Elsevier

Sensors and Actuators A: Physical

Volume 264, 1 September 2017, Pages 289-297
Sensors and Actuators A: Physical

Bioinspired flexible microfluidic shear force sensor skin

https://doi.org/10.1016/j.sna.2017.08.001Get rights and content

Highlights

  • Bioinspired resistive shear force sensor skin for light touch applications.

  • Sensors are located away from contact area enabling multiple sensing mode integration.

  • Skin fabricated from liquid metal filled microfluidic channels and compatible with large strains.

  • Dynamic range is more than 10-bit with shear force up to 5 N.

Abstract

There is a need to gather rich, real-time tactile information to enhance robotic hand performance during haptic exploration and object manipulation. Measuring shear forces is useful for grasping and manipulating objects; however, there are limited effective shear sensing strategies that are compatible with existing end effectors. Here, we report a bioinspired and flexible, resistive microfluidic shear force sensor skin. The sensor skin is wrapped around a finger-shaped end effector and fixed at the location of the nail bed. When the skin is subjected to shear force, one side of the skin experiences tension while the other side experiences compression and bulges similar to a human fingerpad. The tension and compression are measured by liquid metal strain gauges, embedded in PDMS, that are strategically placed adjacent to the nail bed, away from regions of direct finger-object contact. We present the sensor design, a finite element analysis static mechanical characterization model, as well as static response experiments. The resistive shear sensing skin exhibits greater than 10-bit dynamic range (up to 5 N) that is insensitive to the applied normal force. The resistive shear sensing skin is intrinsically flexible and immune to fatigue and other problems of solid-state sensors when subjected to repeated, large strains.

Introduction

Human fingertips are equipped with mechanoreceptors to collect tactile information so that we can precisely control muscle coordination during tasks such as haptic exploration and object manipulation [1]. This information enables humans to deduce some object properties that cannot be inferred from other modes of sensing, such as vision [2]. Despite the importance of tactile sensing, it is not widely implemented in robot hands or prostheses [3]. A robot hand equipped with a deformable multimodal tactile sensor has been used to demonstrate the ability to perceive local shape, which could be used to advance robot autonomy and provide haptic feedback to human teleoperators [4].

Haptic exploration and object manipulation benefit from rich, real-time tactile feedback through physical contact [3], [5]., [6], [7]. It has been shown that human operators tend to exert more gripping force than needed when using visual feedback alone [2]. Damage of the manipulated object and/or the end effector itself due to overloading is also of concern when tactile feedback is not available [8]. Restoring the sense of touch may be valuable in delicate and precise manipulation applications such as tele-operation of minimally invasive surgery [5]. Several tactile sensors and algorithms have been developed to pick up objects of varying mass and texture [9], estimate the contents of a container [10], and distinguish light and firm touch [11].

Spatially and temporally resolved normal and shear stresses are critical mechanical measurements that need to be resolved on artificial fingertips. Sensor design criteria are typically task-specific; however, general sensor specifications can be guided by physiological properties of human mechanoreceptors [6], [12]. Within the scope of in-hand manipulation, normal and shear forces up to 10 N should be resolved with an optimal dynamic range of 1000 [6]. In practice, sensing ranges of 0.3–50 N have been demonstrated in engineered tactile sensors according to their target application [3], [5]., [13], [14]. A normal force spatial resolution of 1.25 mm has been suggested to properly resolve object geometric features [3], [6], [15]. Kyung et al. have shown that humans have high normal force spatial resolution for interactions below 32 Hz [16]. It is less clear how well shear forces need to be spatially resolved, spurring further research of desired resolution during in-hand manipulation. Low latency temporal response contributes to better control stability, faster reflex-like responses, and can enable inference of surface properties such as roughness, shape, or coefficient of friction during haptic exploratory motion [17], [18], [19], [20], [21]. Humans can detect vibrations as high as 700 Hz although frequencies such as 100 Hz, 250 Hz or 1000 Hz have been proposed for vibration sensors for specific applications [3], [6], [13], [14], [15].

