Traditionally, measurements of traffic behaviour for road safety research have been done manually. This has been only partly caused by the lack of sophisticated other means of data collection. We also know that interpretation of human behaviour is very difficult and requires usually the close presence of a qualified human observer. Nevertheless, we can foresee a transition from manual measurements and counts to an increased use of automatic ones for several reasons. Firstly, technological advances have produced new, reliable means of collecting data on traffic behaviour also in a quite comprehensive way. Secondly, automatic counts produce objective data in a repeatable manner. All manual forms of data collection suffer from reliability problems related to the observers’ subjective interpretation of the behaviour. Thirdly, automatic counts provide a means to collect extensive data sets of traffic behaviour usually for a relatively modest cost. Furthermore, automatic counts are the only practical way to collect data on traffic behaviour in rare circumstances, when it is difficult to predict the occurrence of such conditions. An example of this is data collection from slippery road conditions, which is extremely difficult by manual means. If the observers would rely on weather forecasts, they could very seldom make observations on slippery road surfaces as the road maintenance would have spread salt or sand on the road before the observers had reached the location. The road maintenance operators also rely on forecasts and their forecast services are the most sophisticated in any country. Hence manual observers would have to be in constant stand-by close to the road sections to be studied, which is extremely tedious and costly. With automatic counts, comprehensive data collection on speeds, headways and flows is possible with minor costs (see e.g. Rämä et al 1996). Finally, a large number of systems collecting and registering data on traffic behaviour exists already and the number of systems in full operation is rapidly growing. So far, only a few of these are being utilised for collecting data on traffic behaviour for safety research purposes as their main applications lie elsewhere. Nevertheless, these systems produce data that could very easily be also utilised in safety research with very low additional costs and labour. (Author/publisher)
Samenvatting