Fault-tolerant Control of Robot Manipulators with Sensory Faults using Unbiased Active Inference

TL;DR

This paper introduces a novel fault-tolerant control method for robot manipulators using unbiased active inference, enabling robust fault detection and isolation without extra controllers, validated through simulation.

Contribution

It proposes a new formulation of active inference that provides unbiased state estimation and simplifies fault detection thresholds in robotic systems.

Findings

Effective fault detection in simulated 2-DOF manipulator

Unbiased state estimation improves fault tolerance

No additional controllers needed for fault recovery

Abstract

This work presents a novel fault-tolerant control scheme based on active inference. Specifically, a new formulation of active inference which, unlike previous solutions, provides unbiased state estimation and simplifies the definition of probabilistically robust thresholds for fault-tolerant control of robotic systems using the free-energy. The proposed solution makes use of the sensory prediction errors in the free-energy for the generation of residuals and thresholds for fault detection and isolation of sensory faults, and it does not require additional controllers for fault recovery. Results validating the benefits in a simulated 2-DOF manipulator are presented, and future directions to improve the current fault recovery approach are discussed.