Multimodal Wearable Based Automated Driver Inattention State Assessment Using Multi Devices and Novel Cross-Modal Attention Framework

Driver inattention detection remains a critical challenge in driver's well being, requiring robust systems that can distinguish stress-induced mental load during naturalistic driving. Current approaches face limitations in multiple wearable based data fusion and real-time biosignals assessment....

Full description

Saved in:
Bibliographic Details
Published inIEEE sensors letters pp. 1 - 4
Main Authors Pavan, Kaveti, Singh, Ankit, Pawar, Digvijay S., Ganapathy, Nagarajan
Format Journal Article
LanguageEnglish
Published IEEE 2025
Subjects
Accuracy
Biomedical monitoring
Cross squeeze-and-Excitation
driver inattention detection
Electrocardiography
Feature extraction
Frequency modulation
inattention monitoring
Monitoring
multimodal physiological signals
Vehicles
Wearable devices
wearable sensors
Wireless communication
Wireless sensor networks
Online AccessGet full text

Cover

More Information
Summary:Driver inattention detection remains a critical challenge in driver's well being, requiring robust systems that can distinguish stress-induced mental load during naturalistic driving. Current approaches face limitations in multiple wearable based data fusion and real-time biosignals assessment. This study proposes a novel cross-squeeze-and-excitation convolution neural network (crSE-CNN) framework to process simultaneously acquired multiple wearable from 15 participants in controlled driving scenarios. The multimodal signals are applied to multi-stage attention mechanisms (ECG <inline-formula><tex-math notation="LaTeX">\leftrightarrow</tex-math></inline-formula> EDA, ECG <inline-formula><tex-math notation="LaTeX">\rightarrow</tex-math></inline-formula> EDA, EDA <inline-formula><tex-math notation="LaTeX">\rightarrow</tex-math></inline-formula> ECG) with 1D-CNN blocks, optimized for 10-second signal segments. The proposed approach is able to classify drive inattention state. It is observed that ECG <inline-formula><tex-math notation="LaTeX">\rightarrow</tex-math></inline-formula> EDA attention achieves 76.54% average accuracy using Leave-One-Subject-Out Cross-Validation, outperforming unimodal approaches by 12.4% and bidirectional attention by 4.8%. Feature visualizations confirm enhanced pattern discrimination in inattention conditions. This work advances driver health monitoring systems through effective wearable integration and adaptive feature weighting, with potential for edge deployment and clinical stress assessment applications
ISSN:2475-1472
2475-1472
DOI:10.1109/LSENS.2025.3596610