HapticVLA: Contact-Rich Manipulation via Vision-Language-Action Model without Inference-Time Tactile Sensing

2026-03-16

AI Summary

This work addresses the limited generalizability of current vision-language-action (VLA) models in contact-rich manipulation tasks, which often rely on expensive and hard-to-deploy tactile sensors. To overcome this hardware dependency, the authors propose a two-stage approach that eliminates the need for tactile input during inference. First, a tactile-aware action expert is trained using safety-aware reward-weighted flow matching (SA-RWFM). Then, tactile information is distilled into compact tactile tokens via tactile distillation (TD), enabling their prediction from visual and state inputs alone; these tokens are seamlessly integrated into a standard VLA architecture. This method, for the first time, incorporates offline-learned tactile perception into VLA without requiring real-time tactile feedback. In real-world robotic experiments, it achieves an average success rate of 86.7%, outperforming baseline models that use real-time tactile feedback, thereby significantly reducing hardware constraints while maintaining both safety and generalization.

Technology Category

• Intelligent Robots: Manipulation
• Computer Vision: Language and Vision
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Tactile sensing is a crucial capability for Vision-Language-Action (VLA) architectures, as it enables dexterous and safe manipulation in contact-rich tasks. However, reliance on dedicated tactile hardware increases cost and reduces reproducibility across robotic platforms. We argue that tactile-aware manipulation can be learned offline and deployed without direct haptic feedback at inference. To this end, we present HapticVLA, which proceeds in two tightly coupled stages: Safety-Aware Reward-Weighted Flow Matching (SA-RWFM) and Tactile Distillation (TD). SA-RWFM trains a flow-matching action expert that incorporates precomputed, safety-aware tactile rewards penalizing excessive grasping force and suboptimal grasping trajectories. TD further transfers this tactile-aware capability into a conventional VLA: we distill a compact tactile token from the SA-RWFM teacher and train a student VLA to predict that token from vision and state modalities, enabling tactile-aware action generation at inference without requiring on-board tactile sensors. This design preserves contact-rich tactile-aware reasoning within VLA while removing the need for on-board tactile sensors during deployment. On real-world experiments, HapticVLA achieves a mean success rate of 86.7%, consistently outperforming baseline VLAs - including versions provided with direct tactile feedback during inference.

Problem

• tactile sensing
• Vision-Language-Action
• contact-rich manipulation
• haptic feedback
• robotic manipulation

Innovation

• Vision-Language-Action
• tactile distillation
• contact-rich manipulation
• sensor-free haptics
• flow matching