Analysis of the driving behavior of inland vessels using machine learning and development of a multi-level mathematical model to calculate the movement of vessels in six degrees of freedom (SixDoF+AIS)

Scientists

Kathrin Donandt, M. Sc.
Jakob Rzeszutko, M. Eng.
Dr.-Ing. Michael Schröder
Prof. Dr.-Ing. Bettar Ould el Moctar

Project Description

Due to the relatively high traffic density and the maneuvering processes on narrow waterways, it is essential to predict the maneuvering behavior of ships as precisely as possible to optimize traffic, also in connection with the development of driving assistance systems. In the project, the prognosis of the movement and the space requirements of the ships involved in a given traffic situation is to be carried out iteratively in two steps: First, the trajectories of the ships are determined with the help of an artificial intelligence trained with AIS data, for which then in a second step a maneuvering model is determined the respective lane widths of the ships can be determined.

The main project goals are:

  • Assessment of the safety and lightness of traffic situations and determination of collision-free ship trajectories considering the predicted behavior of road users involved by an artificial intelligence trained with AIS data.
  • Further development of mathematical models to simulate the movement of inland vessels during maneuvering processes in all six degrees of freedom, considering sloshing, water currents, machine dynamics and shallow water influences.