Symposium SB03—Robotic Materials for Advanced Machine Intelligence

Polymeric materials possessing stimuli-sensitivity can identify changes in their environment and respond by variations on their properties. Shape-memory polymers belong to this group of materials. The choice of input signals for controlling the shape-memory effect is no longer limited to temperature changes. Developments in synthetic polymer chemistry have led to the incorporation of a diverse range of sensitive molecular switches and were closely followed by materials enabling macroscopic shape shifts induced by a wide variety of stimuli^[1]. Designing criteria for materials that respond to their external environment can also be found on different length-scales of the material organization. Here, principles from nature can be inspiring. Plants have evolved many capabilities to anchor, position their stems and leaves favourably, climb and adapt themselves to different environments by growing and moving. A bio-inspired multi-material system will be presented as an example, which is capable of controlled movements by mimicking the cactus Selenicereus setaceus with is specific geometries and combinations of materials with differing mechanical properties. Its unique star-shaped geometry is 3D printed using a soft elastomer, which is then combined with a hydrogel providing the actuation capability to the artificial cactus system. Through anisotropic swelling, the hydrogel-elastomer system adjusts its configuration and shows a controlled movement^[2]. The Venus flytrap (Dionaea muscipula) represents a finely tuned mechanical bi-stable system. A design approach effectively imitating this mechanism was the starting point for a synthetic material system, which can switch between mono- and bi-stability upon exposure to heat.^[3]

References
[1] A. Lendlein, M. Balk, N. A. Tarazona, O. E. C. Gould, Biomacromolecules 20, 3627-3640 (2019).
[2] A.K. Bastola, N. Rodriguez, M. Behl, P. Soffiatti, N.P. Rowe, A. Lendlein, Mater Design 202, 109515 (2021).
[3] J. Bäckemo, Y. Liu, A. Lendlein, MRS Comm 11, 462-469 (2021).

Conductive redox polymers that change shape in response to chemical reactions can serve as actuators and provide great utility in a variety of technologies, including biomimetic artificial muscles and robotics. Their polymeric nature facilitates direct robot-human interfaces, while the redox-based actuation allows for self-lock in actuation. Despite the benefits of the state-of-the-art redox actuators as artificial muscles, their demonstrated actuation mechanisms, such as the need for liquid electrolytes, external driving voltages and/or extreme environmental pH conditions, poses major challenges for many applications. Herein, we introduce for the first time an all-solid and electrolyte-free redox-induced polymer actuator, which generates motion solely in gas environments. We employed wet-spun polyaniline (PANI), which provides scalable, long, conductive, and flexible micro-size fibers. Catalyst (Pt) coated PANI fibers were exposed to H₂ and O₂ environments, leading to actuation behavior. The actuation mechanism was studied and it was concluded that configurational changes occur in PANI due to H₂ being oxidized to H⁺ in the presence of the catalyst on the fiber surface which then reduced the PANI fiber, generating a volume contraction. On the other hand, O₂ oxidized the polymer by reacting with the reduced PANI in the presence of the catalyst, leading to a volume expansion.
The fibers contracted in H₂ and expanded in O₂ with an actuation magnitude of 3.5% under 1 MPa load. The maximum achieved load was staggeringly high for artificial muscles, a value of 12 MPa (compared to <1 MPa strength of human muscle) with 0.5% actuation. The maximum work capacity the PANI fibers generated was 120 J/kg. The redox-triggered actuation was proven by a change in electrical resistance and polymer color. No temperature change was detected and no external voltage was applied. The gas-driven actuators developed in this work create new avenues for the application of PANI and other redox polymers in soft robotics.

Spiders possess fascinating multi-feats with spiderweb for maximizing adhesion force. Here, we explore an approach to behavioral characteristics of cyclosa japonicas inspired ionic spider webs (ISWs), which is capturable, self-powered sensible and vibrationally dust-cleanable, all of which the principles are solely based on electrostaticity in a single platform [1]. It was made with ionically conducting organogels capsulated with dielectric elastomers as a strand shape. Its thinness and transmittance led ISWs could be camouflageable. Plasma etching and (heptadecafluoro-1,1,2,2-tetrahydrodecyl)trichlorosilane treatment on dielectric elastomer readily boosted cleanability and it even prevented the evaporation of organogel, which led increase of durability of ionic spider web. Its high stretchability and translucency led it endure applied strain upto 50 % to capture targets larger than ISWs, which cannot be realized by rigid one, and realized it can be camouflageable, respectively. The ISW robustly captured a wide range of materials including a leaf, acryl, glass, and aluminum of which weight is 50 times heavier than its own weight. The capturability was ensured with its additional feats including electrostatic sensing ability and resonance cleanability.

