Tactile contact simulation for NVIDIA Isaac Lab. A physically scaled DEX-01 fingertip on the Tesollo DG-5F hand, with normal-force sensing mapped into the sensor's firmware frame.
A 0.1 kg cube loads finger 2. The tactile frames report physical normal force in newtons. Figure provenance.
Documentation · Installation · Watch the examples · API reference · Model behavior
- Read 738 contact samples as a 32 × 32 firmware-mapped force frame.
- Explore contact on a Tesollo hand or a flat pad, with live tactile heatmaps, remote WebRTC viewing, and camera recordings.
- Check normal-force balance with a known 0.981 N load and run geometry, contact, and mapping verification.
The model queries the counter object's PhysX signed distance field and redistributes measured normal contact load across penetrating samples, following TacSL. Output is physical normal force in newtons, verified against known loads and force balance.
Tested on Linux x86_64, Python 3.11, Isaac Sim 5.1.0, GLIBC 2.35+,
and the Isaac Lab revision pinned by setup_env.sh. A supported NVIDIA GPU
and driver are required. For Isaac Sim 5.1 camera/viewport rendering, use
NVIDIA’s validated Linux 580.65.06 (R580) driver. Install Git, Git LFS, GCC,
G++, Make, and
uv; allow tens of
gigabytes of disk space.
git clone https://github.com/analogdevicesinc/dex-01-sim-assets.git
cd dex-01-sim-assets
git lfs install
git lfs pull
./setup_env.sh
source ../env_isaaclab/bin/activate
python scripts/dex01_on_tesollo_dg5f.pyWait for Setup complete.... A gear contacts fingers 2–5 in turn while the
viewport shows their loads and tactile frames. Stop with Ctrl+C. Initial setup
may request NVIDIA's EULA, download extensions, and compile shaders.
See installation options, remote viewing, and known-load validation. The gear is not a calibration load.
Read CONTRIBUTING.md for development checks and required DCO sign-offs. Documentation authoring is covered in the authoring guide.
ADI-authored material uses Apache-2.0. Isaac Lab adaptations and Tesollo geometry retain BSD-3-Clause terms in NOTICE and LICENSE_Tesollo. Demos fetch the gear from NVIDIA's asset server under NVIDIA's terms; it is not redistributed here.
If you use this repository in your work, please cite it as:
@misc{analogdevices2026dex01,
author = {{Analog Devices, Inc.}},
title = {{DEX-01 Tactile Sensor Simulation Asset}},
year = {2026},
howpublished = {GitHub repository},
url = {https://github.com/analogdevicesinc/dex-01-sim-assets}
}Include the commit hash or release tag you used for reproducibility.
