Datenpaket: Multi-Surface Driving Maneuvers: A Dataset from Standstill to Vehicle Dynamics Limits

This dataset presents vehicle dynamics and steering system measurements in various driving conditions. The test vehicle is a rear-wheel-drive Hyundai IONIQ 5 equipped with Michelin CrossClimate 2 tires at nominal inflation pressures. Vehicle dynamics signals were captured using a Kistler Correvit S-Motion optical sensor and an ANavS Multi-Sensor Real-Time Kinematic Global Navigation Satellite System, while left and right tie rod forces were measured with dedicated strain gauges. Furthermore, standard internal vehicle signals were accessed via the vehicle’s CAN bus. These signals were synchronously recorded on a dSPACE MicroAutoBox III at a sampling rate of 1000 Hz.
Testing was conducted at the Karlsruhe Institute of Technology Campus East under dry environmental conditions. The driving maneuvers were performed manually and capture a broad spectrum of dynamic states: standstill, parking procedures, slow-driving transitions, and maneuvers up to 50 km/h at the lateral vehicle dynamics limits. The dataset encompasses a wide range of surface textures, including alignment turn plates, concrete, cobblestone, and two distinct asphalt variants. The resulting dataset contains 41 distinct test runs, providing 42 minutes of recorded signals and a static vehicle parameter set, making it highly suitable for the evaluation of dynamic single-track models alongside steering and tire models.

This dataset provides vehicle dynamics and steering system data collected from a rear-wheel-drive Hyundai IONIQ 5. It comprises 41 distinct test runs totaling 42 minutes of recorded signals of various driving maneuvers and road surface conditions in dry weather. Data was synchronously recorded on a dSPACE MicroAutoBox III at 1000 Hz using a Kistler Correvit S-Motion sensor, an ANavS Multi-Sensor Real-Time Kinematic Global Navigation Satellite System, dedicated tie rod strain gauges, and OEM vehicle CAN bus access.
The dataset is designed for use with MATLAB. The dynamic measurement runs are provided as individual MAT-files containing the time series data. The specifics of the driving maneuver and the road surface are encoded directly into the individual file names. The dataset includes a MATLAB script containing vehicle, steering, and tire parameters. Additionally, the necessary translation vectors between the various sensor reference frames are included.
The dataset includes a comprehensive set of vehicle dynamics, steering system, and global positioning signals. All multi-axis dynamic signals include suffixes indicating their directional axis and the specific reference frame in which the data was recorded. The physical unit is explicitly included at the end of every signal name.
v: Linear velocity
beta: Slip angle
a: Linear acceleration
w: Angular velocity
w_fl: Wheel speed of the front-left wheel
delta_stm: Effective steering angle for a single-track model, derived from the steering wheel angle
F_tr: Tie rod forces, provided as separate signals for the left and right sides
M_eps: Torque of the Electric Power Steering motor
lat: Geodetic latitude
long: Geodetic longitude