Driving-PASS: A Driving Performance Assessment System for Stroke Drivers Using Deep Features
DOI: 10.1109/ACCESS.2021.3055870
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 stroke survivors returning to driving without valid assessment, often due to the limitations and risks associated with traditional on-road tests. The authors propose Driving-PASS, a Driving Performance Assessment System designed to pre-screen stroke drivers using a driving simulator. The system aims to identify invalid drivers before they undergo on-road testing and to pinpoint specific problematic driving behaviors for targeted rehabilitation. The study utilized a STISIM Drive driving simulator to collect data from 27 participants, comprising 18 stroke survivors and 9 healthy controls. Participants completed 13 driving scenarios across urban, highway, and rural environments, generating time-series driving profiles. To establish ground truth labels, ten driving evaluators (seven experts and three researchers) assessed the participants on 11 items: 10 specific driving ability items scored on a Likert scale and one binary driving suitability item. The authors employed a reliability adjustment process using Cronbach’s alpha to select optimal raters for each item. To build the assessment classifiers, the researchers converted segmented driving data into 3D matrix images and extracted deep features using five pre-trained Convolutional Neural Network (CNN) models: ResNet18, InceptionV3, ResNet50, ResNet101, and Inception-ResNet-v2. This approach addressed the limitations of handcrafted feature design and handled imbalanced datasets through resampling methods. The driving ability items were classified into "Proficient" or "Non-proficient" based on a threshold derived from the evaluators' mean scores, while driving suitability was determined by a majority vote of the expert evaluators. The classifiers were trained and evaluated using Automated Machine Learning (AutoML) with stratified holdout validation. The results demonstrate that Driving-PASS successfully constructed 11 accurate assessment classifiers by carefully selecting deep features for each assessment item. The system effectively automates the subjective assessment process, providing a reliable, safe, and cost-effective alternative to manual evaluations. The authors conclude that Driving-PASS serves as a robust pre-screening tool that can improve road safety by identifying stroke drivers who require further rehabilitation or license restriction, thereby addressing the gap in valid screening tools for this population.
Key finding
The proposed Driving-PASS system successfully constructs accurate assessment classifiers using deep features from pre-trained CNN models to evaluate driving abilities and suitability for stroke survivors.
Methodology
simulator
Sample size: 27
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. Discovered via topic_sweep_doaj on 2026-06-01.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | — | — | — | 1 | 2026-06-01 |
| 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 | 2026-06-01 |
| promote | success | — | — | — | 1 | 2026-06-04 |
| 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).
- Methodological Resource: tool software, validation psychometrics
- Theoretical Contribution: computational model