1 July 2021 in Accident Analysis & Prevention
8 February 2023 in IEEE Transactions on Human-Machine Systems
Maintaining situation awareness (SA) is essential for drivers to deal with the situations that Society of Automotive Engineers (SAE) Level 3 automated vehicle systems are not designed to handle. Although advanced physiological sensors can enable continuous SA assessments, previous single-modality approaches may not be sufficient to capture SA. To address this limitation, the current study demonstrates a multimodal sensing approach for objective SA monitoring. Physiological sensor data from electroencephalogram and eye-tracking were recorded for 30 participants as they performed three secondary tasks during automated driving scenarios that consisted of a pre-takeover (pre-TOR) request segment and a post-TOR segment. The tasks varied in terms of how visual attention was allocated in the pre-TOR segment. In the post-TOR segment, drivers were expected to gather information from the driving environment in preparation for a vehicle-to-driver transition. Participants' ground-truth SA level was measured using the Situation Awareness Global Assessment Techniques (SAGAT) after the post-TOR segment. A total of 23 physiological features were extracted from the post-TOR segment to train computational intelligence models. Results compared the performance of five different classifiers, the ground-truth labeling strategies, and the features included in the model. Overall, the proposed neural network model outperformed other machine learning models and achieved the best classification accuracy (90.6%). A model with 11 features was optimal. In addition, the multi-physiological sensor-model outperformed the single sensing model by comparing prediction performance. Our results suggest that multimodal sensing model can objectively predict SA. The results of this study provide new insight into how physiological features contribute to the SA assessment.
19 November 2024 in Applied Ergonomics
15 October 2025 in Ergonomics
Drivers' situation awareness (SA) is often assessed using indirect behavioural indicators such as takeover performance, despite known limitations. Many studies have also focused heavily on visual attention, primarily capturing the perception level of SA. To address this, we developed two physiological SA prediction models. Model 1 links physiological responses to full-level SA (perception, comprehension, and projection), while Model 2 targets comprehension-level SA to examine mechanisms beyond perception. A simulated driving study was conducted with 39 participants, during which eye-tracking and EEG data were collected. In Model 1, both modalities distinguished the degree of full-level SA. In Model 2, eye-tracking metrics differentiated successful from unsuccessful comprehension. In both models, the eye-tracking-based model outperformed the EEG-based model. However, combined model showed better performance, particularly in predicting comprehension-level SA. These findings demonstrate the benefit of integrating cognitive state monitoring beyond perception alone, especially for assessing higher-level SA in automated driving.
30 August 2026 in Proceedings of the Human Factors and Ergonomics Society Annual Meeting
KAKhandakar Ashrafi Akbar
Driver gaze is commonly modeled as a spatial heatmap, but heatmaps alone are difficult for humans to interpret because they do not explain which road object or region is being monitored or why an attention shift may matter. This study examines whether minimal human-grounded supervision can steer a vision–anguage model toward interpretable descriptions of driver attention shifts. Using selected high-change gaze moments from the Berkeley DeepDrive-Attention dataset, we compare zero-shot, one-shot, and LoRA fine-tuned VLM conditions against human-refined reference descriptions and expert ratings. Results show that fine-tuning with 80 expert-refined attention examples improves ROUGE-L, METEOR, Entity Alignment F1, and Human Alignment Score relative to unsteered VLM outputs. The findings suggest that language-based descriptions can complement gaze heatmaps by making driver attention more accessible for human-factors analysis, driver-monitoring review, and situation-awareness support.
29 September 2026 in Journal of Biomechanical Engineering
This study evaluated the biomechanical effectiveness of antirotation load-leg support in rear-facing child restraint systems (CRSs) under proper and improper (misuse) installation conditions. We conducted eight FMVSS 213 sled tests using a CRABI 12-month anthropomorphic test device under four installation configurations-proper installation, loose seat, loose harness, and both loose seat and loose harness-with and without the load-leg support. The use of the load leg reduced head and neck injury metrics in all installation conditions: mean HIC36 decreased from 682 to 571 (by 16%) and AIS ≥ 4 head injury probability from 33% to 21% (p = 0.049); mean Nij decreased from 1.39 to 1.28 (by 8%), and AIS ≥ 4 neck injury probability from 26% to 24% (p = 0.069); and forward CRS rotation by 6 deg–19 deg, with larger reductions under proper installation and loose seat conditions. In contrast, the load-leg use increased chest 3-ms acceleration by 5% (from 62 g to 65 g), and AIS ≥ 4 thoracic injury probability from 58% to 63% (p = 0.049), suggesting that because a 3-ms metric captures only short-duration loading, these findings should be cautiously interpreted. Force-plate measurements showed that the load leg transferred 90% (242 N·s) of the estimated CRS-ATD forward momentum under proper installation and 93% (249 N·s) under loose seat conditions, compared to 70% (189 N·s) under loose-harness conditions. These findings show that load legs provide effective head and neck protection, especially when the harness maintains strong child-CRS coupling.