Measuring variation in driving habits between drivers

This paper explores analysis techniques to distinguish between differing drivers of a vehicle. This is a new research challenge as there are numerous examples where there is a need to measure and monitor the ability of a particular driver in shared vehicles. In older adults, cognitive decline or illness can affect their driving ability, younger drivers that have restricted licenses may need monitoring for compliance, and professional drivers are subject to hours of work restrictions. The use of GPS and engine computer data loggers has simplified the recording of detailed driving information, but the identification of the driver remains a challenge as techniques such as RFID tags and user logs all have errors and omissions. The Candrive project is capturing extensive records for 256 vehicles in Ottawa over a 5 year period where GPS and OBDII logs are captured. This paper presents preliminary analysis of 100 trips for 4 different drivers and shows the potential for time of day, road choice, velocity and acceleration data analytics analysis to provide attributes to distinguish between drivers of a vehicle that will enable the future creation of a driving signature.