ETac: A Lightweight and Efficient Tactile Simulation Framework for Learning Dexterous Manipulation
Zhe Xu, Feiyu Zhao, Xiyan Huang, Chenxi Xiao
AI summary
Problem
Learning tactile-based manipulation policies via reinforcement learning is hindered by the trade-off between simulation fidelity and computational cost, as existing soft-body simulators are either too slow for large-scale training or lack realistic strain propagation modeling.
Approach
ETac uses a lightweight, data-driven hybrid model combining exponential spatial decay with a residual neural network to predict elastomer surface deformations, calibrated against high-fidelity FEM simulations.
Key results
- Achieves FEM-comparable deformation accuracy (0.058 mm RMSE on flat surfaces)
- Supports 4,096 parallel RL environments at 869 FPS on a single GPU
- Trains blind grasping policies with 84.45% success rate across diverse objects
- Provides a scalable, differentiable backend bridging real tactile sensors and RL
Why it matters
It removes the computational bottleneck in tactile robot learning, enabling scalable, high-fidelity policy training for dexterous manipulation in embodied AI and robotics research.
Abstract
Tactile sensors are increasingly integrated into dexterous robotic manipulators to enhance contact perception. However, learning manipulation policies that rely on tactile sens- ing remains challenging, primarily due to the trade-off between fidelity and computational cost of soft-body simulations. To address this, we present ETac, a tactile simulation framework that models elastomeric soft-body interactions with both high fidelity and efficiency. ETac employs a lightweight data-driven deformation propagation model to capture soft-body contact dynamics, achieving high simulation quality and boosting ef- ficiency that enables large-scale policy training. When serving as the simulation backend, ETac produces surface deformation estimates comparable to FEM and demonstrates applicability for modeling real tactile sensors. Then, we showcase its capabil- ity in training a blind grasping policy that leverages large-area tactile feedback to manipulate diverse objects. Running on a single RTX 4090 GPU, ETac supports reinforcement learning across 4,096 parallel environments, achieving a total throughput of 869 FPS. The resulting policy reaches an average success rate of 84.45% across four object types, underscoring ETac’s potential to make tactile-based skill learning both efficient and scalable.