Tactile-based Active Inference for Force-Controlled Peg-in-Hole Insertions

TL;DR

This paper introduces a dual-policy approach combining reinforcement learning and active inference with tactile feedback to improve force-controlled peg-in-hole insertions, achieving high success rates even with tilted pegs and minimal clearance.

Contribution

It presents a novel dual-policy architecture that leverages tactile feedback and active inference for robust peg alignment without extensive training datasets.

Findings

Achieved 90% success rate in real-world peg insertions with less than 0.1 mm clearance.

Outperformed previous methods lacking tactile feedback by a significant margin.

Demonstrated effective generalization across five different peg objects.

Abstract

Reinforcement Learning (RL) has shown great promise for efficiently learning force control policies in peg-in-hole tasks. However, robots often face difficulties due to visual occlusions by the gripper and uncertainties in the initial grasping pose of the peg. These challenges often restrict force-controlled insertion policies to situations where the peg is rigidly fixed to the end-effector. While vision-based tactile sensors offer rich tactile feedback that could potentially address these issues, utilizing them to learn effective tactile policies is both computationally intensive and difficult to generalize. In this paper, we propose a robust tactile insertion policy that can align the tilted peg with the hole using active inference, without the need for extensive training on large datasets. Our approach employs a dual-policy architecture: one policy focuses on insertion, integrating force control and RL to guide the object into the hole, while the other policy performs active inference based on tactile feedback to align the tilted peg with the hole. In real-world experiments, our dual-policy architecture achieved 90% success rate into a hole with a clearance of less than 0.1 mm, significantly outperforming previous methods that lack tactile sensory feedback (5%). To assess the generalizability of our alignment policy, we conducted experiments with five different pegs, demonstrating its effective adaptation to multiple objects.