New driver behaviour models for fleet management based on big data analytics


Terzi R., Sağıroğlu Ş., Cocu O., Arkan R., Tosun M., Tulgar Y.

JOURNAL OF THE FACULTY OF ENGINEERING AND ARCHITECTURE OF GAZI UNIVERSITY, cilt.36, sa.1, ss.543-557, 2021 (SCI-Expanded) identifier identifier identifier

Özet

In this study, shortcomings of fleet management systems have been examined from big data and driver behavior perspectives and big data based new models have been proposed to perform analysis with 6 different scenarios. According to the analysis results, it was determined that among the drivers exceeding the speed limit above 50%, certain drivers have committed 30% of these violations compared to other drivers, there may be a 6-fold difference between drivers in terms of speed violations, even if the average speed is the same, similarly, even if the number of speed violations are the same, there may be a 2-fold difference in violation times, according to the seasonal the highest number of speed violations is in summer, however, speed violation duration occurred in autumn at most, roads where speed limits are exceeded in Ankara in terms of districts, quarter , intercity roads and inner city roads are 23.6% Yenimahalle, 4.62% Saray, 6.85% Eskisehir road and 2.74% Anadolu Boulevard, respectively. Finally, according to 3 specific radar analysis, there may be differences of up to 300% in speed violations before and after radar. As a result, it is evaluated that by using big data analytics, fleets can be used more effectively within the scope of driver / driving behavior, and these models can be used to prevent both cost and work loss.