Identification of critical conflicts and its implementation in microsimulation for effective safety assessment in unsignalised intersections

Traffic safety assessment through microsimulation using Surrogate Safety Measures (SSM) has gained significant attention. Despite researchers proposing effective approaches for critical conflict identification in manual safety assessment, integrating these approaches into microsimulation is yet to be explored. This study provides a methodology for implementing critical conflict identification into microsimulation and calibrating and validating the models with selected SSM's distribution as Measure of Effectiveness (MoE). Post Encroachment Time was used to assess the safety of crossing and merging traffic in unsignalized intersections. Genetic Algorithm was used to calibrate the model with PET distribution of the identified critical conflicts as MoE. The calibrated models were used to investigate the effectiveness of various preventive measures after validation. Roundabout was found to be reasonably effective in terms of crash frequencies estimated using Extreme Value Theory as it reduces the speed of the crossing and merging traffic and channelises the traffic in the study location.