A wide variety of tactile sensors have been demonstrated including whole finger [22] and artificial skin based sensors that are compatible with existing robotic and prosthetic hand actuators [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36]. Sensor skins that can be wrapped conformally around robotic manipulators minimize the need for modification of the end effector. Skin based sensors have been fabricated with liquid-metal-based transducers embedded in elastomers. Polydimethylsiloxane (PDMS) based sensors have been demonstrated as robust, non-toxic, highly flexible, low cost, and easily fabricated [23], [28], [29], [30], [34], [36], [37], [38], [39]. Liquid metal filled channels, such as those filled with eutectic gallium indium alloy (eGaIn), are intrinsically immune to cracks and fatigue, suitable for conformal wrapping and large, repetitive strains [37], [40], [41], [42], [43]. eGaIn is a conductive liquid metal at room temperature and can form microstructures with a 2 μm spatial resolution by injection into micro-channels [23], [39], [40], [44], [45], [46], [47]. eGaIn traces that are 200–300 μm wide can be rapidly patterned by microcontact printing, 3D printing, or stencil lithography [48], [49], [50], [51].

A variety of miniaturized microelectromechanical system soft sensor skins with liquid metal that use capacitive, resistive and inductive sensing modalities have been developed [23], [38], [39], [44], [45], [47]. Capacitive sensors are typically composed of two flat-plate electrodes separated by a dielectric material and benefit from high sensitivity, ease of array fabrication, and high spatial resolution [23], [36], [44]. We previously developed a light touch microfluidic normal force sensor that achieved a spatial resolution of 1 mm with a 5 by 5 array of deformable parallel plate capacitors and was calibrated from 0 to 2.5 N [23]. Roberts et al. presented a soft sensor that used differential measurements in multiple parallel plate capacitor taxels to detect shear deformation in two orthogonal dimensions [44]. Capacitive based sensors are susceptible to electromagnetic interference, crosstalk, and parasitic capacitance.

Resistive sensors, including strain gauges and piezoresistive sensors, convert deformation to change of resistance, which can be easily measured by 4-wire sensing or a Wheatstone bridge circuit. Resistive sensors generally have good sensitivity over a large sensing range, high scanning rate, and are easy to design and implement. Their weaknesses include lower repeatability, high power consumption, and sensitivity to temperature. Park et al. presented an artificial skin capable of measuring uniaxial strain in three directions constructed with serpentine and spiral eGaIn channels; however, the minimal normal force that could be sensed was 7.4 N [47]. Majidi et al. developed a single microfluidic channel resistive sensor skin that measures curvature and could be useful in applications such as angular position feedback of an articulated joint [46].

There are only a few strategies to measure shear force, such as pairs of planar strain gauges [27], vertical cantilevers [49], or clusters of normal force taxels that are strategically positioned in pairs or quadruples such that shear force acting at the center of the sensor cluster results in combinations of normal force readings [36], [40], [41], [50]. Sensor sensitivity can be enhanced by adding a bump or ridge on the sensing surface or adding rigid force-posts embedded in the media over the sensing elements [40], [41], [50], although this approach suffers from reduced spatial resolution since it takes several taxels to construct a single sensing unit. Using a combination of normal force sensors could falsely generate shear force readings from normal forces that are non-uniform or localized on an individual taxel. Previous reports on shear force sensors collectively demonstrate calibration under a single normal force level and rarely discuss the influence of pre-applied normal force. Shear sensors typically measure forces at the point of physical contact between the sensor and object, but, in robotic applications, the location of contact may be arbitrary and would necessitate a large number of distributed sensors. Tactile sensors that resolve shear forces away from the point of contact may be advantageous, especially when faced with constraints on space, such as the small region of a fingerpad. Sensing modalities that do not require localized sensing for direct measurement of stimuli can be displaced to make room for other sensing modalities, such as normal force.

In this paper, we develop a bioinspired, thin and flexible liquid metal filled resistive PDMS microchannel shear force sensing skin. The sensor skin is wrapped around a finger-shaped effector and fixed at the location of the nail bed. When the skin is subjected to shear force it results in one side of the skin in tension and the other side in compression that buckles and bulges similar to a human fingertip. The tension and compression are measured by embedded liquid metal strain gauges that are strategically placed adjacent to the nail bed, away from the point of finger-object contact. We present the sensor design, a finite element analysis (FEA) static mechanical characterization model in addition to an analytical approach, as well as static response experiments. We show that the resistive shear sensing skin exhibits large dynamic range that is insensitive to the applied normal force over a range of shear forces. The resistive shear sensing skin is intrinsically flexible and immune to fatigue when subjected to repeated large strains.

