Text2Touch: Tactile In-Hand Manipulation with LLM-Designed Reward Functions

Harrison Field, Max Yang, Yijiong Lin, Efi Psomopoulou, David A.W. Barton, Nathan F. Lepora

Keywords: Tactile Sensing, Reinforcement Learning, Large Language Models

TL;DR: Text2Touch employs targeted prompt engineering to steer an LLM into automatically generating reward functions that solve a challenging tactile in-hand manipulation task on real hardware, outperforming a human-engineered baseline.

Abstract: Large language models (LLMs) are beginning to automate reward design for dexterous manipulation. However, no prior work has considered tactile sensing, which is known to be critical for human-like dexterity. We present Text2Touch, bringing LLM-crafted rewards to the challenging task of multi-axis in-hand object rotation with real-world vision based tactile sensing in palm-up and palm-down configurations. Our prompt engineering strategy scales to over 70 environment variables, and sim-to-real distillation enables successful policy transfer to a tactile-enabled fully actuated four-fingered dexterous robot hand. Text2Touch significantly outperforms a carefully tuned human-engineered baseline, demonstrating superior rotation speed and stability while relying on reward functions that are an order of magnitude shorter and simpler. These results illustrate how LLM-designed rewards can significantly reduce the time from concept to deployable dexterous tactile skills, supporting more rapid and scalable multimodal robot learning.

Submission Number: 1131