Optimizing Multi-Touch Textile and Tactile Skin Sensing Through Circuit Parameter Estimation

TL;DR

The paper addresses reliable, rapid multi-touch force sensing on textile-based tactile skins by reframing the problem as estimating resistive cell parameters in a grid. It introduces a Regularized Least Squares optimization that enforces circuit laws and mitigates ghosting, after a one-point calibration, enabling accurate multi-touch force predictions. A two-layer skin design is proposed to balance manufacturability and sensitivity, achieving a minimum detectable force down to 1 N. The approach yields substantial improvements in force-sensing accuracy (approximately 26–27% over naive methods) and demonstrates practical viability for human-robot interaction with scalable textile skins.

Abstract

Tactile and textile skin technologies have become increasingly important for enhancing human-robot interaction and allowing robots to adapt to different environments. Despite notable advancements, there are ongoing challenges in skin signal processing, particularly in achieving both accuracy and speed in dynamic touch sensing. This paper introduces a new framework that poses the touch sensing problem as an estimation problem of resistive sensory arrays. Utilizing a Regularized Least Squares objective function which estimates the resistance distribution of the skin. We enhance the touch sensing accuracy and mitigate the ghosting effects, where false or misleading touches may be registered. Furthermore, our study presents a streamlined skin design that simplifies manufacturing processes without sacrificing performance. Experimental outcomes substantiate the effectiveness of our method, showing 26.9% improvement in multi-touch force-sensing accuracy for the tactile skin.

Figures (8)

• Figure 1: An overview of our proposed tactile skin sensing method. (a) Two textile pieces with conductive stripes are separately knitted. (b) The two textile pieces are sewn together orthogonally to create a grid of sensing cells. The light pink fabric is made of Nylon Stretchy Yarn and the red vertical fabric stripes are made of Acrylic Yarn. (c) The skin is modeled as a resistive sensory array. Our approach predicts force applied on the skin by estimating cell resistances $R^{C}$ using the Arduino board.
• Figure 2: (a): Ghosting effects occur when an alternate path of current (shown in green) is formed for the sensing cell at the upper left corner, bypassing it completely. (b): Ohmmeter Configurations. Details are provided in Section IV(C).
• Figure 3: Simulation results for varying wire resistances in tactile skin. Brighter colors denote lower resistances, ranging from $0 \Omega$ to $1.0 M\Omega$. The pressed cells are assigned a lower$0.001 M\Omega$ resistance, while unpressed cells have higher$1.0 M\Omega$. Wire resistances vary as $0.0001-0.041 M\Omega$ from left to right. These values are typical for our setup.
• Figure 4: Simulation results for tactile skin with variable cell resistances. Brighter colors denote lower resistances, ranging from $0 \Omega$ to $1.0 M\Omega$. The pressed cells are simulated with resistances of $0.01- 0.7 M\Omega$ displayed from left to right. Unpressed cells have a fixed resistance of $1.0 M\Omega$, and wire resistances are $0.001 M\Omega$. These values are typical for our setup.
• Figure 5: The Force Estimation experiment setup. (a) flat surface experiment setup and (b)curved surface experiment. (c) two-layer design of the skin allows the detection of tiny forces as low as 1N. Blue dots represent a single-point training set. Red dots represent a multi-touch test set.