Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Mining Marine Vessel AIS Data to Inform Coastal Structure Management
This study demonstrates the use of a multiyear country-scale automatic identification system (AIS) data set to partition a nationwide portfolio of navigation structures managed by the US Army Corps of Engineers (USACE) into affinity groups based on emergent vessel traffic characteristics. The marine vessel AIS was originally intended to prevent the collision of ships at sea. As a remote sensing technology, it provides continuous monitoring for marine vessel traffic and has enabled a variety of unforeseen applications. The methodology presented uses spatial distance criteria to identify vessel traffic local to each structure. Metrics characterizing traffic behavior including traffic composition, spatial position, trip frequency, and traffic seasonality are derived from vessel data. AIS-derived metrics are combined into feature vectors describing each structure. Pearson correlation of feature vectors with r-neighborhood pruning of the affinity matrix is used to identify similar structure pairs. Semisynchronous label propagation is used to partition the structure portfolio graph into prototype groups with strong similarity in underlying traffic characteristics that may be further refined to align maintenance activity with organizational goals.
Mining Marine Vessel AIS Data to Inform Coastal Structure Management
This study demonstrates the use of a multiyear country-scale automatic identification system (AIS) data set to partition a nationwide portfolio of navigation structures managed by the US Army Corps of Engineers (USACE) into affinity groups based on emergent vessel traffic characteristics. The marine vessel AIS was originally intended to prevent the collision of ships at sea. As a remote sensing technology, it provides continuous monitoring for marine vessel traffic and has enabled a variety of unforeseen applications. The methodology presented uses spatial distance criteria to identify vessel traffic local to each structure. Metrics characterizing traffic behavior including traffic composition, spatial position, trip frequency, and traffic seasonality are derived from vessel data. AIS-derived metrics are combined into feature vectors describing each structure. Pearson correlation of feature vectors with r-neighborhood pruning of the affinity matrix is used to identify similar structure pairs. Semisynchronous label propagation is used to partition the structure portfolio graph into prototype groups with strong similarity in underlying traffic characteristics that may be further refined to align maintenance activity with organizational goals.
Mining Marine Vessel AIS Data to Inform Coastal Structure Management
Scully, Brandan M. (Autor:in) / Young, David L. (Autor:in) / Ross, James E. (Autor:in)
08.12.2019
Aufsatz (Zeitschrift)
Elektronische Ressource
Unbekannt
Mining Customer Data to Better Inform Utility Decision-Making
| British Library Conference Proceedings | 2013