Learning Stable Robot Grasping with Transformer-based Tactile Control Policies

TL;DR

This paper introduces a Transformer-based reinforcement learning approach for stable robot grasping that optimizes re-grasp location and gripping force for objects with unknown and moving centers of gravity, demonstrating success in simulation and real-world transfer.

Contribution

It presents a novel end-to-end Transformer-based framework for joint optimization of re-grasp location and force in a model-free setting, extending stable grasp tasks to more dynamic scenarios.

Findings

Successfully trained in simulation and transferred to real-world

Optimized both re-grasp location and gripping force

Analyzed model performance for dual-objective optimization

Abstract

Measuring grasp stability is an important skill for dexterous robot manipulation tasks, which can be inferred from haptic information with a tactile sensor. Control policies have to detect rotational displacement and slippage from tactile feedback, and determine a re-grasp strategy in term of location and force. Classic stable grasp task only trains control policies to solve for re-grasp location with objects of fixed center of gravity. In this work, we propose a revamped version of stable grasp task that optimises both re-grasp location and gripping force for objects with unknown and moving center of gravity. We tackle this task with a model-free, end-to-end Transformer-based reinforcement learning framework. We show that our approach is able to solve both objectives after training in both simulation and in a real-world setup with zero-shot transfer. We also provide performance analysis of different models to understand the dynamics of optimizing two opposing objectives.