Over the years, more digital data is transmitted between Electronic Control Units (ECU) in the car, using the CAN-Bus protocol. By exploiting the data packets, we can discover a wide range of information about who drives the car, what's his status, etc.
In this project, we focused on processing the already extracted raw data, for the usage of classifying the driver, from a set of 10 drivers, while training on road number 1 and testing on road number 2.
Through this project, we compared different types of classifiers, and achieved the best results from Random Forest Tree classifier, implemented with XGBoost algorithm.
Also, we discovered a limitation in the chosen method – we couldn’t classify certain drivers.