Spatial Deep Learning Approach to Older Driver Classification
Author Information
Author(s): Boateng Charles, Ghoreishi Seyedeh Gol Ara, Yang Kwangsoo, Jan Muhammad Tanveer, Tappen Ruth, Jang Jinwoo, Newman David, Moshfeghi Sonia, Jackson Kelly, Resnick Rhian, Furht Borko, Rosselli Monica, Conniff Joshua
Primary Institution: Florida Atlantic University
Hypothesis
Can a spatial deep-learning approach improve the classification of older drivers into normal and abnormal categories using telematics data?
Conclusion
The proposed spatial deep-learning approach significantly enhances the detection of abnormal driving behaviors in older drivers.
Supporting Evidence
- The study demonstrated significant improvements in detecting abnormal driving behaviors compared to traditional methods.
- Data from 200 vehicles over three years was used to validate the proposed approach.
- The combined model outperformed both naive and grid-based approaches in classification accuracy.
Takeaway
This study uses smart technology to help figure out if older drivers are driving safely or not, which can keep them and others safe on the road.
Methodology
The study collected telematics data from 200 drivers over 3.5 years and used a combined model of Simple Neural Networks and Convolutional Neural Networks to classify driving behaviors.
Limitations
The computational demands of large datasets may pose challenges for real-time deployment.
Participant Demographics
Participants were aged 65 and older, including individuals with Mild Cognitive Impairment (MCI).
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website