Compliance with Preventative Traffic Management Strategies
STEINHOFF C, KATES R and KELLER H, Technical University of Munich, Germany
Traffic management systems are often implemented in an environment characterised by dangerous situations such as sudden speed drops occurring due to downstream bottlenecks, traffic instability, or other incidents. Effective traffic management can have a s
Traffic management systems are often implemented in an environment characterised by dangerous situations such as sudden speed drops occurring due to downstream bottlenecks, traffic instability, or other incidents. Effective traffic management can have a significant positive impact on traffic safety if dangerous situations are promptly recognised and the system induces drivers to modify their behaviour appropriately. The latent or inherent danger in a traffic situation is difficult to characterise or quantify. Nonetheless, a substantial body of evidence suggests that any sharp negative downstream speed gradient in the flow of traffic has a high accident potential. Negative speed gradients may be caused by downstream capacity reductions due to incidents, construction, or spontaneous traffic instabilities (HooPs, KATES, KELLER, 1999).
In the case of an existing incident, the strategy is to warn upstream drivers before they reach the speed drop and reduce the chance of secondary accidents. The justification of the warning is usually seen downstream unless the warning is a false alarm.
By an appropriate intervention (reduced speed limit, truck passing prohibition), an impending traffic instability may sometimes be prevented before it actually occurs. Hence, forecasting an impending traffic instability offers two mechanisms for improving safety: warning and prevention. The potential value of preventative strategies is thus quite high -- if one believes that it is possible to forecast an instability -- because the slight increase in trip times due to preventative interventions is a small price to pay for added traffic safety and avoidance of traffic jams. However, the necessity for a preventative intervention is less likely to be confirmed than restrictions due to weather or incidents: For one thing, instability forecasting is probabilistic to begin with. (Note however that automatic incident detection algorithms usually rely on partial traffic information and are thus also subject to uncertainties.) For another, by preventing the occurrence of an instability, a preventative strategy also suppresses the evidence for its own necessity!
A mediating consideration is that preventative management strategies (at least in Germany) typically consist of mild restrictions (e.g., 120 km/h or 100 km/h and/or passing restrictions for trucks). Hence, drivers are unlikely to perceive such strategies as obvious malfunctions even if the underlying instability forecasting algorithms have low specificity (frequent alarms when instability would not have occurred anyway) or low sensitivity (failure to predict spontaneous speed drops that do occur). Nonetheless, an excessive number of "superfluous" restrictions due to preventative strategies is undesirable, because they could induce feedback processes unfavourably influencing the compliance rate.
Association for European Transport