Research Projects

This proposed project would aim at development of a comprehensive scientific methodology to identify the key determinants of road-traffic-accidents and improve the road traffic safety condition in Indian urban roadways based on advanced statistical techniques. For the present study the Hyderabad city has been considered. This proposed project is planned to be carried out through five modules. The first module would include the data collection and subsequent database development. Accident history data of the network, geometric and traffic characteristics of all locations will be collected from both primary and secondary sources and a comprehensive crash database for Hyderabad city would be developed. The following module would aim at identifying the key factors influencing road traffic accidents in urban sections. In this module, the factors influencing crash severities in urban sections will be investigated based on injury severity modelling through unordered and ordered discrete choice models and non-parametric classification tree techniques. The subsequent module develops safety performance functions SPF for segments and intersections for urban areas for total, fatal and non-fatal crashes using negative binomial regression. In this module, a unique segmentation approach will be adopted to divide the case study stretch into a set of homogenous roadways segments, for which separate SPF will be estimated. These SPF's will further be tested for temporal and spatial stability, such that these can be adopted for other cities with similar characteristics. Safety performance functions could be extensively adopted for predicting the crash frequency at a particular location. In the fourth module, the predicted crash frequency from SPF's and observed crash frequency will be combined to estimate the Empirical Bayes (EB) estimate at a particular location and subsequently, segments will be then ranked based on decreasing order of EB or the segments with highest EB will be top ranked or designated as hotspots. The final module of the proposed project would aim at prioritizing the most cost-effective crash countermeasure at a particular hotspot. Initially, a set of countermeasures appropriate for the selected hotspot would be identified and subsequently, their economic appraisal will be conducted to prioritize the countermeasures with highest return for efficient utilization of Government expenditure. Overall, this project would be developing effective road safety interventions through rigorous statistical analysis of accident data. The key conclusions and findings will be instrumental in formulating national level and state level road safety policy since the factual causes and consequences of road accidents are unknown. The results of this study, although case-specific would be tested for transferability. However, the methodology developed could be used with findings serving as basis for comparison across other cities with similar size and characteristics.