Daniel Cunnington Undergraduate Dissertation 2014/15

Monitoring Driver Behaviour

Supervised by F.Ciravegna

Abstract

Monitoring driver behaviour is a complex task often carried out by human driving instructors. When a driver is learning to drive, they are exposed to a large amount of feedback especially regarding specific dangerous manoeuvres that they carry out. This enables them to reflect on their actions and improve the safety of their driving. Experienced drivers, or drivers who have passed their driving test do not have access to the same level of feedback as a learner driver. The only time feedback may occur is if the driver is involved in a near miss or an accident which of course may be too late.

This project aims to produce a system to allow a driver to monitor their own driving behaviour, ensuring they are exposed to regular feedback. The overall solution will be composed of a data collection mechanism and a means of data visualisation. Data will be collected about the internal performance of a vehicle through the On-Board Diagnostics (OBD) port which all European cars manufactured after 2001 support. Travel information such as acceleration or turning events will be detected through the use of smartphone sensors, as these events provide a good indicator for driver behaviour. Data will also be collected regarding the external driving environment such as weather conditions, because a driver exhibiting safe driving behaviour will adapt their driving to the external environment.

On completion of the project, a successful outcome has been reached. In my experiments, the system is able to correctly distinguish safe driving from unsafe driving through the detection and penalisation of specific driving events. The project has also provided a platform for further research into the field of driver behaviour.