TacForeSight: Force-Guided Tactile World Model for Contact-Rich Manipulation

2026-06-09

AI Summary

This work addresses the challenge of enabling robots to continuously perceive and regulate physical interactions under dynamic contact transitions and complex geometric surfaces. To this end, the authors propose TacForeSight, a novel framework that explicitly models the spatiotemporally asymmetric roles of global forces and local tactile signals. TacForeSight introduces a lightweight force-guided tactile anticipation mechanism, integrating a tactile world model (TacForceWM), a predictive tactile-conditioned policy, cross-attention modules, and tactile gating units to achieve efficient multimodal fusion and short-horizon tactile state prediction within a compact latent space. Real-robot experiments demonstrate that the method significantly outperforms existing approaches across five representative tasks and three perturbation settings, exhibiting exceptional robustness and anticipatory reasoning capabilities—particularly in scenarios involving dynamic contact disturbances.

Abstract

Contact-rich manipulation requires robots to continuously perceive and regulate evolving physical interactions under dynamic contact transitions or complex surface geometries. Recent imitation learning methods improve contact-aware control by incorporating tactile or force feedback, but they rarely model the asymmetric spatiotemporal roles of global force and local tactile sensing. To address this, we propose TacForeSight, a lightweight force-conditioned tactile foresight framework for real-time manipulation. The core component is TacForceWM, a tactile world model that predicts short-horizon tactile latent dynamics from dual-finger tactile observations conditioned on high-frequency wrist force and torque signals. Another key component, the Predictive Tactile-Conditioned Policy, leverages the predicted latents as anticipatory contact priors, models the current-to-future tactile evolution via cross-attention, and adaptively fuses visuo-tactile features through a tactile-guided gating module. By forecasting purely within a compact latent space, TacForeSight enables proactive contact reasoning with efficient real-time inference suitable for high-frequency manipulation control. Real-robot experiments on five representative tasks and three in-process perturbation settings show that TacForeSight consistently outperforms existing baselines, particularly under dynamic contact disturbances. All models and datasets will be made publicly available on the project website at https://tacforesight.github.io/ProjectPage.

Problem

• contact-rich manipulation
• tactile sensing
• force feedback
• physical interaction
• asymmetric spatiotemporal roles

Innovation

• tactile world model
• force-conditioned prediction
• contact-rich manipulation
• latent dynamics forecasting
• visuo-tactile fusion