Date of Graduation


Document Type


Degree Name

Doctor of Philosophy in Engineering (PhD)

Degree Level



Civil Engineering


Sarah Hernandez

Committee Member

Heather Nachtmann

Second Committee Member

John Kent

Third Committee Member

Kevin Hall


Automatic Identification System (AIS), Commodity, Data fusion, Inland Waterway Transportation, Multimodal freight, Transportation planning


Within the U.S., the 18.6 billion tons of goods currently moved along the multimodal transportation system are expected to grow 51% by 2045. Most of those goods are transported by roadways. However, several benefits can be realized by shippers and consumers by shifting freight to more efficient modes, such as inland waterways, or adopting a multimodal scheme. To support such freight growth sustainably and efficiently, federal legislation calls for the development of plans, methods, and tools to identify and prioritize future multimodal transportation infrastructure needs. However, given the historical mode-specific approach to freight data collection, analysis, and modeling, challenges remain to adopt a fully multimodal approach that integrates underrepresented modes, such as waterways, into multimodal forecasting tools to identify and prioritize transportation infrastructure needs. Examples of such challenges are data heterogeneity, confidentiality, limitations in terms of spatial and temporal coverage, high cost associated with data collection, subjectivity in surveys responses, etc. To overcome these challenges, this work fuses data across a variety of novel transportation sources to close existing gaps in freight data needed to support multimodal long-range freight planning. In particular, the objective of this work is to develop methods to allow integration of inland waterway transportation into commodity-based freight forecasting models, by leveraging Automatic Identification System (AIS) data. The following approaches are presented in this dissertation:

i) Maritime Automatic Identification System (AIS) data is mapped to a detailed inland navigable waterway network, allowing for an improved representation of waterway modes into multimodal freight travel demand models which currently suffer from unbalanced representation of waterways. Validation results show the model correctly identifies 84% stops at inland waterway ports and 83.5% of trips crossing locks.

ii) AIS and truck Global Positioning System (GPS) data are fused to a multimodal network to identify the area of impact of a freight investment, providing a single methodology and data source to compare and contrast diverse transportation infrastructure investments. This method identifies parallel truck and vessel flows indicating potential for modal shift.

iii) Truck GPS and maritime Lock Performance Monitoring System (LPMS) data are fused via a multi-commodity assignment model to characterize and quantify annual commodity throughput at port terminals on inland waterways, generating new data from public datasets, to support estimation of commodity-based freight fluidity performance measures. Results show that 84% of ports had less than a 20% difference between estimated and observed truck volumes.

iv) AIS, LPMS, and truck GPS datasets are fused to disaggregate estimated annual commodity port throughput to vessel trips on inland waterways. Vessel trips characterized by port of origin, destination, path, timestamp, and commodity carried, are mapped to a detailed inland waterway network, allowing for a detailed commodity flow analysis, previously unavailable in the public domain.

The novel, repeatable, data-driven methods and models proposed in this work are applied to the 43 freight port terminals located on the Arkansas River. These models help to evaluate network performance, identify and prioritize multimodal freight transportation infrastructure needs, and introduce a unique focus on modal shift towards inland waterway transportation.