Source-linked AI summary
A Flexible and Robust Large Scale Capacitive Tactile System for Robots
Perla Maiolino, Marco Maggiali, Giorgio Cannata, Giorgio Metta, Lorenzo Natale
TL;DR
Elastomer-based capacitive skins complicate production and can suffer hysteresis, ageing, and temperature drift. This paper replaces the elastomer with a fabric dielectric sandwich and adds thermal-reference capacitors. The resulting system shows good sensing performance, small hysteresis, and effective temperature-drift compensation.
Problem
Elastomer-based capacitive tactile skins can complicate production and exhibit hysteresis, ageing-related degradation, and temperature drift.
Method
The paper develops a modular capacitive skin using a deformable 3D-fabric, conductive-Lycra, and protective-layer dielectric sandwich with pressure-insensitive thermal-reference capacitors.
Results
The sensor demonstrates good sensitivity and resolution, very low hysteresis, and effective compensation of temperature-induced drift.
Takeaways & Limitations
The fabric-based dielectric simplifies production and improves mechanical properties while supporting thermal compensation in large-scale robot tactile sensing.
Abstract
from arXiv · showhide
Capacitive technology allows building sensors that are small, compact and have high sensitivity. For this reason it has been widely adopted in robotics. In a previous work we presented a compliant skin system based on capacitive technology consisting of triangular modules interconnected to form a system of sensors that can be deployed on non-flat surfaces. This solution has been successfully adopted to cover various humanoid robots. The main limitation of this and all the approaches based on capacitive technology is that they require to embed a deformable dielectric layer (usually made using an elastomer) covered by a conductive layer. This complicates the production process considerably, introduces hysteresis and limits the durability of the sensors due to ageing and mechanical stress. In this paper we describe a novel solution in which the dielectric is made using a thin layer of 3D fabric which is glued to conductive and protective layers using techniques adopted in the clothing industry. As such, the sensor is easier to produce and has better mechanical properties. Furthermore, the sensor proposed in this paper embeds transducers for thermal compensation of the pressure measurements. We report experimental analysis that demonstrates that the sensor has good properties in terms of sensitivity and resolution. Remarkably we show that the sensor has very low hysteresis and effectively allows compensating drifts due to temperature variations.
I. INTRODUCTION
The paper addresses the difficulty of integrating large-area tactile sensors onto curved robotic surfaces and the limitations of elastomer-based capacitive skins. It introduces a revised capacitive skin using a 3D-fabric dielectric and thermal compensation.
- Large-area tactile sensing still requires flexible deployment on curved surfaces while addressing wiring, networking, power, cost, integration, and maintenance.
- Capacitive sensing offers good sensitivity and resolution but commonly requires deformable dielectric and conductive layers that can introduce hysteresis, creep, electromagnetic interference, and thermal sensitivity.
- The previous modular skin covered multiple robots with up to approximately 2000 sensing units but suffered hysteresis, ageing-related sensitivity loss, and poor wear resistance.
- The revised system replaces the elastomer with deformable 3D fabric plus conductive and protective layers, and embeds pressure-insensitive capacitors for temperature compensation.
- Experiments report very low hysteresis, high sensitivity and resolution, and effective compensation of temperature-induced drift.
II. STATE OF THE ART
The related work reviews several large-scale tactile systems using conductive fabric, combined force and pressure sensing, and interconnected modular robot skins.
- Existing large-scale tactile systems include conductive-fabric pressure-switch matrices, covers combining resistive and force sensors, and conformable interconnected modules.
III. ROBOT SKIN TACTILE SYSTEM
The tactile system uses flexible triangular FPCB modules with capacitive pressure sensing, temperature-reference taxels, and a fabric-based dielectric sandwich. The design supports modular coverage while simplifying production and improving robustness.
- Each triangular FPCB module uses a deformable dielectric and conductive ground layer to form capacitive pressure sensors.
- The flexible FPCB contains 12 taxels and a capacitance-to-digital converter that measures, converts, and serially transmits each sensor value.
- Several flexible triangular modules can be interconnected into a mesh covering the desired area and conforming to non-flat surfaces.
- Two FPCB-embedded dummy taxels are pressure-insensitive because the FPCB is bonded to a rigid support, so their temperature-dependent readings compensate thermal drift.
- The dielectric replaces silicone foam and conductive Lycra with a clothing-industry fabric sandwich comprising 3D fabric, Lycra, and protective fabric.
A. Integration of the capacitive skin on the iCub
The revised skin is integrated onto curved robot parts by separately attaching the sensor mesh and a preformed, screw-fixed fabric dielectric cover. This enables conformal mounting, removal, and replacement.
- The dielectric layer is preformed by thermoforming to improve adhesion on non-flat surfaces and avoid strain from bending.
- The external cover uses hemlines and screws, allowing it to be removed for inspection and substituted if damaged.
