What's on your mind? A Mental and Perceptual Load Estimation Framework towards Adaptive In-vehicle Interaction while Driving
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This paper addresses the critical safety issue of driver inattention by proposing a machine learning framework to estimate Mental Workload (MWL) and Perceptual Load (PL) separately during driving. While MWL has been extensively studied, PL—defined as the extent to which a task consumes available perceptual capacity—remains underexplored in applied driving contexts. The authors aim to develop a real-time, adaptive in-vehicle interface system that adjusts to the driver’s cognitive state using non-intrusive, off-the-shelf sensors (ECG and eye-tracking) that are compatible with modern vehicle systems. The study employed a within-subject dual-task design in a driving simulator with 45 participants. Participants performed a primary lane-changing task while simultaneously engaging in one of two secondary tasks designed to manipulate specific load types. An auditory n-back task (1-back, 2-back, 3-back) was used to induce varying levels of MWL without visual interference, while a visual search task with varying set sizes (2x2, 3x3, 4x4 matrices) was used to manipulate PL. Psychophysiological data, including heart rate, heart rate variability (RMSSD), and pupil diameter (processed via the LHIPA algorithm), were collected alongside driving performance metrics (lateral deviation) and secondary task performance. The authors utilized a nested cross-validation approach to train classification models that distinguish between low, medium, and high load levels for both MWL and PL. The results indicate that mental workload significantly influences psychophysiological dimensions, whereas perceptual load shows little effect on these specific measurements. Despite this, the machine learning framework successfully classified load levels using the fused sensor data. The model achieved up to 89% accuracy in classifying mental workload levels. The statistical analysis supported the hypothesis that psychophysiological measures can distinguish between different n-back task levels, reflecting changes in MWL. However, the distinction between PL levels was less pronounced in the physiological data, though the visual search task successfully induced the intended perceptual demands. The significance of this work lies in providing an open-source, minimally intrusive framework for real-time estimation of driver cognitive states. By differentiating between MWL and PL, the study offers a more nuanced approach to adaptive human-machine interfaces than previous methods that treated cognitive load as a single construct. The high classification accuracy for MWL demonstrates the viability of using standard vehicle sensors to monitor driver state, potentially enabling systems that reduce interface complexity or trigger warnings when mental resources are taxed, thereby enhancing driving safety and user experience.
Key finding
Mental workload significantly influences psychophysiological dimensions and can be classified with up to 89% accuracy using machine learning, whereas perceptual load shows little effect on these measurements.
Methodology
simulator
Sample size: 45
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.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | — | — | — | 1 | 2026-05-28 |
| archive | success | canonical_url | — | — | 1 | 2026-06-06 |
| extract | success | cached | — | — | 3 | 2026-06-10 |
| clean | success | clean | — | — | 1 | 2026-06-04 |
| chunk | success | chunk | — | — | 1 | 2026-06-04 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-04 |
| enrich | success | — | — | — | 1 | 2026-05-28 |
| promote | success | — | — | — | 1 | 2026-05-28 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 2 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 15 | 2026-06-11 |
| verify | success | — | — | — | 2 | 2026-06-10 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
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).
- Empirical Findings: physiological data
- Theoretical Contribution: theory or model, computational model