Identifying drivers’ identity by their driving behavioral patterns

16 December 2021, 12:00 
Wolfson Building, Room 206, Tel-Aviv University 
Identifying drivers’ identity by their driving behavioral patterns

Chen Ben-Gal is an M.Sc student in the Department of Industrial Engineering

16  December  2021, 12:00 PM at Room 206 And via zoom



The human factor is a critical component of the Vehicle – Environment – Human triangle. The driver is responsible for operating the vehicle in dynamically changing environments, and adapting it to different road situations. This factor can be analyzed by examining vehicle operation parameters, such as speed, lateral and angular accelerations, safety events, or some other feature combinations. The driving patterns are affected by external conditions, such as weather, traffic loads, time, traffic lights and noise that are hard to model. Understanding the driver identity is beneficial in a variety of use cases related to smart transportation technologies like hybrid auto-insurance, safety application, vehicle thefts and other anomalies.

This study focuses on identifying drivers’ identity by their driving behavioral patterns. Traditionally, the assessment of the driver's identity is based on the client reports, and the most innovative classification models are based on internal sensors such as pedals and wheels angles, some of them takes place in sterilized environment over predefined controlled routes.

This study releases the above restrictions and consider a real life large scale scenario which is less controllable. It is based on a wide range of driving data that includes trips of over 8,000 drivers from a large leasing company. For each drive, GPS data and events log data has been extracted. Some limited number of trips were tagged to tune the learning models using an active learning scheme.

The suggested method extracts various features mainly from GPS coordinates and apply an enrichment and feature engineering process, resulting in 300 new features, and scaling metric based on background trips over the same routs.  We achieved an accuracy of 93.5% classifying the drivers which over perform other conventional methods.


Chen Ben-Gal is an M.Sc student in the Department of Industrial Engineering at Tel Aviv University, under the supervision of Prof. Irad Ben-Gal.

He holds a B.Sc in Statistics and Economics from Tel Aviv University.

He works as a Core Researcher in Forter



Tel Aviv University makes every effort to respect copyright. If you own copyright to the content contained
here and / or the use of such content is in your opinion infringing, Contact us as soon as possible >>