Skip to main content

Mining AIS Data for Improved Vessel Trip Analysis Capabilities

Project Summary

Automatic identification systems (AIS) provide communication between vessels and assist vessel traffic control functions in congested ports, locks and waterways. Vessels digitally broadcast information, including position, call sign, vessel name, course, speed, and navigation status. Publicly available waterway trip data is aggregated to protect business confidentiality of shippers. However, the level of aggregation renders the data unusable for a variety of applications, including waterway congestion, demand and risk analysis. CFIRE researchers at the University of Toledo (UT) and Vanderbilt University (VU) have independently conducted research into collecting and analyzing AIS data on the Great Lakes and inland waterways, respectively. The work completed by these researchers has demonstrated the feasibility of using shore-based AIS receivers to acquire, archive and analyze data on vessel movements. The integration of relational database management systems (RDBMS), geographic information systems (GIS) and custom software routines is a powerful combination of technology that is capable of leveraging the AIS data.

This project will develop a methodology for processing AIS data from multiple sites in near real-time as well as develop a capability to support ad-hoc data query. Such analyses can identify high-risk and high-traffic locations, generate better travel time estimates, detect vessel arrivals, identify key traffic areas for investment and enhancement, enable terminal operators to better manage their operations, and in general lead to a better understanding of vessel traffic within a given area. Benefits of AIS automation include increased levels of information regarding boat location and trajectory which supports safer waterway operations, more efficient lock operation and lead times, and increased shipping efficiencies as waterway traffic is better understood and managed. Key research tasks include installing additional shore-based AIS antennas at strategically selected Great Lakes and inland waterway locations, and developing the server and database management system that can quickly receive and process the data from multiple sites.

Vanderbilt University recently completed a 6-month study of inland AIS data at an Ingram Barge Company facility in Paducah, Kentucky in 2011. Despite the complex river port configuration (Paducah is located at the meeting of the Cumberland, Ohio, and Tennessee Rivers), researchers successfully deployed models to generate useful trip data as well as detect vessel operations such as fleeting, docking and lockages. The Ingram Barge Company is in the process of installing another AIS antenna at Reserve, LA, located between Baton Rouge and New Orleans, LA. This location is expected to produce more data (as AIS carriage is required on this section of the Mississippi River) and additional challenges owing to the mix of inland and deepwater vessel traffic. The Paducah antenna remains operational.
These data assets will provide the necessary raw material for the desired outcomes listed below.

Desired Outcomes

This project is expected to produce a methodology to process large amounts of AIS data in near real-time from geographically diverse AIS antenna locations. The value of such archived AIS data will be demonstrated using a case study approach that will include a range of operational performance measures of the fleet and port infrastructure facilities.
Several analytical approaches will be followed:

  1. Correlation of AIS trip data with significant and extreme weather events (high/low water conditions, reduced visibility, high winds, etc.)
  2. Use of additional data sources (e.g., OMNI XML web service) to generate additional information and trip data beyond AIS coverage areas
  3. Vessel type and size distribution in AIS antenna coverage areas
  4. Compilation and aggregation of vessel traffic figures in key areas to document needs for infrastructure investment
  5. Statistical analysis of vessel speeds through AIS antenna coverage areas
  6. Automated detection of vessel arrivals and departures at dock and terminal facilities along with terminal dwell times
  7. Automated detection of vessel stop events (e.g., lightering, queuing, fleeting areas, etc.) within harbor facilities
  8. Relation of vessel stop events to navigation points of interest

VU will lead the inland rivers component. The 2011 Paducah AIS analysis work was conducted on a 6-month archive. For this reason, VU will focus much of this research on the transition from historical data analysis to efficient data storage, visualization and real-time data processing. The following is a comprehensive list of tasks to be performed during the research:

  1. Development of a normalized relational database design – a normalized design will be generated to efficiently store raw AIS messages. This design will be converted to a data warehouse during the project to efficiently archive the raw data and enable swift queries of the large data repository. The database platform will either be Oracle 11g or Microsoft SQL Server 2012. (2 months)
  2. Develop code to process data as received from the remote AIS servers and add data to the database. (4 months)
  3. Relate real-time AIS data to U.S. Coast Guard static vessel data – the AIS vessel data collected in Paducah and Reserve will be joined to the USCG vessel database. Static particulars about the data (including length, beam, horsepower, vessel type, etc.) will then be available (and query-able) for the AIS case study areas. (2 months)
  4. Spatially-enable live data in GIS – this task will create views based upon the most recent AIS data received and convert them to a GIS layer. This layer will be based upon routines inside the database and Spatial Database Engine (SDE) instance. (3 months)
  5. Serve the vessel layers using Internet GIS in conjunction with additional real-time vessel traffic (lockage), river and weather data sources – additional real-time weather, river condition and vessel lockage data will be integrated with the AIS data, enabling a complete picture of the waterways and current conditions. All data sources will be spatially-enabled and served via Internet GIS software (ArcGIS Server). The map service as well as online analytics and ad-hoc query services will be password-protected. (4 months)
  6. Document results and complete final report. (2 months)

Dr. James Dobbins is the principal investigator of this project.


Try It Out

Use the link below to explore the results of the AIS Data Mining Project.

AIS Web Viewer