UBC TrajReducer AI Solves Critical 30% Vessel Tracking Data Gap Problem

University of British Columbia Okanagan research team has developed TrajReducer, an innovative AI system published in June 2025's Ocean Engineering that accurately predicts vessel destinations and arrival times even when 30% of AIS data is missing, addressing critical supply chain visibility challenges exposed by pandemic disruptions, geopolitical tensions, and incidents like the Suez Canal blockage. Traditional vessel destination prediction methods have been slow and often inaccurate, with approximately 30% of AIS data omitting estimated departure and arrival times, creating blind spots in global maritime logistics. Technical innovation: TrajReducer employs cross-dimension indexing and advanced machine learning to analyze vessel trajectory patterns, enabling accurate predictions from incomplete data. The system represents significant advancement in maritime analytics with applications extending beyond port operations to maritime safety, environmental monitoring, and supply chain optimization. Commercial applications: Canadian ports gain tools to anticipate and prepare for vessel arrivals accurately, positioning them to maintain goods flow even during unexpected global trade challenges. The 80%+ of global trade transported by maritime shipping depends on reliable arrival predictions for efficient operations. Operational resilience: "This is about more than efficiency—it's about resilience," explains lead researcher Professor Chengkai Zhang. "When Canadian ports can anticipate and prepare for the arrival of ships accurately, they're better positioned to maintain the flow of goods that Canadians depend on, even when global trade faces unexpected challenges." The TrajReducer framework provides maritime industry with reliable prediction tools that function effectively despite data gaps that have plagued traditional systems, creating foundation for more resilient global supply chains.
[TechXplore/UBC]

This story is part of the Maritime AI Digest of 5 October 2025.