Thursday, April 3, 2025 | 12:00 – 1:00 PM (ET)
The widespread use of smartphones and wearables in transportation research has enabled cost-effective, non-intrusive methods to improve safety and efficiency. Smartphone sensors, including accelerometers, gyroscopes, and GPS, have been used to study behaviors such as texting while driving and differentiating between drivers and passengers. Understanding these differences enhances travel pattern analysis and traffic studies. However, the challenge lies in balancing classification accuracy with device battery life, ensuring that mobile applications remain practical for long-term use. This study aimed to analyze driver-versus-passenger classification using smartphone sensor data collected in naturalistic settings while optimizing data collection for resource efficiency. Key tasks included evaluating machine learning classifiers, assessing small time windows for classification, and exploring feature selection and dimensionality reduction techniques.
Findings showed that decision trees and random forests achieved high classification accuracy (0.892 to 0.97), with shorter time windows (as brief as three seconds) proving effective. Feature selection outperformed dimensionality reduction, particularly for tree-based classifiers, while feature standardization significantly improved accuracy for larger datasets. Policy recommendations include optimizing data collection periods for efficiency, prioritizing feature standardization for longer time frames, and favoring feature selection over dimensionality reduction when using machine learning classifiers. These insights can guide the development of sensor-based classification systems in transportation research, balancing performance with practical constraints.
About the Presenters:
Dr. Tempestt Neal is an Associate Professor of Computer Science and Engineering at the University of South Florida. She leads the Cyber Identity and Behavior Research (CIBeR) Lab, which primarily conducts quantitative and qualitative research on mobile-based sensing for biometrics and human behavior understanding in interdisciplinary applications. Her research also explores cybersecurity awareness to promote widespread adoption of good cybersecurity practices.
She holds a Ph.D. from the University of Florida (Computer Engineering, 2018), M.S. from Clemson University (Computer Science, 2014), and a B.S. from South Carolina State University (Computer Science with a minor in Mathematics, 2012).
Dr. Neal is currently serving as a Program Co-Chair for the IEEE International Conference on Automated Face and Gesture Recognition. She has served as an Associate Editor for the IEEE Biometrics Council Newsletter and Guest Editor for the MDPI Electronics Special Issue on Recent Advances in Biometric Security in IoT Based on Machine Learning. She has also served on the organizing committee for several workshops in Artificial Intelligence and Biometrics, including the Workshop on Applied Multimodal Affect Recognition and the Workshop on Interdisciplinary Applications of Identity Science and Biometrics.
She was a recipient of the University of Florida Delores Auzenne Dissertation Award and National Science Foundation CyberCorps Scholarship for Service Fellowship. She was also recognized as a 2021-2022 McKnight Junior Faculty Fellow and received an NSF CAREER Award in 2023.
Wilson Lozano is a Ph.D. student in Computer Science and Engineering at the University of South Florida, with a research focus on behavioral biometrics, artificial intelligence, cybersecurity, and data science. He works as a Research Assistant at the USF-CiBER Lab, investigating biometric data collection methods and AI-driven behavioral biometrics applied to authentication methods.
Previously, Wilson served as a Graduate Research Assistant at the Center for Urban Transportation Research (CUTR), where he developed machine learning models for travel behavior analysis and contributed to data integrity evaluations in mobility research.
Wilson has a strong commitment to education and mentorship. He is a former Computer Engineering faculty member at the Inter American University of Puerto Rico, where he founded the Computing Research and Engineering Lab (CoRE Lab)—a space dedicated to interdisciplinary research for undergraduate students.
Wilson is a recipient of the NSF CREST Scholarship for Service in Cybersecurity, a McKnight Doctoral Fellow, and a Sloan Scholar. In addition to his academic pursuits, he enjoys puzzles, dancing Latin music, and outdoor activities.
Wilson holds a bachelor’s in computer systems engineering from Universidad Industrial de Santander (Colombia) and a M.S.Cp.E. from the University of Puerto Rico, Mayagüez.