Section snippets

Sensor design and theory of operation

The shear sensing skin design is inspired by the layered structure of the human fingertip. Skin consists of epidermis (outermost layer), dermis, and subcutaneous fat tissue. While all of these layers are soft relative to the underlying bone, the subcutaneous fat tissue is much softer than epidermis and dermis, thus, skin tends to shear and slide with respect to the underlying bone when shear force is applied to the finger pad. This deformation results in tension on one side of the fingerpad and

Numerical simulations of shear sensor

We developed steady, two-dimensional finite element solid mechanics models (COMSOL Multiphysics, Boston, MA) with linear elasticity to predict the sensor response to various normal and shear force loading conditions. We used these models to determine the appropriate placement of the strain gauges to obtain linear, high sensitivity sensor response that is independent of the applied normal force. The model was composed of two domains: one for the rigid fingertip core and one for the PDMS

Results and discussion

In this section, we present an analytical analysis of the shear sensor skin response and computational predictions of the sensor deformation under various normal and shear forces. The static calibration of the shear force sensor is presented and critical performance criteria are calculated including sensitivity, shear force resolution, and prediction interval.

Summary

Here we report a bioinspired, thin and flexible shear force sensor skin that deforms in response to applied shear forces similar to a human fingertip. The shear sensor skin’s novelty is that the tension and compression are measured by embedded liquid metal strain gauges that are strategically placed adjacent to the nail bed, away from the point of finger-object contact. This design has the advantage over existing skin based shear sensor designs that it is free from confounding effects of normal

Acknowledgments

This material is based upon work supported by the National Science Foundation under Grant No. CBET-1264046 and CBET-1461630. The authors would also like to thank UW NNIN Nanofabrication Facility for training and support in the clean room and Dr. Per G. Reinhall who provided suggestions on the contact mechanics modelling and selection of lubricants. We are grateful to Stefan Foulstone for assistance in generating Fig. 1.

Jianzhu Yin obtained his B.Eng. degree in Thermal Engineering at Tsinghua University in June 2012. He is currently in the Mechanical Engineering Ph.D. program at University of Washington. His work focuses on the design, fabrication, characterization, and application integration of microfluidic tactile sensors.

Reference: (56)

  • J. Monzée et al.

    The effects of digital anesthesia on force control using a precision grip

    J. Neurophysiol.

    (2003)
  • R.S. Dahiya et al.

    Tactile sensing—from humans to humanoids

    Robot. IEEE Trans. On

    (2010)
  • R.D. Ponce Wong et al.

    Haptic exploration of fingertip-sized geometric features using a multimodal tactile sensor

  • A.M. Okamura

    Haptic feedback in robot-Assisted minimally invasive surgery

    Curr. Opin. Urol.

    (2009)
  • B. Preising et al.

    A literature review: robots in medicine

    IEEE Eng. Med. Biol. Mag.

    (1991)
  • Z. Su

    Force estimation and slip detection/classification for grip control using a biomimetic tactile sensor

    2015 IEEE-RAS 15th International Conference on Humanoid Robots (Humanoids)

    (2015)
  • S. Chitta et al.

    Tactile sensing for mobile manipulation

    IEEE Trans. Robot.

    (2011)
  • M. Weigel et al.

    ISkin: flexible, stretchable and visually customizable on-Body touch sensors for mobile computing

  • R.S. Johansson et al.

    Tactile sensory control of object manipulation in humans

  • Johan Tegin et al.

    Tactile sensing in intelligent robotic manipulation –a review

    Ind. Robot Int. J.

    (2005)
  • J. Dargahi et al.

    Human tactile perception as a standard for artificial tactile sensing—a review

    Int. J. Med. Robot.

    (2004)
  • K.-U. Kyung et al.

