The viability of real-time prediction and prevention of traffic accidents

This paper addresses a novel idea of real-time traffic safety improvement on freeways. First by predicting traffic crashes on the freeway mainline using on-line loop detector data, then by proposing ITS remediation strategies to reduce the crash risk in real-time. The results of the predictive models were found to be consistent with the probable mechanisms of crashes under different flow regimes. Using micro-simulation software and the crash potential determined through the models developed in this research to associate traffic characteristics with the risk of a crash occurring, several mitigation strategies were attempted to improve safety in real-time. The before (pre-ITS) and after case were then compared to examine the effects of the crash mitigation methods. The results show that Variable Speed Limits and Ramp metering have positive effects on lowering the crash potential.