Between October 2015 and September 2020, my PhD at AIRO / IDLab – Ghent University – imec was funded by the Human Brain Project (HBP), one of the largest European flagship research programmes. Within the HBP, I belonged to the Neurorobotics subproject (SP10), whose mission was to connect spiking neural networks and simulated brain models to physical or simulated robots in a unified software stack: the Neurorobotics Platform (NRP).
Some of my research visible on this blog — mass-spring networks, Tigrillo, HyQ locomotion control — was carried out under HBP grants. HBP gave me the chance to meet neuroscientists, roboticists and computer scientists across the continent and to initiate cross-disciplinary partnerships and discussions.
Inside the Neurorobotics Platform
The NRP is a simulation environment that connects spiking neural networks with robotic systems. It lets researchers pair different brain models with simulated robots running on high-performance computing clusters, so that embodied experiments can actually be conducted at the scale the HBP aspired to. Within my PhD, I:
- Contributed to the NRP open-source codebase (Python / C++ / ROS), particularly on the robotic side — MuJoCo / Gazebo integration, brain-to-robot coupling, transfer-learning utilities.
- Deployed the Tigrillo compliant quadruped on the NRP as a reference platform for closed-loop experiments with spiking CPGs.
- Acted as a practical “user” during integration phases, reporting bugs, shaping the Python user API, and co-running tutorials with the core platform team (TUM, Fortiss, SSSA, KTH).
The platform has evolved since then and is now part of the EBRAINS infrastructure.
Education, student representation, community
Research consortiums of that size live and die on their education and community programmes. I was elected Student Representative at the HBP Education Programme Committee (a mandate that ran until May 2019), which meant representing ~150 HBP-funded PhDs across Europe in strategic discussions about schools, grants and workshops.
On top of that role, across five years I organized, hosted and contributed to the events below. They are pulled verbatim from my thesis and give a fair picture of what being inside HBP actually looked like:
- Jan 2020, Pisa (IT) — Scientific chair of the 4th HBP Student Conference.
- Jul 2019, Belgrade (RS) — Chair and committee member of the HBP Young Researcher Event.
- Feb 2019, Ghent (BE) — Host and scientific chair of the 3rd HBP Student Conference.
- May 2018, Stockholm (SE) — Technical demonstration and lightning talk at the HBP SGA1 European Review.
- Apr 2018, Tokyo (JP) — Presentation at the Europe–Japan Neurorobotics Workshop.
- Nov 2017, Brussels (BE) — Demonstration at the European Robotics Week 2017.
- Sep 2017, Ghent (BE) — Introduction speech and host of the HBP SP10 Trimestrial Meeting.
- Jun 2017, Obergurgl (AT) — Lightning talk and poster at the 4th HBP School – Future of Artificial Intelligence.
- Apr 2017, Munich (DE) — Lightning talk at the 1st Neurorobotics Platform User Workshop.
- Dec 2016, Obergurgl (AT) — Lightning talk and poster at the 3rd HBP School – Future of Neuroscience.
- Jul 2016, Ghent (BE) — PhD vulgarisation at the Let’s Talk Science! Communication Summer School.
- Jan 2016, Manchester (UK) — Lightning talk and poster at the 4th HBP Education Workshop – Future of Computing.
Each of these events was an opportunity to present technical progress, but also — and more importantly — to negotiate with neuroscientists whether what we roboticists were doing actually deserved the adjective “brain-inspired”. The discipline that imposed on me — proving I had something worth telling a neuroscientist, and learning to listen — is probably the single most useful side-product of those five years.
What it taught me
- Shipping research code in a consortium. The NRP had to run for a neuroscientist who had never touched ROS, on a cluster, with a brain model that her colleague wrote. Releasing code that survives that context is a different craft than releasing code for yourself.
- Running events. Hosting student conferences for ~60 participants taught me logistics, budgeting, speaker wrangling, and how to get academics to actually stop talking at Q&A time.
- Translating across disciplines. Robotics, machine learning, neuroscience, hardware, simulation — being the glue between them is often more valuable than being the best at any one of them.
Going further
- Neurorobotics Platform homepage: neurorobotics.net.
- EBRAINS — where HBP tools live today: ebrains.eu.
- My PhD thesis: Biologically inspired locomotion of compliant robots, UGent, 2021.