A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Proximity-Based Outlier Detection Method for Roadway Infrastructure Condition Data
AbstractThe quality of roadway condition data is critical for the accuracy of infrastructure management decision support systems and, ultimately, for the confidence in these systems. This paper presents a novel outlier detection method for roadway infrastructure condition data. By taking the spatial and temporal attributes of condition data into account, this method is able to detect outliers and differentiate them into gross and pseudo outliers. The method consists of two major steps. In the first step, homogenous clusters of neighboring roadway sections are identified so that sections within each cluster have the most homogeneous condition-versus-time deterioration patterns. In the second step, outliers within each cluster are detected and delineated into gross outliers (i.e., likely errors) and pseudo outliers (i.e., roadway sections affected by isolated local factors, causing their condition data to be dissimilar to their neighboring sections). The developed method was applied to roadway pavement in Texas. In the future, this method can be extended to other linear infrastructure systems such as pipelines and power transmission lines.
Proximity-Based Outlier Detection Method for Roadway Infrastructure Condition Data
AbstractThe quality of roadway condition data is critical for the accuracy of infrastructure management decision support systems and, ultimately, for the confidence in these systems. This paper presents a novel outlier detection method for roadway infrastructure condition data. By taking the spatial and temporal attributes of condition data into account, this method is able to detect outliers and differentiate them into gross and pseudo outliers. The method consists of two major steps. In the first step, homogenous clusters of neighboring roadway sections are identified so that sections within each cluster have the most homogeneous condition-versus-time deterioration patterns. In the second step, outliers within each cluster are detected and delineated into gross outliers (i.e., likely errors) and pseudo outliers (i.e., roadway sections affected by isolated local factors, causing their condition data to be dissimilar to their neighboring sections). The developed method was applied to roadway pavement in Texas. In the future, this method can be extended to other linear infrastructure systems such as pipelines and power transmission lines.
Proximity-Based Outlier Detection Method for Roadway Infrastructure Condition Data
Gharaibeh, Nasir G (author) / Saliminejad, Siamak
2016
Article (Journal)
English
BKL:
56.03
/
56.03
Methoden im Bauingenieurwesen
Local classification TIB:
770/3130/6500
Proximity-Based Outlier Detection Method for Roadway Infrastructure Condition Data
| British Library Online Contents | 2016