Application of Data-Driven Model Predictive Control for Autonomous Vehicle Steering

TL;DR

The paper addresses steering control for autonomous vehicles where traditional MPC relies on accurate, often nonlinear models. It applies Data-driven MPC based on Willems' Lemma and Hankel data to predict future behavior without explicit models. The authors adapt the DDMPC framework to autonomous vehicle steering, derive the optimization problem, and validate it via CarSim-Simulink with comparisons to PID and kinematic MPC, demonstrating improved computation time and tracking performance. The results indicate that data-driven steering control is a feasible and efficient alternative for real-time AV operation, with potential for broader robustness enhancements.

Abstract

With the development of autonomous driving technology, there are increasing demands for vehicle control, and MPC has become a widely researched topic in both industry and academia. Existing MPC control methods based on vehicle kinematics or dynamics have challenges such as difficult modeling, numerous parameters, strong nonlinearity, and high computational cost. To address these issues, this paper adapts an existing Data-driven MPC control method and applies it to autonomous vehicle steering control. This method avoids the need for complex vehicle system modeling and achieves trajectory tracking with relatively low computational time and small errors. We validate the control effectiveness of the algorithm in specific scenario through CarSim-Simulink simulation and perform comparative analysis with PID and vehicle kinematics MPC, confirming the feasibility and superiority of it for vehicle steering control.

Figures (6)

• Figure 1: Algorithm application process of the DDMPC for AV steering control.
• Figure 2: Schematic diagram of a dual-lane switching scenario.
• Figure 3: Trajectories of Data-driven MPC, PID and kinematics MPC with reference line.
• Figure 4: Front wheel steering.
• Figure 5: Tracking error of Data-driven MPC, PID and kinematics MPC.