FTP-1: A Generalist Foundation Tactile Policy Across Tactile Sensors for Contact-Rich Manipulation

2026-06-11

AI Summary

This work addresses the limited generalizability of existing tactile policies across heterogeneous sensors due to hardware-specific constraints. The authors propose the first universal tactile foundation model, which employs heterogeneous encoders to map multimodal tactile signals into a unified morphology-aware latent token space and leverages a shared tactile Transformer for joint modeling. This approach enables cross-sensor and cross-robot policy transfer and, for the first time, supports diverse tactile input modalities while achieving effective generalization to unseen sensors. Experimental results demonstrate a 17.2% improvement in task success rate on seen sensors and a substantial 31% performance gain on two previously unencountered sensor types, establishing a unified benchmark for tactile manipulation.

Abstract

Despite the success of vision-based generalist robotic policies, existing tactile-based policies remain tied to fixed embodiments and sensor setups. This is because tactile signals are highly heterogeneous across hardware, making cross-sensor generalization difficult. We present FTP-1,the first generalist foundation tactile policy pretrained to acquire transferable tactile manipulation abilities across diverse sensors and embodiments. FTP-1 supports varied tactile inputs, including image-, array-, and state-based signals, by using heterogeneous encoders to project them into unified morphology-aware latent tokens that are jointly modeled by a shared tactile Transformer expert. Pretrained on around 3,000 hours of tactile manipulation data aggregated from 26 data sources, spanning human and robot demonstrations across 21 sensors, FTP-1 learns tactile skills that transfer beyond the sensors seen during pretraining. Across downstream finetuning experiments spanning 5 hardware configurations, FTP-1 improves contact-rich manipulation on seen sensor setups by +17.2% and, surprisingly, transfers to two previously unseen tactile-sensor setups, achieving a +31% gain in success rate. FTP-1 establishes the first unified foundation baseline for tactile manipulation, providing future tactile policies with a shared model-level starting point. Pretrained models, datasets, training code and more visualization at https://ftp1-policy.github.io.

Problem

• tactile policy
• cross-sensor generalization
• contact-rich manipulation
• heterogeneous tactile signals
• embodiment transfer

Innovation

• foundation tactile policy
• cross-sensor generalization
• heterogeneous tactile encoding
• morphology-aware representation
• tactile Transformer