ESR 1 – Online identification of vessel dynamics in confined inland waterways

According to the European Commission, freight transport is projected to increase 60% by 2050, and this certainly demands the transportation system to enhance its capacity. Among all transportation modes, inland waterway transportation is characterized by reliability, energy efficiency and a capacity for increased use. Not to mention that more than 37,000 km of waterways connect hundreds of cities and industrial regions in Europe, there is great potential lies in inland waterway transportation. To fully exploit the capacity of inland waterways, we must make it economically competitive. With crew cost accounting for up to 60% of the total cost, autonomous inland vessels represent an exciting disruptive technology that can greatly cut the cost of inland waterway transportation. To the realization of autonomous vessels, ship maneuverability plays an important role. It can be represented with a mathematical model, and based on the model, people can understand how forces are acting on a vessel when it is steered and how a vessel will respond to the control command. Controllers can be developed to navigate the vessel autonomously then. In a word, building a suitable ship maneuvering model is of great significance for the follow-up development of autonomous ship. Inland vessels operate in a rapidly-changing environment, with restrictions of confined waterways and interactions with other vessels. Ship dynamics in the model are not fixed values, but change remarkably under the effect of its surrounding environment. In this case, efforts must be paid to equip the ship maneuvering model continuously adapting ability to fit the drift of ship hydrodynamic characteristics. The objective of my PhD is to build an online updated maneuvering model for inland vessels using sensor data gathered during operation. In order to achieve this goal, the research plan is as following: • Select  a parameterized ship maneuvering model structure suitable for inland vessels by taking into account an optimal balance between complexity and accuracy. • Develop an offline algorithm to identify the ship dynamics from historical sensor data. – Design the inputs and outputs of the identification algorithm, decide what sensor data needs to be collected. Check the validity of the proposed ship maneuvering model structure. • Develop an online algorithm to identify the chosen ship dynamics from sensor data gathered during operation. – Equip the developed offline algorithm with online updating ability.