Symposium F.SM03—Materials and Mechanics Challenges in Haptics for Human-Machine Interfaces

For virtual reality (VR) to be truly immersive, the sense of touch must be stimulated: a tennis ball should feel light and rubbery, a wall must be impenetrable, a lightswitch must click when pressed. Why don’t we have haptic suits with thousands of individual actuators (taxels) when every smartphone display has millions of individually addressed pixels? Generating localized forces on the human body in a comfortable and safe way is a major challenge for soft actuation: both fast motion and high forces are needed, yet the device must conform to the human body, and consume low power.
I will present some approaches developed in my lab for flexible actuators combining high strain and high force, operating at high speed, and that do not rely on external compressed air or vacuum supplies. Due to its high energy density, we have focused on electrostatic actuation, using high electric fields to deform elastomers or textile structures.
For kinesthetic feedback, which can be described as controlling the motion of joint, we have focused on textile-based brakes, only 1 mm thick, that can block the motion of two sliding strips in a few milliseconds[1]. By coating conductive textile strips with high permittivity dielectrics, we developed a clutch that blocks up to 20 N pulling force per square centimeter of active region, for a power consumption of under 2 mW. We integrated these clutches into thin gloves. Users wearing them are much more accurate when manipulating virtual objects and feel deeper immersion. We will report a new fabrication method allowing for even higher forces for use in full-body kinesthetic haptics, blocking shoulder and elbow motion to allow feeling virtual objects as being heavy, and virtual tables as blocking hand motion.
Cutaneous feedback requires dense arrays of actuators to locally stimulate to skin to provide the illusion of touching an object. We use our fingers to identify objects by sliding our fingertips over them, and we know if a glass is about to slip out of our hand by sensing shear forces. Simulate this in VR requires being able to generate normal and shear forces with high spatial resolution. We have developed arrays of sub-mm thick flexible actuators that generate over 60% strain and operate at over 300 Hz. A 6 mm diameter actuator generates 300 mN normal force and 500 µm displacement. This is achieved using fluidically coupled electrostatic zipping and a combination of flexible polymers with high breakdown field and silicone elastomers that allow for high displacement[2]. We have made 5x5 arrays, mounted them directly on the arm and on consumer products, and report on user feedback for notification, control and navigation.
By use of thin dielectrics, high permittivity films, and low stiffness electrodes, we have reduced the drive voltage of Dielectric Elastomer Actuators (DEAs) to 400 V, a level at which we can use SMD components for compact control electronics[3]. We report “feel-through” untethered cutaneous haptics, with DEA actuators only 18 µm thick, so thin the user does not feel them mounted on his or her fingertip when they are off. However, when the 3 mm diameter devices are turned on, the user feels localized pulsation that allows receiving rich haptic information.
I will close by going over the challenge of integrating these technologies in a glove or suit, and the promise of this field.

[1] R. Hinchet and H. Shea, “High Force Density Textile Electrostatic Clutch,” Advanced Materials Technologies, vol. 5, p. 1900895, 2019, doi: 10.1002/admt.201900895.
[2] E. Leroy, R. Hinchet, and H. Shea, “Multimode Hydraulically Amplified Electrostatic Actuators for Wearable Haptics,” Advanced Materials, 2020. in press.
[3] X. Ji et al., “An autonomous untethered fast soft robotic insect driven by low-voltage dielectric elastomer actuators,” Science Robotics, vol. 4, no. 37, 2019, doi: 10.1126/scirobotics.aaz6451.

Haptic interfaces for virtual reality (VR) or augmented reality (AR) require dense arrays of actuators to provide realistic feedback to the user. These actuators must be fast (<20 ms), generate large static forces (>100 mN) and displacements (>300 µm) while being easy to integrate in lightweight, comfortable interfaces. The actuator mechanical compliance should be close to that of skin. Conventional actuation technologies such as electromagnetic, piezoelectric or pneumatic either provide only vibratory feedback or are difficult to integrate in low-profile interfaces. Flexible electrostatic actuators such as our HAXEL devices offer effective solutions to these challenges.
HAXELs are hydraulically amplified zipping electrostatic actuators, less than one millimeter in thickness, designed for integration into textile-based haptic interfaces for VR or AR scenarios. In previous work [1], we reported 6 mm x 6 mm HAXEL actuators that can produce up to 300 mN force and 500 µm displacement (i.e. 60% strain). We report here the modeling and experimental results on scaling HAXELS down to 3 mm and up to 50 mm diameter, for applications in wearable haptics and in soft robotics.
Each HAXEL consists of a fluid-filled cavity whose shell is made of a non-stretchable polymer in the outer regions and of a stretchable elastomer for the central region. Electrodes are patterned in both top and bottom polymer shells. When a voltage is applied to the annular electrodes, they are pulled together by Maxwell pressure, starting to zip at the outer regions where the electric field is highest. As the top and bottom polymer regions are pulled together, the fluid is rapidly forced into the stretchable region, forming a raised bump. To obtain high power density (>100 W/kg), we use high permittivity layers based on PVDF terpolymers.
These actuators can easily be felt by users and provide both normal force and shear force by using segmented electrodes to drive the fluid in-plane. Using a flexible 5x5 matrix of HAXELs, users can obtain information about the location and texture of a virtual object. HAXELs could lead to lightweight and comfortable gloves, sleeves or full-body suits delivering realistic haptic feedback.
Different locations on our bodies have different sensitivity thresholds for feeling force and displacement, and different pitch discrimination. On fingertips we can feel pins 0.8 mm apart, while on our back the distance is roughly 5 mm. It is necessary to adapt the size of the taxels to body location and to the targeted VR application.
The force and displacement characteristics of HAXELs from 3 mm x 3 mm up to 50 mm x 50 mm are discussed and compared to a finite element model that accounts for coupled electrostatics and mechanics using nonlinear material models. Both static and dynamic characteristics are studied as they are equally important for haptic feedback. The measured static displacements range from 100 µm to several mm while the forces range from tens of mN to over 1 N. Good agreement between the model and the experiment is shown, validating the finite element model as a tool for future designs. Optimized electrode shape, cavity shape and effect of materials choice on actuation voltage and generated force, displacement and device dynamics are presented.
Although HAXEL actuators were initially designed for haptics, they can find applications in other areas such as microfluidics as dense arrays of valves or to rapidly reconfigure channels, and for soft robotic as matrices of high force actuators for locomotion or grasping.
[1] E. Leroy, R. Hinchet, H. Shea, “Multimode Hydraulically Amplified Electrostatic Actuators for Wearable Haptics”, Advanced Materials, 2020, in press.

