The reduction in the traffic death rate in the EU has stagnated in recent years, and the goal of halving the number of fatalities in traffic from 2010 to 2020 has not been met yet. The long-term goal of achieving "Vision 0" requires a multidisciplinary approach and shared responsibility between transport infrastructure that "forgives mistakes", traffic regulation, vehicle automation and traffic participants. At pedestrian crossings of the local roads, which do not have traffic lights, reducing the speed of vehicles by applying a traffic sign has not proved to be an effective solution. Vehicle speed is proved to be a key parameter that determines the outcome of a vehicle-pedestrian conflict. The pedestrian island spatially separates the opposite directions of vehicle flows, thus enabling a better perception of the conflict zone, which is important for vulnerable traffic users. In this paper, an analysis of the effectiveness of horizontal discontinuities and pedestrian island, as infrastructural measures for traffic calming, was made. The evaluation was made using microsimulation traffic modeling, and the calibration of the model was made using a neural network based on measured field data. Travel time is a traffic indicator used in the model calibration process. The calibrated model was then used for the analysis of the effects of the planned reconstruction of the road network segment near the children's playground in a residential area in the city of Osijek, Croatia. Statistical analysis was done to determine whether vehicle speeds on the observed road network segment before and after reconstruction differ significantly. Applied methodology showed that the application of microsimulation traffic modeling provides a possibility to assess the effectiveness of the reconstruction of transport infrastructure in the planning phase

reconstruction assessment; pedestrian safety; reducing vehicle speed; traffic modeling; VISSIM