Although studies of speed behaviour and accidents on two lane rural roads are available, they are based on different methods and are not comparable. Nor do they reflect present driving behaviour, driver population and vehicle technology. Speed recordings and accident surveys were made on selected two lane rural roads in 1991 in order to demonstrate correlations between accidents, traffic and roadside data. From studies of papers concerned with the "driver-vehicle-load" loop system, two research groups can be formed: Studies of (1) correlations between accidents and elements of road outline; and (2) individual driving behaviour in relation to boundary conditions of traffic and road distribution. Parameters judged relevant were width of carriageway, longitudinal gradient, curvature or alignment, traffic volume, accident rate, accident cost rate (accident severity), speed. No clear correlation was found between accident parameters and speed parameters. Lack of correlation may be explained by small number of accidents - ca 38 per route - and the fact that accidents were often concentrated at individual critical sites such as curves and summits. One practical result is that increasing the width of carriageway reduces the accident parameter; the same applies to good alignment. Although stretches of high standard have the highest speeds, they also have the smallest number of accidents.
Abstract