< Back to previous page


Perception for Autonomous Inland Waterways Navigation

Shipping on inland waterways can be a critical contributor in handling increasing transportation volumes. However, it is faced with a lack of ship crews, which could be alleviated by increased vessel automation, or deployment of unmanned ships. To realise autonomous shipping, prior conceptual research suggests the combined use of several sensor types. While conceptually similar to autonomous road vehicles, unmanned shipping differs in key aspects, such as higher vehicle inertia and lesser manoeuvrability or the resulting need for object detection and identification at greater distances. However, established navigational tools beyond GNSS and radar are available, such as AIS and IENC. While these systems are robust, they have relatively low positional accuracy, and IENC maps are updated only in large time intervals. Aboard unmanned vessels, LIDAR and stereo cameras can be used to complement these tools. They offer greater resolution for object detection, but also require calibration. The purpose of this project is to develop such a calibration procedure for 3D-LIDAR and stereo camera aboard an autonomous ship, as well as to develop object detection and identification algorithms compatible with ECDIS. Further, sensor fusion between available GNSS, AIS, IENC and Radar data (ECDIS map data) and the stereo camera and 3D-LIDAR aboard the vessel is to be implemented and used for ECDIS map updates. Sensor calibration, object detection and sensor fusion are additionally to be validated experimentally on a scale model vessel.

Date:27 Jul 2020 →  14 Dec 2021
Keywords:Robotics, Automation, Perception, Object Detection, Navigation
Disciplines:Field and service robotics, Sensing, estimation and actuating
Project type:PhD project