Data-based Ship Modeling for Marine Automation

Cargo ship in harbour
Photographer: Jonas Linder

How can sensor data and data-based modeling be used to make cargo ships more efficient, with the long-term goal to develop fully autonomous ships? This is the topic of an ongoing industry-oriented research project.

Marine transportation systems are vital components of the global economy and there is an increasing interest in support systems for marine automation, for example, due to new requirements concerning vessel performance, energy efficiency and operational safety. Many approaches rely on accurate information about the properties of the marine vessel and its behavior over time. Many of these properties are time-varying and a common approach to monitor the changes over time is to equip ships with various sensors.

Hence, a central problem is to extract useful information from the sensor data, including properties that cannot be measured directly during normal operation. For example, methods for system identification and sensor fusion can be used to obtain information about the loading conditions of a ship and to detect possible problems such as load shifts, damages, and suboptimal ballasting.

The research at Linköping University about these topics is carried out within the competence center LINK-SIC and in close collaboration with ABB Corporate Research in Västerås, Sweden and ABB Marine & Ports in Helsinki, Finland. The main result is a novel system identification framework that can be applied also when only a model of a smaller part of the ship dynamics is desired and when some of the important input signals are unknown. This framework can be applied in many other application areas, e.g., automotive, aerospace, and process industry.

Jonas Linder successfully defended his PhD thesis in March 2017 and he and the other researchers involved are now working on new problems that are relevant for the marine automation industry.

Contact

Staff at Automatic Control

Organisation