Magnetic Tactile-Driven Soft Actuator for Intelligent Grasping and Firmness Evaluation

TL;DR

This work tackles the lack of integrated tactile sensing in soft actuators and the distortion caused by actuator deformations. It introduces SoftMag, a magnetically sensing soft actuator co-designed with a soft pneumatic layer, and validates it with multiphysics simulations and a neural decoupling strategy to remove parasitic signals. The authors extend the actuator into a two-finger SoftMag gripper, enabling real-time tactile inference via a multi-task network and introducing a probing-based firmness estimation method tested on apricots. The results demonstrate accurate, non-destructive firmness assessment during grasping and highlight the potential of sensorized soft robotics for intelligent, material-aware manipulation and sorting tasks.

Abstract

Soft robots are powerful tools for manipulating delicate objects, yet their adoption is hindered by two gaps: the lack of integrated tactile sensing and sensor signal distortion caused by actuator deformations. This paper addresses these challenges by introducing the SoftMag actuator: a magnetic tactile-sensorized soft actuator. Unlike systems relying on attached sensors or treating sensing and actuation separately, SoftMag unifies them through a shared architecture while confronting the mechanical parasitic effect, where deformations corrupt tactile signals. A multiphysics simulation framework models this coupling, and a neural-network-based decoupling strategy removes the parasitic component, restoring sensing fidelity. Experiments including indentation, quasi-static and step actuation, and fatigue tests validate the actuator's performance and decoupling effectiveness. Building upon this foundation, the system is extended into a two-finger SoftMag gripper, where a multi-task neural network enables real-time prediction of tri-axial contact forces and position. Furthermore, a probing-based strategy estimates object firmness during grasping. Validation on apricots shows a strong correlation (Pearson r over 0.8) between gripper-estimated firmness and reference measurements, confirming the system's capability for non-destructive quality assessment. Results demonstrate that combining integrated magnetic sensing, learning-based correction, and real-time inference enables a soft robotic platform that adapts its grasp and quantifies material properties. The framework offers an approach for advancing sensorized soft actuators toward intelligent, material-aware robotics.

Figures (24)

• Figure 1: Concept of the SoftMag actuator: Sensorizing the M-PAM actuator with the SoftMag sensor through a soft, integrated design enabled by a shared porous material.
• Figure 2: Results of the multiphysics simulation validation by applying a $0-31.4$ kPa pressure to the SoftMag actuator: (a) Stress and deformation under pressure load; (b) Side view and the A-A cross-section plane; (c) Magnetic flux density map on the A-A plane; (d) $\Delta B_y$ and $\Delta B_z$ at the sensing locus; (e) Tip displacement in the Y and Z directions; (f) Bending angle of the actuator.
• Figure 3: Fabrication process of the SoftMag actuator.
• Figure 4: Indentation testing setup and protocol: (a) Overall testing setup; (b) Close-up of the indentation platform; (c) Grid map of $9×9$ indentation points used for normal indentation tests; (d) Grid map of $9×9$ indentation points used for shear indentation tests.
• Figure 5: Spatial analysis of sensing performance during the normal indentation test: (a–c) Global distributions of peaks and troughs in normal force, $B_x$, and $B_z$ magnetic flux responses; (d) Global distribution of PC1 scores from principal component analysis of magnetic flux responses; (e–g) Global slope distributions (first derivatives) of normal force, $B_x$, and $B_z$; (h) Global mean signal-to-noise ratio (SNR) map for $B_z$ magnetic flux under unloaded conditions.