Validation and Sensitivity Analysis of a Developed Midblock Vehicular Delay Model

Abstract

This paper develops and validates a deterministic model that estimates the delay incurred by platoon vehicles due to pedestrian activity at midblock crosswalks on urban street segments. The developed model is validated using midblock delay data measured at an urban street segment in Newark, New Jersey. The model validation involves both quantitative and qualitative statistical analyses to compare the delay values measured in the field with those estimated using the developed model. The quantitative statistical analysis involves performing the t-tests using commercially available statistics software. The qualitative analysis involves the use of a diagonal plot to compare the measured and estimated delay data sets. Sensitivity analysis is performed to study the relationship between estimated midblock delay and the model’s variables and parameters. The sensitivity analysis involves estimating the delay by increasing and decreasing the baseline values of the model variable and parameters by 50%. Results of the t-test show no statistically significant difference between the measured and estimated delays. Result of the sensitivity analysis shows pedestrian walking time significantly impacts vehicular delay.