Multifingered Force-Aware Control for Humanoid Robots
Pasquale Marra, Gabriele Mario Caddeo, Ugo Pattacini, Lorenzo Natale
AI summary
Problem
Tactile sensor outputs vary widely across technologies, hindering transferable manipulation strategies, while existing multifingered control methods remain limited to grasping or single-finger feedback.
Approach
The authors train a neural network to estimate 3D contact forces from magnetic tactile sensors, then deploy a model-based controller that dynamically adjusts hand pose to align the estimated Center of Pressure with the fingertip contact centroid.
Key results
- Calibrated dataset of 3D forces and tactile outputs for custom Xela sensors
- Characterization of sensor repeatability, cross-device variability, and gravity-induced artifacts
- 82.7% success rate in balancing single objects and 80% accuracy in multi-object scenarios
- Force-domain control framework enabling sensor-agnostic in-hand object balancing
Why it matters
Enables robust, transferable tactile feedback integration for humanoid robots performing complex non-prehensile manipulation tasks.
Abstract
In this paper, we address force-aware control and force distribution in robotic platforms with multi-fingered hands. Given a target goal and force estimates from tactile sen- sors, we design a controller that adapts the motion of the torso, arm, wrist, and fingers, redistributing forces to maintain stable contact with objects of varying mass distribution or unstable contacts. To estimate forces, we collect a dataset of tactile signals and ground-truth force measurements using five Xela magnetic sensors interacting with indenters, and train force estimators. We then introduce a model-based control scheme that minimizes the distance between the Center of Pressure (CoP) and the centroid of the fingertips contact polygon. Since our method relies on estimated forces rather than raw tactile signals, it has the potential to be applied to any sensor capable of force estimation. We validate our framework on a balancing task with five objects, achieving a 82.7% success rate, and further evaluate it in multi-object scenarios, achieving 80% accuracy. Code and data can be found here https://github.com/ hsp-iit/multifingered-force-aware-control.