An Unsupervised Approach for Driving Behavior Analysis of Professional Truck Drivers

Modern vehicles can generate up to several Gigabytes of data per day which are mostly used only for aspects directly related to the proper functioning of the vehicle itself. However, these data have an enormous value as they can be collected and analyzed to better understand additional aspects of the driving experience, such as classifying the driver's behavior and driving style. In this paper, we present a simple yet novel unsupervised methodology that is able to classify the behavior of a driver in a certain geographical area on the basis of the data collected from all the drivers in the same area. The proposed methodology has been tested on two different datasets involving professional truck drivers and it has been verified using human labelled ground truth data. The results obtained demonstrate the feasibility of the proposed solution. To our knowledge, this is the first study to classify driving behaviours of professional truck drivers and validate their performance on such large-scale data with actual safety scores.