Cerebellar-Inspired Residual Control for Fault Recovery: From Inference-Time Adaptation to Structural Consolidation

TL;DR

This paper presents a cerebellar-inspired residual control framework that enables real-time fault recovery in robotic policies without retraining, by augmenting existing policies with online corrective mechanisms inspired by cerebellar principles.

Contribution

It introduces a novel inference-time residual control method inspired by cerebellar mechanisms, allowing fast, localized fault correction without altering the original policy parameters.

Findings

Achieves up to 66% performance improvement under faults in MuJoCo benchmarks.

Provides robust fault recovery with graceful degradation under severe disturbances.

Demonstrates effective consolidation of residual corrections into policy parameters.

Abstract

Robotic policies deployed in real-world environments often encounter post-training faults, where retraining, exploration, or system identification are impractical. We introduce an inference-time, cerebellar-inspired residual control framework that augments a frozen reinforcement learning policy with online corrective actions, enabling fault recovery without modifying base policy parameters. The framework instantiates core cerebellar principles, including high-dimensional pattern separation via fixed feature expansion, parallel microzone-style residual pathways, and local error-driven plasticity with excitatory and inhibitory eligibility traces operating at distinct time scales. These mechanisms enable fast, localized correction under post-training disturbances while avoiding destabilizing global policy updates. A conservative, performance-driven meta-adaptation regulates residual authority and plasticity, preserving nominal behavior and suppressing unnecessary intervention. Experiments on MuJoCo benchmarks under actuator, dynamic, and environmental perturbations show improvements of up to $+66\%$ on \texttt{HalfCheetah-v5} and $+53\%$ on \texttt{Humanoid-v5} under moderate faults, with graceful degradation under severe shifts and complementary robustness from consolidating persistent residual corrections into policy parameters.