Jean-Marie Yazbeck

Monday, March 24, 2025, 11:00 - 12:00

SR 01-012

Abstract

This thesis investigates the application of Model Predictive Control (MPC) to the reverse maneuvering of articulated vehicles, specifically trucks with single and double-trailer configurations. Reverse driving of articulated vehicles introduces considerable control complexities, primarily due to their nonlinear dynamics, high-dimensional state spaces, and nonholonomic constraints, which significantly elevate the risk of instability phenomena such as jackknifing. The intricacy is further compounded when extending the model from a single to a double-trailer configuration.

The primary objective of this thesis is to formulate, design, and validate a comprehensive MPC control tailored for both single and dual-trailer vehicles. Key components of this research encompass trajectory generation, focusing on ensuring kinematic feasibility and smoothness, precise modeling of the nonlinear vehicle-trailer dynamics, and efficient implementation of the derived models into an MPC controller.
This MPC controller is obtained by time-discretization the corresponding Optimal Control Problem (OCP) via the direct multiple shooting method which is implemented in the real-time capable optimization software acados.

Simulations conducted on two distinct benchmark scenarios, a V-shaped track, characterized by a sharp 90-degree turn, and a snake-shaped track featuring multiple alternating curvatures, were executed in both forward and backward driving directions. Results reveal that the MPC controller tracks the specified track even under challenging reverse-driving conditions, effectively avoiding instability risks such as jackknifing. Additionally, detailed numerical simulations underline the practicality of deploying this MPC control in real-time operational contexts.

Overall, the research presented in this thesis represents an advancement in the application of advanced control schemes in the domain of autonomous vehicle maneuvering, providing a foundational step toward future enhancements of multi-trailer vehicle control.