Abstract

In maritime logistics, ML can support operations and strategic decision making by turning data into insights. Several use cases from research projects are presented, covering improved yard management in intermodal terminals, extracting traffic networks as graph data from ship position data, associating ship position data with image data to enable autonomous shipping, and transcribing marine radio communication. In each case, the domain problem is in the center of modelling and more or less sophisticated off-the-shelf ML methods are used as one component of the proposed solution. Insights from the development process are shared in three separate talks.