Learning-Based Robust Control: Unifying Exploration and Distributional Robustness for Reliable Robotics via Free Energy

TL;DR

This paper introduces a novel learning-based control framework inspired by neuroscience, combining exploration and robustness to uncertainties, validated through simulations and real-world robotic manipulation tasks.

Contribution

It unifies exploration and distributional robustness in robotic control using a free energy principle, enabling reliable policies with minimal fine-tuning.

Findings

Improves robustness to epistemic uncertainties in control policies.

Reduces sim-to-real gap in robotic manipulation.

Enables zero-shot deployment on real robots.

Abstract

A key challenge towards reliable robotic control is devising computational models that can both learn policies and guarantee robustness when deployed in the field. Inspired by the free energy principle in computational neuroscience, to address these challenges, we propose a model for policy computation that jointly learns environment dynamics and rewards, while ensuring robustness to epistemic uncertainties. Expounding a distributionally robust free energy principle, we propose a modification to the maximum diffusion learning framework. After explicitly characterizing robustness of our policies to epistemic uncertainties in both environment and reward, we validate their effectiveness on continuous-control benchmarks, via both simulations and real-world experiments involving manipulation with a Franka Research~3 arm. Across simulation and zero-shot deployment, our approach narrows the sim-to-real gap, and enables repeatable tabletop manipulation without task-specific fine-tuning.