Traffic flow volumes present key information needed for making transportation engineering and planning decisions. Accurate traffic volume count has many applications including: roadway planning, design, air quality compliance, travel model validation, and administrative purposes. Traffic counts also serve as an important input in highway safety performance evaluation. However, collecting traffic volume on all rural non-state roads has been very limited for various reasons, although these roads constitute a great portion (60 to 70%) of road mileage in the roadway network of any state in the U.S. For example, out of 61,335 miles of roadway in Louisiana, 73% of the roadways are non-state roads. Due to limited resources, traffic volume information on non-state roadways has not been systematically collected in Louisiana. Generally, traffic volumes on these roads are fairly low, and VMT on these roads is much less compared with that on interstate or arterial roads. Thus, regularly conducting traffic count is not economically feasible for non-state roadways. This study develops an AADT estimation methodology by using modern statistical and pattern recognition methods. By using available traffic counts on non-state roadway and four variables (namely: population, job, and distance to intersection and to major state highways at block level), a training set to estimate roadway AADT for eight parishes were obtained by a modified support vector regression (SVR) method. This pattern recognition method yields better AADT estimates than the conventional parametric statistical methods. Sensitivity analyses were also conducted in this study, which indicates a parish-specific model works better than an aggregated single model. With the estimated AADT, the DOTD and local government agencies can make better decisions on funding allocations for safety improvement projects and pavement maintenance actions. The estimated parish-specific AADT on non-state roads can also improve statewide travel demand forecasting models and air quality assessment.
Developing a Method for Estimating AADT on all Louisiana Roads
2015
97 pages
Report
No indication
English
Application of Neural Networks to Estimate AADT on Low-Volume Roads
British Library Online Contents | 2001
|Application of Neural Networks to Estimate AADT on Low-Volume Roads
Online Contents | 2001
|AADT: providing truth in comparative traffic safety
Automotive engineering | 2014
|