Artificial neural networks are increasingly employed in prediction modeling and are particularly advantageous when the relationship between the response and the predictor variables is complex. For the purposes of prediction, neural networks are to be trained with data that are accurately compiled. Frequently, the data collected either from field or laboratory observations are noisy in nature. The effect of noisy data on the predictive capability of neural networks has been studied. Present serviceability rating (PSR) of pavements is the attribute to be predicted. Six noisy databases are created and are employed to train the neural networks to predict PSR. Regression equations are developed with the same noisy databases, and the predictions from neural networks are compared with those of regression. The results show that the neural networks predict PSR as accurately as regression models with a given noisy data. In addition, neural networks are trained with data containing no noise. If no noise is present in the data, neural networks predict PSR accurately while properly capturing the effect of each explanatory variable on the response variable.
(LTPP). With the structural details, falling weight deflectometer deflection data, and the distress information compiled from the LTTP information management system data base, the authors use two and four methods for rigid and flexible pavements, respectively, to determine the remaining lives. The remaining lives calculated by two methods for rigid pavements are comparable. Three of four methods for flexible pavements also generated comparable remaining lives. The authors were encouraged by the results and recommend that the survivor curve approach be explored further for network-level remaining life calculations. The reliabilities of various techniques currently available for the remaining life calculation are discussed.
Pavement deterioration models are indispensable for many purposes; as a result, a number of models are in use. Models with simple equation forms are easier to use, but frequently such models may not suffice for many purposes. Consequently, complex nonlinear forms of models are to be considered. However, determination of the solution to a complex model form is not an easy task. There are various methods of obtaining solutions to such models, with each method having its own advantages and disadvantages. The use of genetic algorithms for model development is examined in this study. A very brief description of genetic algorithms is included, and their application for the development of a model is illustrated. Five models of varied complexities, extracted from the literature, are employed to create databases in which the relationship between the response and the predictor variables is known. The solutions to the models are developed employing genetic algorithms. The results indicate a high degree of accuracy, which suggests that genetic algorithms are useful as a tool for development of solutions to pavement deterioration models.
The evaluation of remaining life is necessary to make optimal use of the structural capacity of in-service pavements. It simply represents the useful life left in the pavement until a failure condition is reached. Knowledge of remaining life facilitates decision making in regard to strategies for reconstruction-rehabilitation of roads, thereby leading to the efficient use of existing resources. Several methods proposed or used by various agencies to estimate the remaining lives of pavements are reviewed. They are classified under two categories: functional and structural. Making use of the Mississippi Department of Transportation pavement management system data base, survivor curves are developed for seven classes of flexible pavements with from thin to thick structures. By using these survivor curves a novel method for estimating remaining life is proposed. The reasonableness of the selected methods is examined by putting them to use in calculating the remaining lives of each of eight rigid and flexible pavement sections, all of them from the Mississippi global positioning system sections of the Strategic Highway Research Program–Long-Term Pavement Performance project (LTPP). With the structural details, falling weight deflectometer deflection data, and the distress information compiled from the LTTP information management system data base, the authors use two and four methods for rigid and flexible pavements, respectively, to determine the remaining lives. The remaining lives calculated by two methods for rigid pavements are comparable. Three of four methods for flexible pavements also generated comparable remaining lives. The authors were encouraged by the results and recommend that the survivor curve approach be explored further for network-level remaining life calculations. The reliabilities of various techniques currently available for the remaining life calculation are discussed.
In the Long-Term Pavement Performance (LTPP) program, 35-mm, black and white, continuous-strip photographs are used as a permanent record of pavement distress development for archival purposes and to quantify the distress severity and extent for pavement performance analysis. The traditional method of interpreting distress from LTPP film utilizes a relatively small image projected onto a digitizing tablet. From quality control checks performed on the interpreted data, it was found that some low severity types of distress, identified from larger magnified images projected onto a wall or projection screen, could not be seen in the smaller image used for distress interpretation. The variability in distresses interpreted directly off of the large format, wall-image projection was assessed through analysis of interpretations performed on six asphalt concrete and six portland cement concrete pavement sections used in the LTPP distress rater accreditation workshops. The data set included distress ratings from eight individuals, four two-person rater teams, and an experienced rater team. Also available were distress ratings performed in the field by the experienced rater team, which are used as reference values which represent the best estimate of ground-truth. Statistical tests show that the film-interpreted distresses from individual raters exhibit much larger variability than those from the rating teams. The most significant contributor to this finding is outlier observations in which one of the individual raters had significantly different ratings than the rest of the group. The spread in the rating teams was much lower. The film interpreted distresses from the experienced group correlated very well with the field-derived reference values.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.