This paper presents a calibration and validation procedure for microsimulation models, which used metrics (mean, variance, and mean absolute percentage error) with statistical tests (t-test, Shapiro-Wilk, Kolmogorov-Smirnov and Wilcoxon), to compare empirical and simulated data. A genetic algorithm was used to identify calibration parameters set. The paper justifies the approach using VISSIM microsimulations to analyze two safety countermeasures: Advance Detection System and Advance Warning System, which can be applied at signalized intersection. The end result was a calibrated and validated model, which could be used to compare different safety countermeasures at rural signalized intersections in the state of Nebraska (USA). The proposed approach could be utilized in similar studies.
