Tube-based Robust Model Predictive Control for Tracking Control of Autonomous Articulated Vehicles

Abstract

Articulated vehicles play a critical role in the transportation industry, but the rise in truck-related accidents necessitates effective solutions. Autonomous driving presents a promising approach to enhancing safety. Among autonomous technologies, this paper presents a framework for an autonomous vehicle tracking control algorithm utilizing tube-based robust model predictive control (RMPC). The primary objective is to achieve precise path tracking while ensuring performance, safety, and robustness even with modeling errors. The framework adopts a lumped dynamics model for articulated vehicles, which reduces computational complexity while preserving linearity. Specific constraints of articulated vehicles are integrated to guarantee stability, safety, and adherence to actuator limits. The tube-based RMPC technique reliably satisfies constraints under worst-case scenarios, thereby addressing robustness against modeling errors. The proposed algorithm employs tube-based RMPC to ensure the safety and robustness of autonomous articulated vehicles. In the design of the tracking controller, error tube analysis between the actual plant and the prediction model plays a vital role. An error tube analysis method and framework are introduced through simulation. Performance evaluations of the proposed algorithm and previous tracking controllers are conducted through comparative simulations. Previous algorithms exhibited tracking errors exceeding 50 cm, posing potential safety risks. In contrast, the proposed algorithm demonstrates tracking errors of less than 50 cm. Furthermore, the proposed algorithm exhibits notable stability. The results demonstrate that the proposed algorithm enables accurate and safe tracking of complex autonomous articulated vehicles.