Annual Average Daily Traffic (AADT) is an important parameter for traffic engineering analysis. Departments of Transportation continually collect traffic count using both permanent count stations (i.e., Automatic Traffic Recorders or ATRs) and temporary short-term count stations. In South Carolina, 87% of the ATRs are located on interstates and arterial highways. For most secondary highways (i.e., collectors and local roads), AADT is estimated based on short-term counts. This paper develops AADT estimation models for different roadway functional classes with two machine learning techniques: Support Vector Regression (SVR) and Artificial Neural Network (ANN). The models predict AADT from short-term counts. The results are first compared against each other, using the 2011 ATR data, to identify the best models. Then, the results of the best models are compared against both the regression-based model and factor-based model. The comparison reveals the superiority of the SVR model for AADT estimation for different roadway functional classes over all other methods. Among models for different roadway functional classes, developed with the 2011 ATR data, the SVR-based models show minimum errors in estimating AADT compared to the ANN-based, regression-based, and factor-based models, depicting the superiority of the SVR-based model for all roadway functional classes over other models in terms of AADT estimation accuracy. SVR models are validated for each roadway functional class using the 2016 ATR data and short-term count data collected by the South Carolina Department of Transportation (SCDOT). The validation results show that the SVR-based AADT estimation models can be used by the SCDOT as a reliable option to predict AADT from the short-term counts.
Improving Group Assignment and AADT Estimation Accuracy of Short-term Traffic Counts using Historical Seasonal Patterns & Bayesian Statistics
