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

arXiv 2026 50.1 method, application

TLDR

A force-conditioned tactile world model for real-time contact-rich manipulation, predicting tactile latent dynamics and outperforming baselines in real-robot experiments.

Reasoning

The paper presents a novel tactile world model that effectively integrates global force and local tactile sensing, with strong empirical validation on real robots across multiple tasks. However, the abstract lacks discussion of limitations or failure cases, and the approach is not explicitly compared to model-based RL methods.

Read-first score

Read-first score 50.1, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 32.

Recency 5%

100

Uses a gentle age decay so recent papers surface without erasing older foundations. 2026

Citation impact 15%

Uses OpenAlex-shaped citation metadata as a bibliometric attention signal, separate from paper quality. citation_normalized_percentile=0.92958021

Methodology quality 15%

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=baseline,dataset,experiment,metric

Reproducibility 15%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=dataset,github

Topical relevance 25%

45.7

Uses existing LLM keyword relevance scores normalized to 0-100. world model,world simulator,generative world model,interactive world model,video world model,world dynamics prediction,model-based reinforcement learning world model

Graph prestige 15%

5.7

PageRank-style prestige over the local citation graph built from available OpenAlex references. local_pagerank=0.0016

Citation velocity 10%

Citation velocity estimates citations per publication-year to reduce old-paper bias. velocity=0.00

Field roles

FoundationFrontierBridgeMethodology anchor

Rank sensitivity

Stability: volatile; rank range: 332.

Keyword Scores

world model

world dynamics prediction

generative world model

interactive world model

model-based reinforcement learning world model

world simulator

video world model

Deep Analysis

Innovations

Force-conditioned tactile world model (TacForceWM) that predicts short-horizon tactile latent dynamics from dual-finger tactile observations conditioned on high-frequency wrist force and torque signals.
Predictive Tactile-Conditioned Policy that leverages predicted latents as anticipatory contact priors, models current-to-future tactile evolution via cross-attention, and adaptively fuses visuo-tactile features through a tactile-guided gating module.
Lightweight latent space forecasting enabling real-time inference for high-frequency manipulation control.

Methodology

TacForeSight comprises two components: TacForceWM, a tactile world model that predicts future tactile latent states conditioned on wrist force/torque, and a Predictive Tactile-Conditioned Policy that uses these predictions as anticipatory priors via cross-attention and tactile-guided gating. The framework operates in a compact latent space for efficient real-time inference. Real-robot experiments were conducted on five representative contact-rich manipulation tasks and three in-process perturbation settings, comparing against existing baselines.

Key Results

TacForeSight consistently outperforms existing baselines across five tasks and three perturbation settings, particularly under dynamic contact disturbances.

Resources

Paper