The problem of poor quality of traffic accident data assembled in national databases has been addressed in European project InDeV. Vulnerable road users (pedestrians, cyclists, motorcyclists and moped riders) are especially affected by underreporting of accidents and misreporting of injury severity. Analyses of data from the European CARE database shows differences between countries in accident number trends as well as in fatality and injury rates which are difficult to explain. A survey of InDeV project partners from 7 EU countries helped to identify differences in their countries in accident and injury definitions as well as in reporting and data checking procedures. Measures to improve the quality of accident data are proposed such as including pedestrian falls in accident statistics, precisely defining minimum injury and combining police accident records with hospital data.

The paper presents findings from research project Mobis which is aimed at developing a method of assessing safety of unsignalised pedestrian road crossings using video image analysis. Pedestrian and vehicle traffic has been recorded at selected zebra crossing sites in Warsaw and Wrocław, before and after installation of active signage systems SignFlash and Levelite. Speeds of approaching vehicles were measured and drivers’ behaviour was classified using video analysis.

The paper presents a comparison of effectiveness of systems such as SignFlash and Levelite based on changes in the mean and standard deviation of vehicle spot speeds as well as changes in speed profiles of vehicles approaching the crossings.

Results indicate that both SignFlash and Levelite active signage reduce mean vehicle approach speeds and have a positive impact on drivers’ behaviour.