A microsimulation approach was carried out in this paper to evaluate the safety performance of turbo roundabouts in which the "CAVs" of connected autonomous vehicles are mixed with the "CVs" of conventional vehicles the research aims to evaluate the advantages in terms of safety and performance of turbo roundabouts. The paper shall also lead to describe the methodological path followed to build VISSIM models of turbo roundabout changing O-D matrix as real case applications, to calibrate the simulation models, and to estimate the potential conflicts when the percentages of CAVs are introduced into the traffic mix. The results, in accordance with the existing bibliography, have shown that the safety levels and the parameters that determine an improvement in the service level in a turbo roundabouts are significantly influenced not only by the geometric characteristics, but also by the distribution of vehicular flows. Therefore, it follows that in absence of crash data including CAVs, the surrogate measures of safety must be considered a strong approach to evaluate the safety performance of a roundabout so far, any road entity.

Safety Performance Functions and Crash Modification Factors are statistically-based prediction methods that require significant efforts and long periods in crash data collection. Traffic conflict studies can mitigate this issue using a short time survey to measure the number and severity of traffic conflicts, which are regarded as surrogate safety measures. Unfortunately, they are empirical studies that can be carried out only after the implementation of a treatment. The overall objective of the present research is to investigate the performance of different methods for conflict detection and classification, considering the observed conflicts on 2+1 roads in Poland. Observations were compared with conflicts detected in simulated environments. The latter include either the Agent-Based Microsimulation (ABM) approach, or the virtual reality simulation using a Driving Simulator (DS). Conflicts were detected and classified based on video recording and analysis of vehicle trajectories in the merging area of 2+1 roads. The studies focused only on lane-changing conflicts. Locations, Post Encroachment Time and Time to Collision values of observed conflicts between vehicles were subsequently identified. Observed conflicts were compared with the ones resulting from ABM and DS, to determine whether there is a correlation between them.