- On the iCub forearm, the triangular sensor mesh is glued to the cover with bicomponent glue and vacuum assistance to improve adhesion on the 3D surface.
- The fabric dielectric layers are glued, cut, shaped to the robot part, and fixed with screws; the procedure was repeated across iCub parts and a WAM arm.
A. Experimental Test Setup
The sensor was characterized with a Cartesian-robot setup using two prototypes: a flat patch and a 3D cover integrated on robot hardware.
- Experimental apparatus: Measurements used a three-axis Cartesian robot, an off-center load cell, and cylindrical probes attached to apply controlled contacts.Robot position and load-cell values were collected at 25 Hz.
- Robot integration: The sensor was integrated on an iCub and a WAM arm from Barret Technology.
- Prototypes: The flat prototype comprised 16 triangular modules on a flat structure, while the 3D prototype covered the iCub forearm with 16 modules.The flat prototype used a 2 mm dielectric layer.
- Test allocation: Sensitivity, repeatability, hysteresis, and spatial resolution were evaluated with the flat prototype; thermal drift was evaluated with the 3D prototype.
B. Sensitivity and Repeatability
The sensor showed two approximately linear sensitivity ranges and higher low-pressure sensitivity than the previous silicone-foam design.
- Protocol: Sensitivity and repeatability experiments used central taxels, 2 mm and 7 mm probes, repeated incremental loading cycles, and unloading to the no-contact position.
- Sensitivity: The sensitivity was 2.50 fF/kPa from 2 to 45 kPa and 0.86 fF/kPa from 65 to 160 kPa.Sensitivity decreased after the dielectric layer became more compressed and less compliant.
C. Hysteresis and Relaxation
Repeated load–unload tests showed low hysteresis, while relaxation testing quantified the dielectric’s time-dependent response.
- Protocol: The hysteresis experiment used the same taxels and load–unload cycle as the preceding test, with 15 repetitions and one-minute waits between repetitions.
- Hysteresis: The hysteresis result was described as remarkable for a capacitive sensor.
- Relaxation: The stress-relaxation experiment held the dielectric at fixed deformation for 10 minutes at constant room temperature and compared sensor, load-cell, and indenter responses.
- Relaxation: The relaxation constant was 1 h 18 min, providing a parameter for estimating stress relaxation after a duration T under constant deformation.
D. Spatial Resolution
Spatial resolution was measured by moving indenters along a line in 0.2 mm increments and recording responses across activated taxels. The responses were bell-shaped with overlapping receptive fields, and their widths varied with indenter diameter.
- D. Spatial Resolution: The 2 mm probe activated taxels whose average responses and standard deviations were recorded at successive positions.The figure identifies the activated taxels by their numbering in Figure 15(a).
- D. Spatial Resolution: Responses formed bell-shaped curves with overlapping receptive fields across the excited taxels.The response pattern was observed for the spatial-resolution measurements.
- D. Spatial Resolution: Receptive-field width changed with indenter diameter, confirming that deformation affects the spatial response.The comparison used 2 mm and 7 mm indenters.
E. Temperature Compensation
Temperature compensation uses embedded thermal sensors and taxel-specific calibration gains to correct capacitance readings. Experiments across heating and cooling cycles found the compensation effective over the tested temperature range and transferable across triangles.
- E. Temperature Compensation: Temperature compensation adds a thermal sensor’s capacitance change, ΔC, multiplied by a gain to the taxel output.The flexible FPCB embeds two pressure-insensitive thermal sensors with nearly identical behavior.
- E. Temperature Compensation: The compensated taxel value subtracts the calibrated thermal contribution from the raw taxel value relative to the startup thermal baseline.Ki is estimated separately for each taxel during calibration.
- E. Temperature Compensation: Heating and cooling experiments varied temperature between 15°C and 40°C using a programmable oven and a 3D prototype.The experiments recorded all taxels in a single triangle, including the two thermal sensors.
- E. Temperature Compensation: Compensation was effective over the tested temperature range, and the same Ki values could be used across different triangles without individual calibration.Different Ki values were determined for each taxel, while thermal drift was similar among groups of taxels.
V. CONCLUSIONS
The revised tactile skin replaces the elastomeric dielectric with a fabric-based layered structure and adds thermal sensors for drift correction. Experiments showed good sensing performance, small hysteresis, and effective compensation of temperature-induced drift.
- V. CONCLUSIONS: The revised sensor replaces the elastomer dielectric with a sandwich of deformable 3D fabric, conductive Lycra, and protective fabric.The new dielectric simplified production and improved mechanical properties, mounting, and replacement.
- V. CONCLUSIONS: The sensor was characterized for repeatability, sensitivity, hysteresis, and spatial resolution, showing good performance and small hysteresis.The conclusion summarizes the experimental characterization of the revised system.
- V. CONCLUSIONS: Thermal sensors embedded in the FPCB effectively compensated drift induced by temperature variations.The thermal units were introduced as part of the revised system.