Unified tactile sensing model improves robot dexterous manipulation
Learning Heterogeneous Tactile Representation for General Dexterous Manipulation
Robotics
Summary
Tactile sensing helps robots understand touch when they handle objects, but different sensors work differently and have unique limits. The authors developed LingBot-Touch, a model that learns to combine and understand touch data from multiple sensor types into one common representation. This lets robots use different kinds of tactile sensors together more effectively, improving their ability to manipulate objects with skill. The approach was tested both in simulations and on real robots, showing better performance and efficiency.
What this means in practice
- •For robotics engineers: Use a unified tactile representation to improve robot hand control across various tactile sensors, enabling more precise manipulation tasks.
- •For industrial automation teams: Integrate heterogeneous tactile data in robotic systems for efficient and adaptable handling of complex objects on production lines.
Authors
Bing Han, Yichuan Peng, Haoyu Dong, Yuanmeng Hu, Yongyi Jia, Fangjing Wang, Kecheng Zheng, Yiming Zhang, Yanhong Zeng, Xiaoqing Zhang, Chen Song, Rundong Zou, Anqi Dong, Yuan Guan, Ka Leong Cheng, Xing Zhu, Yujun Shen, Nan Xue, Wei Ning
Abstract
Tactile sensing is an indispensable modality for contact-rich dexterous manipulation, a capability that hints at the ultimate frontiers of embodied intelligence. However, commercially available tactile sensors are still at a relatively early stage of development, with each type of sensor offering irreplaceable strengths alongside its own drawbacks. A key challenge is thus to free tactile understanding from sensor-specific physics, so that embodied models can effectively leverage heterogeneous tactile data. To this end, we develop LingBot-Touch, a tactile foundation model characterized by the following properties. (i) Our model yields a unified representation across different tactile sensors (i.e., piezoresistive, magnetic, and visuotactile sensors), and is easily extensible to new sensor types. (ii) The resulting tactile tokens are plug-and-play: they can be seamlessly integrated into policy models across paradigms, from diffusion policy to vision-language-action (VLA) models such as $π_{0.5}$ and LingBot-VLA, consistently enhancing manipulation dexterity. (iii) Extensive experiments, spanning simulation and real-robot deployments across multiple hardware platforms, demonstrate that our tactile representation substantially benefits complex manipulation tasks while maintaining strong computational efficiency. We believe the findings in this work demonstrate the potential of leveraging cross-sensor data to scale tactile learning, and offer encouraging insights for the design of future hardware and embodied models.