Driving-Style Assessment from a Motion Sickness Perspective Based on Machine Learning Techniques

Ruiz Colmenares, Jon Ander; Asua Uriarte, Estibaliz; del Campo, Inés · 2023 · Crossref

DOI: 10.3390/app13031510

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study addresses the challenge of improving ride comfort by identifying driving behaviors that induce passenger motion sickness. As automated driving systems advance, passenger comfort becomes a critical design parameter, yet motion sickness remains difficult to evaluate due to its subjective nature and dependence on various external factors. The authors propose a methodology that utilizes measured vehicle signals and machine learning techniques to identify specific driver actions negatively affecting comfort, aiming to develop a recommendation system for smoother driving. The research employs the Uyanik instrumented car dataset, consisting of signals from 18 drivers traveling a fixed 25 km route in Istanbul. The data includes CAN-bus variables (e.g., speed, pedal pressure, steering angle) and IMU accelerations. To quantify motion sickness, the authors calculate the Motion Sickness Dose Value (MSDVxy), a metric derived from ISO 2631 standards that combines lateral and longitudinal acceleration effects. The data processing pipeline involves cleaning outliers using the Local Outlier Factor algorithm, normalizing signals, and reducing dimensionality. Feature selection was performed using correlation analysis and Recursive Feature Elimination (RFE), identifying five key variables: longitudinal acceleration (positive and negative), lateral acceleration, steering wheel angle relative speed, and brake pedal pressure. To address class imbalance, the dataset was categorized by MSDVxy levels and augmented using oversampling techniques like SMOTE. Clustering algorithms, specifically K-Means, Spectral Clustering, and Agglomerative Clustering, were applied to group driving patterns. Model performance was evaluated using the silhouette coefficient to ensure distinct cluster separation. The results indicate that the clustering models successfully identified interpretable driving patterns associated with varying levels of motion sickness. The analysis revealed that specific combinations of driver-controlled variables, particularly aggressive acceleration and steering maneuvers, correlated with higher MSDVxy values. The study validated the methodology by demonstrating that the identified clusters correspond to distinct driving styles, allowing for the differentiation between comfortable and uncomfortable driving behaviors. The significance of this work lies in its provision of a data-driven framework for assessing driving style from a comfort perspective. By linking specific vehicle dynamics to motion sickness metrics, the authors enable the creation of a recommendation system that can advise drivers on how to adjust their behavior to reduce passenger discomfort. This approach supports the development of Advanced Driver Assistance Systems (ADAS) and automated driving features that prioritize ride comfort, offering a practical tool for optimizing driving patterns in both human-driven and autonomous scenarios.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• passenger motion sickness comfort
• simulator sickness

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Methodological Resource: tool software
• Theoretical Contribution: computational model