Estimation of Primary Task Demand using Naturalistic Field Operational Test Data

Abstract: Driver’s attention on the driving task is vital for safe travel. Intelligent vehicle systems(IVSs) are systems which assist the driver by, for instance, providing extra information aboutthe vehicle, the environment, or the driver to make driving more comfortable and safe. IVSshave a high potential to improve driving, however they should not distract the driver by, forexample, inducing the driver to take his/her eyes off the road in critical situations; rather IVSsshall direct the driver to shift focus to the direction of immediate danger.The objective of this thesis was to estimate ‘primary task demand’ (i.e. create a mathematicalmodel describing the demand of the current driving situation) using naturalistic driving data.Such estimation may be used to increase the effectiveness of IVSs by 1) preventing suchsystems from distracting the driver when the primary task demand is high and 2) helping suchsystems providing information according to the current driving situation.Naturalistic driving data from the SeMiFOT field operational test was used to select severaldriving situations which were ranked by 40 drivers based on their perceived primary taskdemand. Further, this study focused on roundabouts since they are a very common scenariofor low speed crashes and require continuous maneuvering from the driver. Finally, this studyalso elucidates the extent to which perception of primary task demand is influenced bycultural difference by comparing results from 20 Indians and 20 European drivers.