Utilizing Real-time Traffic and Weather Data to Explore Crash Frequency on Urban Motorways: A Cusp Catastrophe Approach

Abstract

Abstract The investigation of crash frequency with freeway traffic and weather data has recently received significant attention by researchers. This paper extends previous research by proposing nonlinear models for modeling crash frequencies which incorporate real-time traffic and weather data collected from an urban motorway in Athens, Greece. Cusp catastrophe theory is applied and compared with traditional statistical models such as the negative binomial model. The results of crash frequency models provide evidence of the potential applicability of the cusp catastrophe theory to road safety, however it seems that linearity of the system is preserved. Hence, traditional models such as the negative binomial model are proved equally capable of describing the underlying phenomenon, even though the goodness-of-fit is not as good as that of the cusp model. Therefore, the explanation of crash frequency phenomenon only with nonlinear model can be supported. A number of interesting findings have also been disclosed. Firstly, is that rainfall intensity has a strong linear impact on crashes (high rainfall intensity causes more crashes). On the other hand, average flow is indicated to have a strong non-linear relationship with crash frequency. Finally, more research is needed to further understand the applicability of cusp catastrophe theory in road safety.