuel‑economy gains and superior tracking while satisfying driver car‑following preferences.

Abstract

This paper presents a novel vehicular adaptive cruise control (ACC) system that can comprehensively address issues of tracking capability, fuel economy and driver desired response. A hierarchical control architecture is utilized in which a lower controller compensates for nonlinear vehicle dynamics and enables tracking of desired acceleration. The upper controller is synthesized under the framework of model predictive control (MPC) theory. A quadratic cost function is developed that considers the contradictions between minimal tracking error, low fuel consumption and accordance with driver dynamic car-following characteristics while driver longitudinal ride comfort, driver permissible tracking range and rear-end safety are formulated as linear constraints. Employing a constraint softening method to avoid computing infeasibility, an optimal control law is numerically calculated using a quadratic programming algorithm. Detailed simulations with a heavy duty truck show that the developed ACC system provides significant benefits in terms of fuel economy and tracking capability while at the same time also satisfying driver desired car following characteristics.