The accuracy of pavement condition prediction is a major concern associated with a pavement management system (PMS). The PMS used by the Kansas Department of Transportation (KDOT) includes a project-level optimization system that requires models that estimate the probability of a given level of distress occurring. These models are based on historical pavement condition data and specific project-level data concerning pavement structural characteristics, traffic, and climatic conditions. Multiple linear regression and two artificial neural network (ANN) structures are used to predict roughness distress level probability for bituminous pavements as defined by the KDOT PMS. Since the response variable is the probability of being in a given roughness distress level, within the historical database the probability values are binary (1 if the pavement exists in a given roughness distress level or 0 if the pavement is in any other roughness distress level). This produces poorly conditioned data for regression analysis. Therefore, results indicate that ANNs have a superior ability to predict the probability of roughness distress level compared with multiple regression methods.
Roughness Level Probability Prediction Using Artificial Neural Networks
Transportation Research Record
Transportation Research Record: Journal of the Transportation Research Board ; 1592 , 1 ; 89-97
1997-01-01
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
SURFACE ROUGHNESS PREDICTION USING HYBRID NEURAL NETWORKS
British Library Online Contents | 2007
|Predicting Roughness Progression in Flexible Pavements Using Artificial Neural Networks
British Library Conference Proceedings | 1994
|Road Traffic Prediction Using Artificial Neural Networks
IEEE | 2018
|Prediction of Runoff Using Artificial Neural Networks
HENRY – Bundesanstalt für Wasserbau (BAW) | 2010
|DGPS Correction Prediction Using Artificial Neural Networks
Online Contents | 2007
|