The effectiveness of variable message signs (VMS) in improving safety certainly depends on a chain of requirements including the quality of traffic data collection and processing, the performance of automatic incident detection algorithms, correct driver perception of the meaning of the signs, and finally appropriate driver behaviour in response. The present paper focuses on the question of correct driver perception of VMS and the relation to appropriate behaviour. In a controlled perception test performed on a sample of Munich drivers, most drivers correctly interpreted the meaning of typical and proposed combinations of pictograms and messages. Nonetheless, statistical evidence for a direct response in the speed level of a highway to variable speed limits is very weak if confounding factors such as traffic context are properly taken into account. Hence, since the signs are being perceived properly and accidents are in fact decreased, the mechanism for safety improvement does not appear to be speed level decrease itself, but rather more subtle effects of variable messages on traffic flow such as more farsighted driving due to the perception of increased danger. If this is the mechanism behind improved safety, then more effort should be concentrated on characterising and enhancing this response mode.
Abstract