GenTact Toolbox: A Computational Design Pipeline to Procedurally Generate Context-Driven 3D Printed Whole-Body Artificial Skins

TL;DR

This work tackles the lack of context-aware, form-fitting tactile skins for robots by introducing GenTact Toolbox, a three-stage pipeline that combines procedural generation, task-driven simulation, and multi-material 3D printing to autonomously design capacitive tactile skins tailored to a robot’s geometry and application. The approach yields skin units that conform to specific embodiments, optimize sensor placement via a density heat map derived from simulated contact data, and fabricate functional skins through 3D printing with conductive nodules and RC delay sensing. Key contributions include an open-source pipeline, Blender-based procedural skin generation, a simulation-driven optimization loop, and real-world validation on a Franka FR3 arm for a human-robot interaction scenario, with demonstrations on additional platforms to show generality. The work advances rapid customization and deployment of whole-body tactile skins across diverse robots and tasks, enabling safer and more effective physical interaction in unstructured environments.

Abstract

Developing whole-body tactile skins for robots remains a challenging task, as existing solutions often prioritize modular, one-size-fits-all designs, which, while versatile, fail to account for the robot's specific shape and the unique demands of its operational context. In this work, we introduce GenTact Toolbox, a computational pipeline for creating versatile whole-body tactile skins tailored to both robot shape and application domain. Our method includes procedural mesh generation for conforming to a robot's topology, task-driven simulation to refine sensor distribution, and multi-material 3D printing for shape-agnostic fabrication. We validate our approach by creating and deploying six capacitive sensing skins on a Franka Research 3 robot arm in a human-robot interaction scenario. This work represents a shift from "one-size-fits-all" tactile sensors toward context-driven, highly adaptable designs that can be customized for a wide range of robotic systems and applications. The project website is available at https://hiro-group.ronc.one/gentacttoolbox

Figures (6)

• Figure 1: The computational pipeline presented in this work, GenTact Toolbox, generates form-fitting and adaptable whole-body tactile sensors. GenTact Toolbox uses the 3D model of a given robot (a) and a user-generated heat map (b) to create digital meshes of sensor arrays (b) that can be 3D printed as functional tactile sensors (c).
• Figure 2: The GenTact pipeline for designing form-fitting and adaptable tactile skins is composed of three stages: procedural generation, simulation, and fabrication. The procedural generation stage (left) generates an initial distribution of sensors that are then passed into the simulation stage (bottom right) to be evaluated and improved based on the task they are used for. Finally, the sensors are connected internally in the fabrication stage (top right) to be printed and deployed on the real robot.
• Figure 3: Snapshots of link 5 for the FR3 as it goes through the digital skin generation and fabrication stages of the design pipeline.
• Figure 4: Top: Sensor distribution is streamlined and configurable using various scalar parameters such as the cutoff tolerance, fill tolerance, and minimum distribution distance. Bottom: The original skin heat map can produce jagged edges that require smoothing.
• Figure 5: In this example scenario, a Unitree H1 humanoid robot is covered with unoptimized skin units and tasked to move storage bins in simulation. At the end of the simulation, four concentrated regions of high contact were identified on the chest plate. The heuristic described by \ref{['eq:BW']} uses the contact points to generate a new density heat map that can be tuned by $\alpha$ and $n$. The resulting optimized configuration has significantly fewer sensors in less critical regions while maintaining a high density near the more likely contact areas.