24th European Control Conference · Reykjavík, Iceland

Free Energy Principle
for Control

From Neural Dynamics to Robotic Embodiments

◆ 7–10 July 2026 ◆ Half-Day Workshop ◆ University of Iceland
Register for ECC 2026
About the Workshop

Bridging Neuroscience, Control Theory & Robotics

The Free Energy Principle (FEP) from computational neuroscience suggests that natural agents make decisions by creating a model of the world and minimizing surprise between data received from the environment and this internal model. It provides a unified framework for perception, action, and learning by minimizing variational free energy.

This half-day workshop explores the application of the FEP and active inference to control, emphasizing the connection between neural dynamics and robotic embodiments. We investigate how FEP-based controllers can handle uncertainty, enable compliance and safe interaction, and learn from demonstrations.

Participants will gain both conceptual understanding and practical insights for implementing FEP controllers on real control systems. No prior knowledge of the Free Energy Principle is assumed.

Key Topics
● Variational inference for state estimation
● Active inference for motion planning
● Distributionally robust formulations
● Impedance control for human-robot interaction
● Neural dynamics & biologically plausible networks
● Connections to MPC & inverse optimal control
Invited Speakers & Talks

Five Perspectives on the FEP

Programme

Workshop Schedule

Half-day workshop (approximately 4 hours). The schedule may be adapted to morning or afternoon slots.

Workshop Organizers

Meet the Team

Call for Spotlight Presentations

Students & Early-Career Researchers

We invite graduate students and early-career researchers to submit abstracts for spotlight presentations during the concluding session of the workshop (12:30–13:00). This is an excellent opportunity to present your work on topics related to the Free Energy Principle, active inference, neural dynamics, or their applications to robotics and control.

Submission format: Extended abstract (1–2 pages) describing the contribution, including preliminary results where applicable.

Topics of interest include: Variational inference for state estimation, active inference for planning, FEP-based robot control, neural circuit models, distributionally robust control, impedance control under uncertainty, learning from demonstration via FEP, and connections between FEP and established control paradigms.

Submission deadline: TBA — Please check back or contact the organizers for updates.

Submit Your Abstract Contact Organizers
Registration

Join Us in Reykjavík

Registration for the workshop is handled through the main ECC 2026 conference registration system. We expect 40–60 attendees.

Register at ECC 2026 ↗