Simulator Evaluation of Drivers' Performance on Rural Highways in relation to Drivers' Visual Attention Demands

Abstract

Aim of the study is to investigate, by means of a driving simulator experiment, drivers' performance in terms of lateral position, speed, deceleration, steering angle, and breaking times on a divided two-lane rural highway in relation to drivers' visual attention (VD). In the experiment, the virtual scene of twenty different geometric alignment sections without traffic and the VD testing were designed. Twenty-three experienced drivers with the calibration of attention capacity participated in a 30 km drive in an interactive fixed-base simulator. Each participant was required to drive with the controlled speed of 60 km/h along the central lane as repeating random number and was evaluated on VD and driving performances. Three different data analysis techniques were used: (a) statistical tests and hypothesis test of curvature change rate (CCR) of the geometric alignments, visual attention demands, and driving performance data, (b) correlation analysis of VD, CCRs, and driving behaviors, and (c) regression analysis of the VD and CCRs. Results have showed that the driving performance can be effectively influenced by the highway alignment and a prediction model built in this study can evaluate the drivers' visual attention demands before the highway constructed. The interactions among VD, driving behavior, and CCRs were also found.