Actual Driving Data Analysis for Design Consistency Evaluation

Abstract

In recent years, researchers have proved that a consistent highway design ensures that successive elements are coordinated to produce harmonious and homogeneous driver performances and does not provoke unexpected events. Knowledge and practice show that drivers make fewer errors near geometric features that conform to their expectations. On this basis, the importance of identifying inconsistencies on highways for its significant contribution to road safety is emerging as an important feature in highway design. Although several techniques and models for evaluating the consistency of a design in a quantitative way have been identified and some countries have implemented the design consistency concept in their road design guidelines mainly in a qualitative way, there have been only a few efforts to measure actual driving behavior. The aim of this paper was to determine design inconsistencies on existing two-lane rural roads with the use of actual driving behavior by means of field data measurements and to verify their agreement with a consistency evaluation model. Furthermore, suitable equipment and a procedure for surveying driving dynamics and driver workload have been developed. In particular, driving behavior was assessed through direct measurements and parameters taken from data collected on a selected sample of test drivers by using a purposely designed instrumented vehicle. The vehicle, named the driver instrumented vehicle acquisition system (DIVAS), was driven under real traffic conditions on a two-lane rural road. The design classes of consistency of the test courses also were evaluated with a well-known safety criteria model. Data collection and treatment procedures are presented, and data analysis and results from this first experiment are given.