Annual Average Daily Traffic Prediction Model for Minor Roads at Intersections

Abstract

Annual average daily traffic (AADT) is an important element for maintenance, safety, environmental analysis, finance, engineering economics, and performance management. Most previous studies were conducted to estimate AADT on the road segment instead of considering the intersection, and did not well consider the possible difference of AADT between the major road and minor road at intersections. The present research was conducted to develop a model that can estimate AADT for minor roads at intersections using available Highway Performance Monitoring System (HPMS) data owned by state departments of transportation and Census data. The performance of multiple linear regression, random forest, and neural network were compared in the study. Multiple regression analysis was selected to develop an estimation function for the minor road AADT. The AADT on the major road, the functional class of the major road and minor road, and the number of traffic lanes on the major road and minor road were selected as the input of the regression model, which was based on the statistical analysis. A multiple regression model with logarithmic transmission was selected for AADT estimation. The cross-validation showed the high accuracy of the developed model. The equation generated in this paper can be easily used by transportation agencies for AADT estimation on minor roads at intersections.