Data-Driven Steering Dynamics Modeling and Steering Angle Tracking Control for Self-Driving Vehicles: Simulation and Experiments on 2025 Nissan Leaf Electric Vehicle

Engineering notes

Simulation and on-vehicle tests demonstrate that the proposed data-driven model and controllers can be deployed on the vehicle: the PID controller achieves a steady-state tracking root-mean-square error of 1.23° across the full ±450° operating range, and the neural controller tracks in closed loop with a characterized direction asymmetry identified for future refinement.

Chinese explanation / 中文解读

中文解读待补充：本站会优先为端到端自动驾驶、BEV感知、3D目标检测、轨迹预测、路径规划、LiDAR感知等高价值论文补充中文说明。

Original abstract

Autonomous vehicles (AVs) increasingly rely on accurate steering dynamics models and high-precision steering angle tracking to achieve safe and reliable control. Additionally, the increased attention from automakers and academia emphasizes the potential of AVs in improving transportation safety, convenience, energy efficiency, and ride comfort. This paper presents a data-driven modeling framework using both simulated and real-world driving data collected from a 2025 Nissan Leaf SV Plus. This test vehicle is equipped with an in-house developed drive-by-wire system. Besides the data-driven steering dynamics model, this paper also presents a data-driven steering angle tracking control and a proportional-integral-derivative control. Simulation and on-vehicle tests demonstrate that the proposed data-driven model and controllers can be deployed on the vehicle: the PID controller achieves a steady-state tracking root-mean-square error of 1.23° across the full ±450° operating range, and the neural controller tracks in closed loop with a characterized direction asymmetry identified for future refinement. The proposed data-driven steering dynamics model and control can be used in the development of next-generation autonomous driving systems for their accuracy and simplicity.

Links and sources

Comments