    Perceptual and biomechanical frequency response of human skin: implication for design of tactile displays

    Eurohaptics Conference, 2005 and Symposium on Haptic Interfaces for Virtual Environment and Teleoperator Systems, 2005. World Haptics 2005. First Joint

    (2005)
  • J.A. Fishel et al.

    Bayesian exploration for intelligent identification of textures

    Front. Neurorob.

    (2012)
  • M.R. Tremblay et al.

    Estimating friction using incipient slip sensing during a manipulation task

    IEEE International Conference on Robotics and Automation, 1993. Proceedings

    (1993)
  • R. Howe et al.

    Sensing skin acceleration for slip and texture perception

    IEEE International Conference on Robotics and Automation, 1989. Proceedings

    (1989)
  • H. Zhao et al.

    Optoelectronically innervated soft prosthetic hand via stretchable optical waveguides

    Sci. Robot.

    (2016)
  • N. Wettels et al.

    Biomimetic tactile sensor array

    Adv. Robot.

    (2008)
  • M.L. Hammock et al.

    25th anniversary article: the evolution of electronic skin (E-Skin): a brief history, design considerations, and recent progress

    Adv. Mater.

    (2013)
  • Cited by (63)

    • Textiles in soft robots: Current progress and future trends

      2022, Biosensors and Bioelectronics
      Citation Excerpt :

      Many efforts have been devoted to the development of electronic textiles based on electrically conductive materials, which mainly include CNTs (Ryu et al., 2015; Wang et al., 2017d), nanofibers (Wang et al. 2017a, 2017c), graphene (Cheng et al., 2015; Mousavi et al., 2011; Yang et al., 2018b), metal nanowires (Chen et al., 2018; Lee et al., 2018) and conducting polymers (Grancarić et al., 2017). Many groups have already developed different manufacturing techniques for the fabrication of soft sensors, particularly in the form of embeddable device (Cheng et al., 2010; Guan et al., 2016; Shen et al., 2021; Sonar and Paik, 2016; Yin et al., 2017). Guan et al. fabricated a sensor by functionalizing fibers in a wristband using liquid ion solution and found that a firm liquid/solid interaction during the stretching process can contribute to a stable liquid bridge between the fibers, and this sensor can detect movement and pressure and has potential to track human motion for sport injury (Guan et al., 2016).

    • Research Progress and Prospect of Gallium-Based Liquid Metals in Electrical-Thermal-Mechanics Field

      2023, Xinan Jiaotong Daxue Xuebao/Journal of Southwest Jiaotong University
    View all citing articles on Scopus

    Jianzhu Yin obtained his B.Eng. degree in Thermal Engineering at Tsinghua University in June 2012. He is currently in the Mechanical Engineering Ph.D. program at University of Washington. His work focuses on the design, fabrication, characterization, and application integration of microfluidic tactile sensors.

    Veronica J. Santos is an Associate Professor in the Mechanical and Aerospace Engineering Department at the University of California, Los Angeles, and Director of the UCLA Biomechatronics Lab. Dr. Santos earned her Ph.D. in mechanical engineering with a biometry minor from Cornell University (2007). While a postdoc at the University of Southern California, she contributed to the development of a biomimetic tactile sensor for prosthetic hands. Her research interests include human hand biomechanics, human-machine systems, haptics, tactile sensors, machine perception, prosthetics, and robotics for grasp and manipulation. She currently serves on the Editorial Boards for the ASME Journal of Mechanisms and Robotics and the IEEE International Conference on Robotics and Automation.

    Jonathan D. Posner is an associate professor joint in Mechanical Engineering and Chemical Engineering departments as well as adjunct professor in Family Medicine at University of Washington. He directs UW’s Engineering Innovation in Health program that focuses on developing technical solutions to pressing challenges in health and healthcare. Dr. Posner earned his Ph.D. (2001) degree in Mechanical Engineering at the UC Irvine, was a postdoctoral fellow at the Stanford University, and spent 18 months at the von Karman Institute for Fluid Mechanics. Jonathan’s group works on diverse set of need-driven research projects including medical devices, point-of-care biochemical diagnostics, improved cookstoves for the developing world, and helmets that reduce the risk of concussion.

    View full text