PREVIOUS RESEARCH DEVELOPED TECHNIQUES FOR PREDICTING THE LIKELIHOOD OF COLLISIONS WITH ROADSIDE OBJECTS. THESE TECHNIQUES HAVE SINCE BEEN TESTED IN BOTH SYDNEY AND MELBOURNE. THE PAPER DESCRIBES THE PROCEDURES USED IN MELBOURNE FOR COMPARING PREDICTED ACCIDENT RATES WITH RATES DETERMINED FROM REPORTED ACCIDENT STATISTICS. THE TECHNIQUES WERE FOUND TO BE REASONABLY EFFECTIVE FOR ROADSIDE OBJECTS AT MID-BLOCK LOCATIONS ON MAJOR ROADS, BUT QUITE INEFFECTIVE FOR OTHER LOCATIONS. THE TECHNIQUES HAVE A NUMBER OF THEORETICAL WEAKNESSES, AND SEVERAL FEATURES WHICH MAKE THEIR APPLICATION DIFFICULT OR INACCURATE. AN APPROACH IS OUTLINED WHICH ALLOWS THE EXPLANATORY POWER OF THE TECHNIQUES TO BE UTILISED, IN ASSOCIATION WITH HISTORICAL ACCIDENT DATA, FOR THE IDENTIFICATION OF HAZARDOUS ROADSIDE OBJECTS AND THE EVALUATION OF ALTERNATIVE REMEDIAL MEASURES (A). THE NUMBER OF THE COVERING ABSTRACT OF THE CONFERENCE IS IRRD NO 255423.
Samenvatting