Bridging Academia and Industry: My Secondment at RH Marine

One of the most valuable parts of being a MSCA fellow working with the AUTOBarge project is the opportunity of secondments at companies: short-term stays at one of our partner organizations. It’s our chance to step out of the academic world and see how research meets reality. Earlier this summer, I spent two months at RH Marine in the Netherlands, a leading innovator in maritime electrical and automation systems. My goal was to bridge the gap between my PhD research and the cutting-edge technology being used on ships today.

Figure 1: The working environment in RH Marine

The Challenge: When a Ship’s “Eyes” Can Fail

In my research, I focus on improving how autonomous vessels “see” and perceive the world around them. Cameras are fantastic sensors—they provide rich, detailed images of the environment. But what happens when it’s foggy, rainy, or dark? Just like our own eyes, camera performance can degrade significantly in bad weather or due to sensor noise. For an autonomous ship, being unable to reliably see and identify other vessels is simply not an option.

My work during the secondment tackled this exact problem. Don’t rely on just one pair of eyes. Instead, we can fuse camera data with information from other sensors that are robust under poor visual conditions.

My Mission: Building a More Robust Perception System

My time at RH Marine was an intensive period of learning and development, focused on a few key activities:

• Hands-On with Industry-Grade Simulators: The first thing I did was dive into RH Marine’s advanced ship simulators and Electronic Chart Display and Information Systems (ECDIS). Getting hands-on experience with the systems that are used by captains today was invaluable. It gave me a clear understanding of the practical challenges and what the industry truly needs from new technology.

Figure 2: Getting hands-on experience with RH Marine’s advanced maritime simulation platform

• A New Way to Fuse Sensors: My main project, developed in collaboration with my colleague Yunjia Wang (ESR4), was to design a new framework for combining camera images with Automatic Identification System (AIS) data. To train and test this new AI model, we needed the right data. A part of my secondment was dedicated to creating a new dataset called FVessel-S. I took an existing public dataset and enhanced it by adding high-quality segmentation masks—creating precise, pixel-level outlines of every ship in the images. This foundational work is crucial for teaching our model to be as precise as possible.

Figure 3: An example from the FVessel-S dataset, showing a ship with the instance segmentation mask

Gaining New Perspectives

Beyond my own project, the best part of the secondment was the exchange of knowledge. I had fascinating discussions with the experts at RH Marine, who shared insights from their own R&D. Learning about their advanced work, such as how they combine camera and radar data to estimate a ship’s 3D pose, was incredibly enlightening. It showed me how industry leaders are tackling similar problems and gave me a much broader perspective that will surely influence the next steps of my PhD.

A Bridge Between Two Worlds

My two months at RH Marine were a fantastic success. I returned not only with a novel framework for my research but also with a deeper appreciation for the practical challenges and innovative spirit of the maritime industry. This direct industry perspective is now guiding my PhD work, ensuring that my research is aimed at solving tangible, real-world problems.

A huge thank you to my supervisors at RH Marine, Ehab el Amam and Tianlei Miao, for their exceptional guidance and for making this secondment such a rewarding experience!

An article by Zhongbi Luo