A nautical chart shows what an area of sea looks like for the purposes of ensuring safe navigation. Ordinary land maps can show how a road curves and progresses, but this is not possible when it comes to the sea. It takes large quantities of precise data – which need to be processed and corrected – before it is possible to produce a nautical chart that is safe for use in navigation.
Collecting the necessary quantities of data to produce a nautical chart is a resource-intensive process. The Danish Geodata Agency is therefore constantly looking for and exploring new possibilities within data collection.
Strength in Numbers
One of the options currently being explored is known as crowd-sourced bathymetry (CSB). This involves collecting depth measurements from multiple vessels, and specifically from vessels which already operate in a given area.
Many ships are already equipped with echo sounders and it is this potential that is to be leveraged through the collaboration between the Geodata Agency and Sternula, which we have previously described, to develop Gavian, a TCSB box which collects valuable data from these echo sounders.
Danish Cooperation in a European Framework
This is where Aalborg University comes into the picture. Both Aalborg University and the Geodata Agency are a part of the EU MobiSpaces project, which places focus on developing data management platforms for mobile data that can be processed decentrally and to such an extent that it can return usable data, thus improving data processing for all.
The Department of Computer Science at Aalborg University contributes the necessary expertise within mobile analytics and big data, in particular spatio-temporal data, which is highly relevant to the focus areas of the Geodata Agency.
The Geodata Agency contributes to the project in the form of a specific case whereby a CSB box on board a ship shall collect depth measurements en route to its destination, and compare the collected data with a nautical chart stored in the box.
Artificial Intelligence Bolsters Nautical Charts
If the box discovers that there are discrepancies between what is on the nautical chart and what it finds in the real world, the Geodata Agency will need to receive this information quickly.
The box on board will be trained via artificial intelligence, or more specifically, through federated learning, to assess whether the data collected should be given a low or high priority. If the data is considered to be of a high priority, the box will automatically transmit the data to the Geodata Agency via a satellite connection. This is a relatively expense means of transmitting data, however. For that reason, the box is designed to hold off on transmitting non-critical data until a faster and cheaper connection is available in those cases where the box assesses that the data in question is of a lower priority. It is researchers at the Department of Computer Science who will develop the algorithms to ensure that data is only transmitted when it is of a high priority.
Similarly, the box will also be able to model nautical charts if it sails into an uncharted area It will simply store the recorded depths. By undertaking repeated sailings in an uncharted area, it will then be possible, in the longer term, to build up a foundation of data that is sufficiently solid to use in navigation. That is the vision for the future, in any case.
There are also potential cost savings to be made by placing such boxes on multiple vessels. This comes in part from attaining greater coverage, with data being collected considerably faster and the subsequent data processing also becoming more streamlined and efficient.
But there is also another and perhaps more tangible cost saving to be had. For obvious reasons, ships typically prefer not to sail into areas with minimal chart coverage. But if data can now be collected from multiple smaller vessels, which do not go as deep as larger ships, it can become possible to find new, shorter and faster routes. This represents a potential to make savings not only on fuel, but on time and money as well.
Project Is Well Under Way
The project was commenced on the first of September and the first results are now beginning to trickle in. And there is clear excitement about this at the Geodata Agency:
“We are very pleased with this collaboration with Aalborg University which allows us to jointly explore the use of new technologies across national borders. At the same time, the EU project is also a great opportunity for us to exchange experiences with other agencies that are faced with the same challenges as us,”
explains Niels Tvilling Larsen, Head of Office at Nautical Charts and Marine Data at the Danish Geodata Agency.
Kristian Torp, Professor with Special Responsibilities at the Department of Computer Science, Aalborg University, also feels that the project is of great value:
“The Geodata Agency provides domain-specific expertise which allows us, as researchers, both to acquire knowledge on concrete problems previously unknown to us, and to deliver research results that are useable in practice,”
The Danish Geodata Agency is also collaborating with Aalborg University on the DIREC project which aims to define a future data processing platform for the management of bathymetric and geographic marine data.
The purpose of the platform will be to support a broad variety of data sources – from traditional bathymetric depth measurement data collected via echo sounders through to newer and different technologies such as satellite data, LiDAR data and AIS ship data. The idea is that the platform will link these very different technologies together on a single platform where data can be integrated together to deliver new products for navigation, nautical charts and marine analyses with a focus on safety, sustainability and environmental footprint.