Armira Kontaxi defended her PhD thesis on the Driver Behavior Telematics Feedback Mechanism, February 2025

Armira Kontaxi defended her PhD thesis titled “The Driver Behavior Telematics Feedback Mechanism” on Friday February 21st 2025 at the Railways Amphitheatre of the Department of Transportation Planning and Engineering of the School of Civil Engineering of NTUA, and online via a teleconference at the following link:

https://centralntua.webex.com/meet/geyannis

This PhD thesis was carried out at the Department of Transportation Planning and Engineering at the School of Civil Engineering of the National Technical University of Athens under the supervision of  Professor George Yannis.

This dissertation examines the driver behavior telematics feedback mechanism, addressing gaps in feedback’s lifecycle effects—pre-feedback, feedback, and post-feedback phases. To that end, A 21-month naturalistic driving experiment involving 230 drivers across six feedback phases generated a robust dataset of 106,776 trips, covering 1.3 million kilometers. The tailored feedback interventions concerned scorecards, gamification, and peer comparisons. Advanced statistical and machine learning models, including Generalized Linear Mixed-Effects Models (GLMMs), Structural Equation Models (SEMs), and Survival Analysis methods (e.g., Weibull AFT, Cox-PH with frailty, and Random Survival Forests), were utilized to analyze behavioral metrics such as speeding, mobile phone use, harsh braking, and accelerations which demonstrated substantial impacts on reducing risky behaviors. Results demonstrated that the overall impact of feedback significantly improved driving behavior and safety, with notable variations across user groups and driving contexts. Urban environments demonstrated the most substantial reductions in mobile phone use and harsh events, likely driven by the heightened complexity and demands of navigating urban settings. Feedback features also influenced outcomes differently; scorecards were particularly effective in reducing risky behaviors like speeding, while gamification elements motivated sustained engagement among professional drivers. Despite these successes, survival analyses revealed significant relapse tendencies once feedback was removed, with survival probabilities for maintaining improved behaviors, such as reduced speeding and harsh braking, falling below 50% within 150 trips post-feedback. These findings highlight the need for continuous and adaptive engagement strategies, incorporating diverse features tailored to the specific needs of different user groups and driving contexts, to ensure long-term effectiveness and sustained safety improvements.

Invitation 

Abstract