Particular challenges for ships with automated navigation include maneuvering in narrow channels and docking and undocking in ports and locks. While guidelines for planning a ship’s course and speed during transit through a canal or in flowing waters are relatively straightforward to define, this is no longer the case for port and lock maneuvers. A given target position, such as a pier, can usually be approached via various maneuvers. The best strategy depends not only on the ship and its current load condition but also on the current environmental conditions.

In the ELLA project, a development platform specifically designed for automated maneuvering is being developed, built, and deployed in a test field. The vessel and test field serve as a learning environment for an artificial intelligence system that, based on manually performed maneuvers and its own test runs, learns step by step to plan and execute the required maneuvers. At the end of development, the system should be capable of reaching a specified destination independently and safely. The project results can then be transferred to real inland waterway vessels, taking into account the laws of similarity: Through the scale-accurate representation of the hull shape as well as the propulsion and maneuvering systems, the handling characteristics of a typical cargo motor vessel are realistically simulated.