Soft valves for hydraulic actuators that based on meso-scale channels have been designed in various forms, primarily fluid-driven or electrically-driven, with promise for the control of soft robots. Fluid-driven valves are often tethered to external pressure sources that limit, for example, robot locomotion. Electrically-driven valves based on electrostatic forces exhibit a large hysteresis and slow response times. In our work, we present a dynamic electrically-driven soft valve made of ultra high power density dielectric elastomer actuators (DEAs). These dynamic DEAs are actuated at 200Hz or above, with low damping (tan(δ) < 0.34 at 100Hz) and a short response time (in less than 10 milliseconds). Dynamic actuation produces 300% greater blocked force compared to quasi-static operation with the same DEAs due to higher driving voltages and greater than five times increase in free displacement at the actuator resonance. The channel, on the contrary, is cast with a dissipative material. The viscoelasticity difference between the actuator and the channel materials makes it possible to achieve both pressure and flow control with the dynamic DEAs. The valve performance can be tuned by the input voltage, the driving frequency, and the active area of the DEAs. This dynamic electrically-driven soft valve, as a component of a soft fluidic robot, enables effective control of 1.5psi flow pressure in soft channels. The novel dynamic design represents a substantial performance improvement over existing quasi-steady DEA valves, and will enable effective control of future meso-scale hydraulic robots.

Can we use robots to help humans learn a new motor skill, or to improve performance of a motor task? If so, what are the neural substrates that enable humans to improve motor skill learning or execution when they interact physically with external agents? These are fundamental motor neuroscience questions, which have or will have special relevance in multiple domains, such as human augmentation, surgical training, athletics, and neurorehabilitation. In this talk, I will present the development of methods that can help address these questions, which combine MRI-compatible robotics with functional neuroimaging and advanced biosignal processing, and will present two applications of these methods in fundamental motor neuroscience research.

Over the past years, my lab has developed a new class of MRI-compatible robots that can be used in conjunction with functional Magnetic Resonance Imaging to study the neural substrates involved in the neural control of movements during physical interaction with external agents. Specifically, I will present two systems, the MR-SoftWrist, a 2 DOF wrist exoskeleton capable of force feedback designed to study neural substrates involved in the control of wrist movements during externally imposed forces, and the MR-StretchWrist, a 1DOF wrist robot designed to apply velocity-controlled perturbations to the wrist to condition stretch reflexes and study their neural correlates using fMRI.

The development of these robots is complicated by the magnetic fields of MRI, which pose challenges from the standpoint of materials selection for robot components, electrodes for measuring biosignals associated with motor tasks, as well as biosignal processing. To achieve MRI-compatibility, our current solution is to include piezoelectric ultrasonic motors for actuators, shielding and filtering of sensor signals, and use non-magnetic metal for all robot parts. However, because piezoelectric motors are intrinsically high-impedance actuators, we have developed custom solutions based on series elastic actuation and nonlinear passivity-based force control for reducing actuator impedance. Moreover, we have recently developed a novel method for measuring artifact-free electromyographic signals during MRI which enables for the first time collection of biosignals during arm movements under fMRI using a novel set of electrodes and nonlinear interference estimation algorithms. Validation experiments demonstrate that the biosignals measured during MRI are not significantly different than those measured in normal lab conditions.

In this talk, I will also present the application of MRI-compatible robots to establish the neural correlates involved in the neural control of movements. In one experiment, we exposed 30 healthy individuals to lateral force perturbations during point-to-point wrist movements to elicit adaptation to this condition. By analyzing resting state fMRI data measured before and after motor execution and before and after adaptation, we were able to identify areas involved in creating and processing the memory related to movement execution under force perturbations. In a second experiment, we used the MR-StretchWrist to apply perturbations to the wrist along the flexion/extension axes and condition long-latency responses. Long-latency responses are interesting components of motor control because they gracefully blend fast reaction times of spinal reflexes with sophisticated modulation capabilities of voluntary control. In our first experiment with the MR-StretchWrist, we were able to determine the somatotopic organization of long-latency responses of flexor and extensor muscles in the brainstem for the first time in-vivo in humans. The observed organization is partially consistent with animal models, with activity primarily in the ipsilateral medulla for flexors and in the contralateral pons for extensors, but also include other areas, such as the midbrain and bilateral pontomedullary contributions.

Electroactive artificial muscles have been a scientific novelty soft actuator for many years, but of little practical use because they lack substantial force output or large-enough displacement. Significant advancements in constituent materials, actuator design and assembly of component artificial muscles for devices have now allowed the transition of this interesting technology into actual consideration for real-world applications in haptics, human-machine interactions and soft robotics. The electroactive artificial muscles discussed here are, fundamentally, flexible capacitors that give a mechanical response when an electrical potential is applied to them. As with any capacitor, greater force is expressed with an inverse squared relationship to the thickness of the electrical insulation used to space the positive and negative electrodes. As insulation thickness is reduced, output force from the artificial muscle is increased. We have determined that particular acrylic polymer blends exhibit an impressively high resistance to electrical breakdown per-micron-thickness of material in these systems, when compared to other polymers including: polypropylenes, polyesters, polyimides and silicone polymers. Electrical insulation polymer optimization was found to be able to reduce electrode spacing by x4, which can manifest in up to a x16 increase in force. It was found that this increased force is lost if the artificial muscle stiffness is not tuned correctly. Too much stiffness immobilizes the muscle whereas a lower value would result in the muscle not able to transfer the elevated force being generated. This optimization was accomplished through specific layering of biaxially oriented polypropylene (BOPP). With these marked material improvements, forces of nearly 7N were generated for 0.5mm of displacement from an artificial muscle that is only ~350 microns thick in a resting state are achieved. A single stack of artificial muscles is rarely capable of generating the necessary force for real-world applications. The exceptional artificial muscles discussed here are uniquely capable of functioning by arranging the individual artificial muscles are arranged in an overlapping closest-packed tessellated pattern. This original system-level optimization still further improves artificial muscle force output by an additional x4; setting these original ensembles of artificial muscles apart from related technology in the field of soft robotics research. These flexible actuators hold the promise to proliferate the automotive industry and beyond for haptics in general because they are light weight, quiet, compliant, thin, simple to implement and new.

Wearable technologies for human-machine interaction (HMI) require on-board sources of energy to power sensors for tactile or physiological monitoring, actuators for haptic feedback, and electronics for signal processing and communication. For many emerging haptic/HMI applications, miniature battery technologies (e.g. coin cells, LiPo) are inadequate – operation over sustained durations instead depends on larger batteries or methods of harvesting energy from the body or environment. However, in order to be compatible with the human body, such energy storage and harvesting systems should be soft and stretchable so that they can deform with the surrounding tissue and not impair natural motion.

In this talk, I will highlight progress in creating soft, highly stretchable batteries and energy harvesting devices using liquid metal (LM) alloys. Eutectic gallium indium (EGaIn) is selected as the LM alloy due to its enabling combination of high electrical conductivity and fluidic deformability. In particular, I will focus on the following LM architectures: (i) soft-matter batteries composed of EGaIn and ionic gel that can be stretched to >300% strain; (ii) LM-polymer composites for thermoelectric and triboelectric energy harvesting; and (iii) dielectric elastomer generators composed of an LM-polymer dielectric and thin-film EGaIn coating. To demonstrate their potential role in future haptics/HMI, I will present examples of how these LM-based technologies are used for powering wearable sensors and electronics.

Gallium-based liquid metals have shown great potential in applications such as soft and stretchable electronics due to their high conductivity and intrinsic softness. However, liquid metal-based bulk conductors are typically sensitive to strain and therefore not be suitable for stretchable interconnects. Additionally, reliably interfacing liquid metal with conventional, rigid electronics often requires cumbersome fabrication processes. In this work, we present a liquid metal-based biphasic conductor with high conductivity, extreme stretchability (>1000%), negligible resistance change when strained, cyclic stability, and a reliable interface with rigid electronics. The biphasic material is a mixture of liquid metal and crystalline solids, which are produced by thermally processing liquid metal particles. We use a scalable transfer-printing process to create various multilayer circuit board assemblies that maintain their performance when stretched. This technique enables facile and direct conversion of existing PCB designs into soft and stretchable forms for applications in soft robotics and wearable devices.

Conductive polymers have been extensively used as flexible and transparent alternatives to inorganic conductors in transistors, light-emitting diodes, and solar cells. These “classic” applications, however, do not take advantage of some of the unique features of conductive polymers such as their ionic conductivity and our ability to tune their physical and mechanical properties via chemical modification. These features really shine in the developing area of bioelectronics: the interface between electronic materials and devices, and biological systems. In particular, I will highlight in this talk our efforts in using conductive polymers for haptics, to study or stimulate the human sense of touch.
The most widely-used conductive polymer is doped poly(3,4-ethylenedioxythiophene) (PEDOT), commercially available in water as a polyelectrolyte complex with poly(styrene sulfonate) (PSS). PEDOT:PSS, however, often requires the use of additives to modify its electronic and mechanical properties leading to potentially unstable and unsafe materials. Instead, we are using PSS copolymers to modify the intrinsic properties of PEDOT:PSS without additives. In particular, we have synthesized stretchable conductive polymers which served as electro-tactile materials to simulate rough and smooth surfaces in virtual reality. I will also share our preliminary research in the development of PEDOT:PSS conductive copolymers that can dynamically change their mechanical properties under applied voltages: electro-active actuators.

Smart fluids are suspensions of functional particles in which an applied field drastically changes the fluid’s properties. For example, magnetic microparticles in solution can form magnetorheological fluids (MRFs) that solidify upon the application of a magnetic field because of the formation of rigid particle chains. By virtue of being a fluid, these materials could provide an empowering path to realizing tunable actuators in soft robotic systems for human interactions. To test this idea, we successfully use an MRF as the active fluid in a soft actuator and gripper. Despite this success, the efficacy of MRF is inversely proportional to its density, which is commonly high on account of containing a high volume fraction of iron. As a path to realizing more efficient MRFs, we hypothesized that the inclusion of ultra-high aspect ratio 2D materials could provide the needed increase in yield stress without requiring a high volume fraction of dense iron particles. Thus, we explored 2D material sheets as additives in an MRF and tested its response to magnetic fields in shear mode, where the fluid is sheared between two plates, and in valve mode, where a pressure gradient pushes the fluid through a tube. While the addition of the 2D material did not affect performance in shear mode, its presence increased the pressure difference seen in valve mode at low flow rates. To explain these seemingly contradictory results, we propose a model based on the 2D material improving the yield stress by averaging out the shear stress over length scales commensurate with the particle size. This discovery can be used to rationally design MRFs for valve mode operation, which is important to their successful incorporation into soft robotics. These advances motivate further exploration of MRFs in soft robotics and reveal the influence of highly anisotropic additives on MRF performance.

Nature inspires a wide range of bio-mimetic systems ranging from soft robotic actuators to perceptive electronic skins that enhance and support our life. The growing demand on assistive, medical and bioelectronic technologies however raises concerns on the ecological footprint of this emerging platforms, as they are often designed for a defined, limited operational lifetime. Introducing a key feature essential to nature - biodegradability - will enable soft electronic and robotic devices that reduce (electronic) waste and are paramount for a sustainable future.
We here introduce materials and methods such as tough yet biodegradable biogels for soft systems that facilitate a broad range of applications, from transient wearable electronics to metabolizable soft robots. These embodiments are highly stretchable, are able to heal and are resistant to dehydration with unchanged mechanical properties for more than one year on the shelf. Our forms of soft electronics and robots are built from resilient gelatin-based gels with tunable, extreme mechanical properties that uniquely combine performance and durability with degradability. They are engineered for long-term operation in ambient conditions without fatigue, sustaining more than 350 000 actuation cycles and even operate under water for extended periods of time, but fully degrade after use through biological triggers. Electronic skins that measure pressure, strain, temperature and humidity serve as human-friendly on-skin interfaces or equip robotic systems with sensory feedback. Such advances in the synthesis of biodegradable, mechanically tough and stable gels that do not compromise in performance when compared to their non-degradable counterparts may bring bionic soft systems a step closer to nature and enable human-friendly technologies with reduced ecological footprint.

M. Baumgartner, F. Hartmann, M. Drack, D. Preninger, D. Wirthl, R. Gerstmayr, L. Lehner, G. Mao, R. Pruckner, S. Demchyshyn, L. Reiter, M. Strobel, T. Stockinger, D. Schiller, S. Kimeswenger, F. Greibich, G. Buchberger, E. Bradt, S. Hild, S. Bauer, M. Kaltenbrunner, “Resilient yet entirely degradable gelatin-based biogels for soft robots and electronics”, Nature Materials (2020)

For the sense of sight, we can purchase high definition screens to recreate nearly any image. There is no equivalent technology for the sense of touch. However, the accurate and rich recreation of tactile sensations could have broad implications in human machine interfaces, virtual reality, and health.

Touch is a complex event. As a finger and object make contact, a rich variety of mechanical vibrations are produced, which contributes to the subtlety and object identification behind every day tactile experiences. In this talk, we will discuss how we can use materials control at interfaces to create new tactile sensations by controlling friction and adhesion. We will share our work on screening and characterizing tactile materials, demonstrating that tactile perception cannot be predicted from material properties alone. We also share our complementary work on building a fundamental basis for tactile perception, enabled by materials control over tactile stimuli. Establishing a fundamental basis for touch could help unify our understanding of touch and help recreate arbitrary tactile sensations through a combination of more basic units.

Stimuli responsive polymers and composites have garnered attention in the development of novel actuators due to their ability to respond to various stimuli such as light, heat, electric field, magnetic field, and more. Compared to traditional electric motors, these actuators offer advantages such as conformability, better biomimicking ability, lower cost, and higher power-to-weight ratio. These advantages also are ideal for the field of soft robotics, creating flexible morphing structures that is able to move with more degrees of freedom then conventional stiff robotics. Piezoelectric materials, with external stimuli of electric field, the morphing device in use can respond with great speed and accuracy.
The current study investigates the development of a nanocomposite with 1D nanofillers single-walled carbon nanotubes (SWCNT) and boron nitride nanotubes (BNNT) in a piezoelectric polymer Polyvinylidene Fluoride (PVDF) matrix. The addition of nanotubes will: (i) increase piezoelectric property for enhanced actuation with its alignment, (ii) lower applied power due to conductive nanofiller percolation, and (iii) enhance mechanical property to withstand the environmental conditions of various actuation applications. Due to its high aspect ratio and conductivity, PVDF/SWCNT samples after stretching had high polar β crystal content of 80% and high conductivity (>10-6 S/cm). This resulted in high deflection magnitude at low applied electric field. At 0.5 wt.% PVDF/SWCNT stretched resulted in a maximum deflection of 570μm at 5.88 V/mm. With PVDF/BNNT, the samples after electric poling whilst lower in β crystal content, had higher deflection magnitude due to the higher overall dipole moment. PVDF/BNNT stretched and poled had β crystal content 12% lower than that of only stretched but 520% increase in deflection magnitude (from 50μm to 260μm) With an increase in conductivity and piezoelectric property, the material’s actuation can be performed with greater displacement and lower power.
With the addition of nanotube which provided constraints for alignment, percolation, and enhancement of mechanical properties, flexible morphing devices were developed. The choice of nanotube is important to optimize specific properties (i.e., electrical) but both nanofillers exhibited great morphing magnitude to various electric fields applied. This study provides insight to future research utilizing piezoelectric materials for actuation application in soft robotics.

Recently, ferromagnetic soft continuum robots – a type of slender, thread-like robots that can be steered magnetically – have demonstrated the capability to navigate through the brain's narrow and winding vasculature, offering a range of captivating applications such as robotic endovascular neurosurgery. Composed of soft polymers with embedded hard-magnetic particles as distributed actuation sources, ferromagnetic soft continuum robots produce large-scale elastic deflections through magnetic torques and/or forces generated from the intrinsic magnetic dipoles under the influence of external magnetic fields. This unique actuation mechanism based on distributed intrinsic dipoles yields better steering and navigational capabilities at much smaller scales, which differentiate them from previously developed continuum robots. To account for the presence of intrinsic magnetic polarities, this emerging class of magnetic continuum robots provides a new type of active structure – hard-magnetic elastica – which means a thin, elastic strip or rod with hard-magnetic properties. In this work, we present a nonlinear theory for hard-magnetic elastica, which allows accurate prediction of large deflections induced by the magnetic body torque and force in the presence of an external magnetic field. From our model, explicit analytical solutions can be readily obtained when the applied magnetic field is spatially uniform. Our model is validated by comparing the obtained solutions with both experimental results and finite element simulations. The validated model is then used to calculate required magnetic fields for the robot’s end tip to reach a target point in space, which essentially is an inverse problem challenging to solve with a linear theory or finite-element simulation. Providing facile routes to analyze nonlinear behavior of hard-magnetic elastica, the presented theory can be used to guide the design and control of the emerging class of magnetically steerable soft continuum robots.

Human-machine interfaces require soft, biocompatible materials that can provide high spatiotemporal resolution-coupling with bioelectronic devices. Here, we introduce an organic mixed-conducting particulate composite material (MCP) that enables highly-resolvable coupling by varying particle size and density. MCP-based high performance anisotropic films are pattern-free, scalable, and biocompatible. MCP enabled facile and effective electronic bonding between soft and rigid electronics, permitting recording of neurophysiological data at the resolution of individual neurons from freely moving rodents and from the surface of the human brain through a small opening in the skull. We also acquired high spatiotemporal resolution electrophysiological signals non-invasively by directly interfacing MCP with human skin. MCP provides a single material solution to facilitate development of bioelectronic devices that can safely acquire, transmit, and process complex biological signals.

Soft pressure sensors with high sensitivity over a wide pressure range are required for various applications such as electronic skins for human-mimetic robotics and electronic tattoos for digitizing human body. In the last decade, most research aiming at increasing the sensitivity of capacitive pressure sensors focused on developing dielectric materials with added air gaps and/or higher dielectric constants. After extensive research, sensitivity has been significantly improved at low pressure range, e.g. 1 kPa, but drops drastically as the pressure increases. To overcome this challenge, we present a novel soft capacitive pressure sensor employing an electrically conductive foam capable of both piezoresistive and piezocapacitive responses. The foam is made out of functionalized carbon nanotubes and Ecoflex and can be inexpensively fabricated without MEMS technology. The foam is 600-µm thick, 85% porous, and open cell with tubular ligaments. An ultrathin dielectric layer was added between the conductive foam and the electrode to ensure the whole device is still capacitive. The sensor has a modulus of 2 kPa and an initial impedance of 47 MΩ with a phase angle of -86°. This capacitive sensor exhibits a sensitivity of 1.95 kPa^-1 within 0-1 kPa, 1.06 kPa^-1 within 1-5 kPa, 0.88 kPa^-1 within 5-10 kPa, 0.52 kPa^-1 within 10-30 kPa, and 0.35 kPa^-1 within 30-50 kPa of pressure ranges. The hybrid response is fully understood through a simplified circuit model, which has been validated by the experimental measurements. We have successfully applied this sensor to measure subtle mechanophysiology on human body, including the pulse waves on the neck and the seismocardiogram on the chest.

There are two main strategies to develop stretchable conductors based on the materials used – nanocomposites with synthesized nanomaterials dispersed into a polymer matrix and deposited thin films exploiting mechanics guided structural design. Both strategies have seen remarkable progresses in developing highly conductive and stretchable conductors. However, so far the mechanical design strategy has been mainly applied to the deposited thin films. It would be interesting to exploit this strategy for the nanocomposites. Here we present two mechanical designs – delaminated buckle and horseshoe – for developing highly conductive and stretchable silver nanowire conductors. In the first design, a polymer composite film (with a thin layer of silver nanowires encased) was bonded to a pre-stretched elastomer substrate. Upon releasing the prestrain, a partially delaminated wavy geometry of the composite film is created. During the evolution of the buckle delamination, the blisters pop up randomly but self-adjust into a uniform distribution, which effectively reduces the local strain in the silver nanowires. The resistance change of the conductor is less than 3% with the applied strain up to 100%. In the second design, the polymer film was patterned into a horseshoe shape and bonded to an elastomer substrate. We found that the substrate thickness and bonding strength can dictate the deformation modes of the polymer film. For thin substrate and weak bonding, the resistance change of the conductor is less than 1% with the applied strain up to 100%.

An important problem in the field of haptics is the recording and playback (or similarly, composition and display) of texture. This problem is by no means unique to haptics: texture is a vital component of computer-generated graphics and it has been studied in the auditory domain as well (think of crickets chirping or wind blowing through leaves). Haptics, however, presents unique challenges.

Consider the problem of recording: what signals should be measured? While an argument can be made for measuring material and geometric properties of the textured surface, very little progress has been made toward realistic playback on the basis of such information. One reason is that the perception of texture requires contact and, typically, relative movement. This interaction causes the skin to deform whereupon various specialized mechanoreceptors capture aspects of that deformation including localized strain, skin stretch, and vibration. Arguably, the right signals to record are the same ones that the mechanoreceptors respond to. Indeed, measurements of normal and lateral force, movement speed, and skin vibration, have generally proved more useful than measurements of surface properties when it comes to playback. Playback, however, is limited in two additional respects: one, the ability of the haptic display to physically create spatiotemporal patterns of skin deformation similar to those caused by real world interaction; two, the ability of the display to control those patterns as a function of state variables such as finger pressure and speed.

In this talk, I’ll discuss the issues of recording and playback in the context of electroadhesive variable friction haptic displays. The talk will necessarily begin with a discussion of fingertip mechanics, including the roles of moisture and stratum corneum plasticization, in regulating friction; and then go on to a discussion of electroadhesion in which the material properties of both the skin and the countersurface are highly significant.

Electroadhesion is a technology that provides very good temporal control (bandwidth from DC up to several tens of kHz … well beyond the tactile range), but limited spatial resolution (the touch surface remains flat at all times and actuation is typically synchronous across the full fingerpad). This puts rather stringent limitations on playback, the perceptual consequences of which are sometimes surprising. I’ll review a number of studies we have performed to explore and expand the gamut of variable friction textures.

Ultimately, the gamut of an electroadhesive display, especially one built on a rigid surface, is limited. Properties such as hardness/softness, warmth/coolness, as well as coarseness (i.e., 2.5D shape), are unlikely to be realized without advances in the underlying display technology. While it is certainly possible that electroadhesion could be combined with other existing actuation methods (e.g., vibrotactile, pin arrays, thermoelectric, etc.) to expand the range of stimuli, the pathway toward an elegant, robust, and cost-effective solution is not evident. Instead, a pathway rooted in the development of novel “haptic materials” is more appealing. I will close this talk by discussing some of the characteristics such materials should possess.

Continuum robots are promising for medical interventions due to their ability to traverse around obstacles in confined spaces and conform to highly curved paths. In addition, their inherent structural compliance offers improved safety and adaptability, which is critical for interactions with sensitive structures inside the body. Real-time information about robot shape and applied forces is important for enabling more advanced control methods, ensuring safety, and creating new human-machine interfaces with haptic feedback. Yet the small size, and in particular, the highly limited space for routing wires through these robots, introduces a number of challenges for sensing and makes the integration of standard sensors difficult, especially if measurements along the entire robot length are required. To begin addressing these challenges, we are developing both wireless localization and force sensing methods compatible with continuum robots. This talk will first focus on a permanent-magnet based method for the localization of the tip of a growing robot (also known as a vine robot). The approach combines analytical and machine learning techniques to localize the robot tip in 5 degrees of freedom without requiring any significant modifications of the robot. The second part of the talk will focus on a new wireless force sensor based on backscatter tags. When a force is applied, the sensor deforms, creating contact between two conductive traces, which results in the reflection of electromagnetic waves on the edges of this contact area. The reflected waves will have a phase change that is dependent on the amount of contact between the traces, and can be used to determine the magnitude and location of the applied force. The development of these sensors is important to enable user interfaces with improved control and haptic feedback.

As eidetic signal recognition has become important, displaying mechanical signals visually has imposed huge demands for simple readability and without complex signal processing. Such visualization of mechanical signals has been used in delicate urgent medical or safety-related industries. Accordingly, chromic materials have been considered to facilitate the visualization with multiple colors and simple process, yet the response and recovery time is very long, so that rapid regular signals are unable to be detected, i.e., physiological signals, such as respiration. Here, we suggest the simple visualization of low strain ~2 %, with ultrasensitive crack-based strain sensors with a hierarchical thermochromic layer. The sensor shows a gradient color change from red to white color in each strain, which is attributed to the hierarchical property, and the thermal response (recovery) time is dramatically minimized within < 0.3 s, as the hierarchical membrane was inspired by termite mounds for efficient thermal management. The fast recovery property can be taken advantage of in medical fields, such as monitoring regular respiration, and the color changes can be delicately monitored with high accuracy by software on a mobile phone.

Recent developments in wearable devices brought technology much closer to the user's skin in the form of interactive fabrics, electronic skins, interactive tattoos, and more. However, these devices are typically only used for user input and do not provide haptic feedback in response to interactions. In fact, while we have seen many advances in the materials that make up wearable devices, the haptic feedback included in most wearables is, typically, based on off-the-shelf actuators such as linear resonant actuators that, due to their rigidity, cannot conform to the soft, flexible, and dynamic human body. To tackle this, we created a haptic device based on stretchable magnets. Our haptic device has two components: (1) a silicone skin doped with permanent magnetic micro powder (NdFeB) allowing it to display tunable magnetic polarity, and (2) a fingernail-worn device that includes an electromagnetic coil, 9-DOF IMU, and microcontroller. Users interact with our wearable haptic device by touching or dragging their finger over the magnetic silicone skin. Our device senses the proximity to the magnetic layer (via its IMU) and actuates the coil, causing the magnetic skin to oscillate—creating haptic vibrations on the user's skin. Using this mechanism, we can render a wide range of haptic sensations such as detents, button presses, or textures.
We believe our device is especially suited for on-body gestures, such as thumb-to-index pinching (popularized by Google's Soli or Microsoft's HoloLens, to cite a few) or to expand haptic output to different areas of the body where rigid devices are uncomfortable. Lastly, we studied our device’s dynamic performance via modal analysis and users’ perception of tactile feedback by means of psychophysical studies.

Merging human body with electronics and machines can enable internet of health (IoH), human-machine interface (HMI), as well as augmented human capabilities. However, bio-tissues are soft, curvilinear and dynamic whereas wafer-based electronics are hard, planar, and rigid. Over the past decade, stretchable electronics have emerged as a result of new materials development, structural design and manufacturing processes. In particular, epidermal electronics, a.k.a. electronic tattoos (e-tattoos), represent a class of stretchable circuits, sensors, and stimulators that are ultrathin, ultrasoft and skin-conformable. This talk will introduce a dry and freeform (subtractive) “cut-solder-paste” process for the rapid prototyping of modular and reconfigurable e-tattoos. This method is applicable for thin film metals, polymers, ceramics, as well as 2D materials and can easily incorporate integrated circuits to the stretchable e-tattoos. Electrically conductive open cell foams with hybrid piezoresistive and piezocapacitive responses has been applied as high-sensitivity mechanical transducer. The e-tattoos can laminate at different locations of the human skin for the synchronous and continuous tracking of electrophysiology (ECG, EEG, EMG, EOG), mechanophysiology (respiration, seismocardiogram, blood pressure, etc.), as well as thermophysiology etc.. In addition to sensing, the e-tattoos can also be used as personalized treatment and therapeutic devices. For wireless operation, we leverage near field communication (NFC) for wireless and Bluetooth low energy (BLE) for wireless data transfer. A wireless charging-on-the-go strategy enables the long-term (e.g., 3 days) operation of the e-tattoo.

Wearable haptics refers to systems worn on the body or integrated into clothing that can provide haptic feedback to the wearer in a wide range of forms, such as vibrotactile, cutaneous, skin stretch, pressure, and kinesthetic feedback. The potential for wearable haptics is supported by technological advancements in sensing and actuation that allow for the design and deployment of flexible materials, fabrics, and small-footprint actuators that can augment and enhance human performance. While rigid and grounded haptic feedback devices have been explored as a means to train new motor skills, rehabilitate movement coordination after neurological injury, and even augment human force output capabilities in tasks such as lifting or running, the application space is limited due to the weight and scale of such devices. Wearable haptic devices seek to overcome the limitations of rigid and grounded haptic devices.

Vibrotactile haptic feedback is ubiquitous for notifications and alerts in portable devices such as phones, smart watches, and fitness trackers, but the set of possible notifications is limited by the nature of the eccentric motors so often used to deliver such cues. In recent years, my group has been shifting to wearable haptic systems that provide multiple haptic modalities of feedback to the user, including vibration, skin stretch, pressure, and kinesthetic feedback, all packaged in bands that can be worn on the arm, or in hybrid rigid-soft exosuits that can apply forces directly to the limbs. We are particularly interested in applications where the provision of such feedback has a measurable impact on human performance.

In this talk, I will present some of our wearable haptic systems, describe our target applications, and highlight the distinguishing features our hardware while identifying opportunities for improvement. Two systems that I will discuss leverage multi-sensory haptic cues that are displayed directly on the user’s arm. Tasbi is a wearable haptic device that provides squeeze and vibration feedback for haptically enhanced augmented and virtual reality interactions. MISSIVE incorporates skin stretch, squeeze and vibration cues presented simultaneously to the user in distinct patterns. The use of multisensory cues allows us to design large discrete cue sets while maintaining a small and wearable form factor. With MISSIVE, we demonstrated language transmission via haptic phonemes, or units of sound encoded as haptic cues consisting of vibration, radial squeeze, and lateral skin stretch components. I will also present two wearable exosuit systems for upper limb movement assistance and rehabilitation, a glove, and an upper limb garment. Using cable-based actuation systems that enable remote placement of actuators, rigid spines precisely located for load transfer and cable management, and soft goods for compliant human-machine interfacing, these devices offer the potential for improved wearability compared to their rigid robotic counterparts, and feasible integration into the performance of everyday functional tasks.

Wearable haptic devices present many challenges in actuator selection and design. In addition to providing a sufficiently strong signal to be easily felt by the user, the actuators must also not encumber the user’s motion. Ideally the actuators should be small, thin, lightweight, flexible, low power, and provide multiple degrees of actuation. An actuator with all these properties is not currently available commercially, forcing haptic designers to make decisions regarding trade-offs between magnitude, resolution, degrees of freedom, and workspace among many others. These limitations motivate our research towards providing complex haptic signals using minimal hardware. We approach this problem from two directions by using haptic illusions and by creating novel actuators that can provide multiple haptic sensations simultaneously.

Haptic illusions seek to fool our sense of touch into thinking we are feeling something different than what is being displayed by taking advantage of the inexactness of our perception. In our research, we employ haptic illusions to allow small, lightweight actuators to create sensations that would normally require actuators that are impractical for a wearable device. I will discuss two haptic illusions and their uses: asymmetric vibrations for generating ungrounded forces and sequential indentation for creating lateral motions.

I will also present our work in creating multi-modal hardware, which allows for complex haptic sensations using minimal actuator size and weight. Our PATCH (Pump-Actuated Thermal Compression Haptic) device presents simultaneous thermal and pressure cues. The PATCH system is actuated by using pumps to fill fabric actuators with water of varying temperature. The actuators are lightweight, flexible, and provide a wide range of haptic cues.

Integrated wearable electronics capable of providing biophysical information on complex and dynamic systems have attracted interests in diverse fields such as healthcare and robotics. Gallium and gallium-based liquid metals stand out for their excellent combination of electrical conductivity and mechanical deformability. However, liquid metals feature complex physical and chemical properties that pose challenges to the fabrication of miniaturized soft and stretchable circuits. Herein, we present a manufacturing approach to form ultrathin (1.5 µm) and high-resolution (2.2 µm) liquid metal features with high surface density over large surface areas (wafer scale). Based on a combination of soft lithography, selective wetting, Ion Beam Etching and thermal evaporation of gallium, the process enables a range of channel designs and geometries with unprecedented aspect ratios (> 10'000) that can be used to form liquid metal based stretchable electronic conductors with unrivaled transparency.
We demonstrate conducting grids of w = 2.2 µm width and variable pitch (p = 100.2, 50.2, 29.2, 15.2 µm) that can reversibly stretch to 100 %. Adjusting the density of the narrow lines across the PDMS substrate offer tunability in the patterned metallic grid. In the relaxed state, the liquid metal grids display optical transmittances at 550 nm wavelength of 69.3 %, 80.2 %, 86.9 %, and 89.6 % for p = 15.2 µm, 29.2 µm, 50.2 µm, 100.2 µm with sheet resistance of 4.8 Ohm/sq., 6.4 Ohm/sq., 8.8 Ohm/sq., 12.5 Ohm/sq. respectively. The accuracy and the robustness of the patterning technique are demonstrated by stretching the most transparent structure (p = 100.2 µm) to 100 % deformation. The combination of high electrical conductivity (12.5 Ohm/sq.), optical transparency (89.6 %) and mechanical deformability (100 %) ranks these soft and stretchable liquid metal grids amongst the best performing transparent conductors reported to date.

Piezoelectric materials are suitable for haptic technology as they can convert mechanical stimuli into electrical signals and vice-versa. However, owing to their disadvantageous mechanical properties such as brittleness (in ceramics) and a low piezoelectric coefficient (in polymers), it is still a challenge to use them in haptic technology. Here, we present a truss-like 3D hollow zinc oxide (ZnO) nanostructure that shows a drastically improved elastic strain limit while maintaining a piezoelectric coefficient similar to that of single-crystal ZnO. The ZnO hollow nanostructures were fabricated by proximity field nanopatterning (PnP) and atomic layered deposition (ALD) at four different process temperatures from 90 to 300°C. The piezoelectric properties were visualized and analyzed through dual AC resonance tracking piezoresponse force microscopy (PFM). The piezoelectric coefficient was approximately 9.2 pm/V. Furthermore, we measured the measured elastic strain limit of approximately 10% using the nanopillar compression test, which was at least 3 times greater than the prior report. The extended elastic limit of the 3D hollow structure was further supported by finite element simulations. We envision that our ZnO hollow nanostructure can be used in enhanced haptic devices, which mimic the human sense of touch.

Degradable semi-crystalline polymers are a class of materials that modern regenerative medicine heavily relies on. While their formulation into nanostructured devices can be advantageous to tailor functional properties such as cellular uptake kinetics and degradation behaviour, it is difficult to assess and control the morphology, crystallinity and organization of nanostructured polymer materials by common preparation methods.

The Langmuir technique overcomes most of the drawbacks concerning preparation and investigation of nanostructured polymer materials. Here, ultrathin films at the air-liquid interface serve as model systems for polymer devices and interfaces, which can be manipulated and observed in a straightforward way. In particular, the semi-crystalline morphology, including crystal size, number density, thickness, and melting temperature can be precisely adjusted. By transferring the layers to solid substrates, functional biomaterial interfaces with unique properties can be created. For example, via enzymatic degradation of the amorphous chains, a semi-crystalline Langmuir film can converted into an almost entirely crystalline film with a thickness of a few nanometers. The preparation and analysis of these highly functional, entirely crystalline thin films based on oligo(ε-caprolactone)s (OCL) with different end-groups is of particular interest. Poly(ε-caprolactone) (PCL) is established in clinical applications with a relatively low melting point point and well documented enzymatic degradability¹. The melting temperature of PCL or OCL crystals can be adjusted via the crystallization temperature to be very close to physiological conditions, producing a polymer coating that changes a broad variety of its functional properties upon application of a small thermal stimulus.
For example, highly crystalline films of hydroxyl and methacrylate end capped linear OCL are prepared by degrading its amorphous phase by the enzyme Pseudomonas cepacia lipase (0.007 mg_^.mL^-1) at constant surface pressure of ~7 mN^.m^-1 on air-water interface at room temperature. The thickness of the OCL crystals in the film can be adjusted in the range of ~ 6 to 8 nm, resulting in a melting temperature between ~30 to 40 °C at the air-water interface. Upon melting, the elastic moduli of the film are drastically altered. Further, functional end groups are expelled from the crystals and therefore accumulated at the crystalline surface. Upon melting, these end-groups are buried in the amorphous phase, resulting in a change of the chemical composition of the material surface. Moreover, the degradability of the thin film increases significantly upon melting, meaning that the lifetime of the material can be adjusted via its thermal history. In addition, the barrier properties of the thin film depend on its crystallinity, with polymer crystals being considered impenetrable for small molecules.

Here, we demonstrate the fabrication and behaviour of these thermally switchable OCL crystals coatings, which are of interest for applications such as barrier layers, substrates modulating stem cell adhesion, or coatings with on-demand acceleration of degradability. Another possible application are friction reducing coatings, for example on sutures. Here, the temperature at the suture-tissue interface does not increase as long as the heat generated by frictional forces is absorbed by the melting enthalpy of the crystals.

The correct balance of forces on the hip joint during total hip arthroplasty (THA) is essential for implant longevity and to prevent the need for revision surgery. This has become increasingly important since the average age of a THA patient is decreasing, hence these implants must survive greater activity levels and correspondingly higher stresses than those typically found in older, less active patients. Currently, the lifetime of an implant positioned purely by a surgeon's judgement is approximately 15 years. Quantitative force feedback during surgery could provide more accurate implant positioning, increasing the average lifetime. However, there are currently no force sensors capable of providing this information within the small and complex geometry of the hip joint.

Here, we solve this unmet clinical need by presenting a thin, conformable and flexible microfluidic capacitive force sensor, which can be incorporated into the surgical process without significant modification to existing implants. The sensor consists of a deformable microfluidic chip bonded to a flexible substrate, on which interdigitated electrodes are incorporated, and that can operate in confined spaces to give accurate force readings from joints. The sensor operates by monitoring the change in capacitance of the electrodes as the fluid in the microfluidic channel is displaced in response to an applied force. Multiple sensors were incorporated into a model hip implant geometry, and a bespoke mechanical testing rig was built to simulate loading of the acetabular cup from the femoral head at different angles. The experimental results were validated using finite element modelling. This was used to optimise the sensor geometry for the production of sensitive devices that can withstand the high forces required for surgery. Our novel microfluidic-based force sensing technology can potentially be adapted for application in joints either for sensor-assisted orthopaedic surgery, or for gathering telemetric data from joints before, during and/or after surgery. Such biocompatible and implantable force sensors will contribute significantly to several clinical drivers, through providing quantitative feedback to the attending clinician, and could also serve to underpin robot-assisted orthopaedic surgery in the future.