Evaluating the impact of music tempo on drivers and their performance using an artificial intelligence model: a multi-source data approach
DOI: 10.1007/s00521-025-11077-w
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 complex relationship between in-car music listening and driving performance, motivated by the significant global health and economic burden of traffic accidents. While music can enhance driver arousal and mood, it may also increase mental workload and compete for cognitive resources, potentially impairing safety. Previous research has yielded mixed results regarding how music tempo affects driving, and no prior studies have utilized artificial intelligence (AI) models to analyze these effects using combined physiological and performance data. This research aims to fill that gap by investigating how slow-tempo and fast-tempo music influence driver physiology, eye behavior, and performance metrics using a multi-source data approach. To achieve this, the researchers conducted a simulated driving experiment involving 26 participants using a motion platform and the Euro Truck Simulator software. Each participant completed three 4-minute driving scenarios: listening to slow-tempo music (60 bpm), fast-tempo music (140 bpm), and no music. Data was collected via Tobii Pro eye-tracking glasses (gaze, pupil diameter, head motion), Equivital sensor belts (heart rate, breathing rate, galvanic skin response, skin temperature), and a software SDK (vehicle speed, acceleration, steering angle, throttle, brake). The multi-modal dataset was preprocessed through cleansing, normalization, and timestamp alignment. A Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM) model was developed to classify driver states based on auditory conditions. The model was optimized using hyperband hyperparameter tuning and Chi-squared feature selection to maximize accuracy and minimize computation time. The CNN-LSTM model demonstrated superior performance, achieving an average accuracy of 99.29% with minimal variance across precision, recall, and F1 score metrics. This performance significantly outperformed nine other classifiers, including a densely layered deep feed-forward neural network (98.78% accuracy), Gradient Boosting (98.44%), and classical models like Support Vector Machines and Logistic Regression. Feature selection reduced processing time by approximately 20% without substantially compromising accuracy, with model performance plateauing at 16 selected features. The confusion matrices confirmed high true positive rates and few misclassifications across all three music conditions. The findings indicate that AI models can effectively classify driver behaviors under varying auditory conditions using multi-source data. This has implications for enhancing vehicle safety systems, suggesting the potential for future in-vehicle technologies that proactively adjust to driver states to mitigate risks associated with auditory distractions. The authors note limitations regarding the small sample size and controlled experimental setup, which may affect generalizability to real-world driving. They recommend future research explore additional musical characteristics, such as genre and lyrics, and account for individual preferences to further refine understanding of music’s impact on driving performance.
Key finding
A CNN-LSTM model achieved 99.29% accuracy in classifying driver behaviors under different music tempo conditions, outperforming other machine learning models.
Methodology
simulator
Sample size: 26
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 author_sweep on 2026-05-07.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | — | — | — | 1 | 2026-05-07 |
| 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 | 2026-05-07 |
| promote | success | — | — | — | 1 | 2026-05-07 |
| 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
- Methodological Resource: tool software
- Theoretical Contribution: computational model