JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Classifying Advanced Driver Assistance System (ADAS) Activation from Multimodal Driving Data: A Real-World Study

  • 0Department of Computer Engineering, Hallym University, Chuncheon 24252, Republic of Korea.
Sensors (basel, Switzerland) +

|

October 16, 2025

|

October 16, 2025

View abstract on PubMed

Summary

This summary is machine-generated.

This study demonstrates detecting advanced driver assistance systems (ADAS) status in real-world driving using synchronized vehicle and smartphone data. Findings show ADAS use impacts driving behavior, offering insights for safety and future research.

Area Of Science

  • Automotive Engineering
  • Human-Computer Interaction
  • Transportation Safety

Background

  • Accurate detection of advanced driver assistance systems (ADAS) status is vital for safety, liability, and accident reconstruction.
  • Existing research often uses simulations or unsynchronized data, limiting real-world applicability.

Purpose Of The Study

  • To develop and evaluate methods for distinguishing between ADAS-enabled and manual driving modes using real-world data.
  • To analyze behavioral differences associated with ADAS usage.

Main Methods

  • Collected synchronized Controller Area Network (CAN)-bus and smartphone Inertial Measurement Unit (IMU) data from drivers on highways.
  • Developed lightweight statistical and deep learning classification models to detect ADAS status.
  • Analyzed driving behavior, focusing on speed regulation and steering stability.

Main Results

  • Systematic differences in speed control and steering variability were observed between manual and ADAS modes.
  • ADAS usage was associated with reduced steering variability and more stable speed control.
  • The developed classification pipelines demonstrated moderate accuracy in identifying ADAS operational status.

Conclusions

  • This study presents one of the first data-driven detections of ADAS status under naturalistic driving conditions.
  • The findings confirm ADAS influences driver behavior, providing valuable data for driver monitoring and adaptive systems.
  • The released dataset serves as a resource for advancing automotive safety research.

Keywords: