In 2016, ETH in Zurich organised the first Cybathlon: a large event where 66 international teams competed in six disciplines, all focused on developing assistive technologies suitable for everyday use with and for people with disabilities. I was part of the exoskeleton team from IHMC Robotics (FL, USA), which won the silver medal at the event.
With the team, we developed a one-off exoskeleton that allowed our pilot to perform the challenging tasks in the exoskeleton race. These tasks were designed to advance the capabilities of contemporary exoskeletons, and included sitting down and standing up, navigating a slalom, going up and down slopes and stairs, and opening and closing a door. Our exoskeleton was the first with actuated ankle joints, enabling larger steps.
I contributed to the exoskeleton's low-level actuation and sensor system, and developed some of the task-specific behaviours. Furthermore, I designed and built the human-machine interface: a combination of (electronic) hardware and software.
After finishing a master's in mechanical engineering at TU Delft, I worked at one of the fastest-growing robotics companies in the Netherlands: Fleet Robotics. They offer robotic solutions for underwater ship hull maintenance in large ports. I joined the company in the start-up phase, developing the software for their prototype hull-cleaning robot. During my time at Fleet Robotics, the company grew from a single prototype to a multi-robot, 24/7 operation with hull-cleaning and inspection robots.
I returned to academia for a PhD on interactions between autonomous vehicles and human-driven vehicles. Focusing on highway interactions such as merging, the goal of the project was to gain a better understanding of human-human interactions to improve interactions between humans and automation. This project was a collaboration between TU Delft and Nissan and was therefore an application-driven scientific project.
During the project, I developed the Communication-Enabled Interaction (CEI) framework, an alternative to the prevalent Game Theory for modelling interactions between humans. A model based on the framework was the first unified model to explain driver interactions in terms of high-level decisions, safety margins, and low-level control input. It therefore captured the underlying mechanisms that lead to eventual interaction outcomes.
Besides the modelling work, I continued developing software during my PhD. I co-developed JOAN, an open-source driving simulator for human-in-the-loop experiments with automated vehicles. And I've developed TraViA, a software package for visualising and annotation large naturalistic (i.e., recorded in the real world) traffic datasets. I was awarded my PhD cum laude in May 2024.