References
[1] Lee, Y., Song, W. J., Jung, Y., Yoo, H., Kim, M. Y., Kim, H. Y., & Sun, J. Y. (2020). Ionic spiderwebs. Science Robotics, 5(44).

Creating spatially distributed deformations is a unique feature of soft actuators which allow them to reversibly morph their shape, enabling a variety of novel applications. Dielectric elastomer actuators (DEAs), that are among the fastest and most energy-efficient soft actuators, can morph their shapes using one or both of two mechanisms: one creates spatially distributed actuation magnitude using a distribution of internal electric fields and the other creates spatially distributed actuation direction determined by a pattern of stiffening elements, such as produced by 3D-printing. Although the actuation shapes can be analyzed for a given configuration of electrodes and stiffening elements (the forward problem), the design of the actuator for morphing into a target shape upon applying the voltage (the inverse problem) has not been fully addressed. In this presentation we introduce a simple analytical solution for the inverse problem to design actuators that will morph into specific target shapes. It is based on combining the two shape-morphing mechanisms to control the actuation magnitude and direction locally. This enables the generation of actuator designs for morphing into any shapes defined by z=f(x,y) using a simple analytical solution. In this manner, the pattern of number of active layers and stiff rings and required electric field can be determined in order to morph to a desired target shape. To illustrate our approach to the inverse problem, DEA actuators that morph into simple target shapes, including cone, cosine, and sine, are designed and fabricated. Their actuation surface profile measurements are compared to the target shapes. Our approach is further illustrated by morphing a flat sheet into complex, non-symmetrical shapes, such as a face.

Direct ink writing of liquid crystal elastomers (LCEs) offers a new opportunity to program geometries for a wide variety of shape transformations. However, the degree of shape change is highly dependent on the underlying heating stage: once out of the hot stage, the shape tends to return back to the original stage. Therefore, the ability to spatially actuate the 3D printed LCEs in free-forms is highly desired. Herein, we develop a 3D-printable gold nanorod (AuNR)/LCE composite ink, allowing for photothermal actuation of the printed filaments even at a low AuNR concentration (0.1 wt%). By fine-tuning the printing pressure, speed, and temperature, as well as the ink formulation, we show that the printed filament has superior photothermal response with 27 % actuation strain upon irradiation to near-infrared (NIR) light (808 nm) at 1.4 W/cm², corresponding to 160 °C. As a result, each 3D printed AuNR/LCE composite lattice can be actuated into different shapes with high amplitudes by locally control of its exposure to NIR laser.

Herein, we present a new class of liquid crystalline elastomers (LCEs) crosslinked with poly(ether-thiourea) comprising a triethylene glycol spacer (LCE-TUEG) wherein thiourea bonds impart a hydrogen bonding capability as well as permit dynamic covalent bond (DCB) exchange at elevated temperatures. While hydrogen bonding enhances the mechanical properties of LCE-TUEG, the DCB allows the network rearrangement of the LCEs, generating various useful properties that are not present in conventional LCEs, including the ability of welding, melt and solution reprocessing, reprogrammable actuation, and self-healing. By harnessing such dynamic features, we fabricate electrically-driven artificial muscles that can be actuated by Joule heating using a resistive wire. In particular, an excellent specific work is demonstrated (~61 J kg⁻¹) for the artificial muscle, and full recyclability of both the LCE matrix and the metallic heating wire is achieved. Furthermore, a biomimetic artificial hand is created by welding and assembling multiple LCE-TUEG films embedded with heating wires, followed by mechanical alignment. The integration of a microcontroller to the artificial hand enables the selective actuation of each finger, and various hand gestures are successfully demonstrated.

One of the important goals in soft robotics is to build a system that can carry out the whole function of living organisms, spanning their birth, growth, and death. Among these, the death of soft robots is nowadays in great interest due to the growing concern towards the disposal of soft robots. An on-demand disintegration of soft robots can figure this out, but materials up-to-date are unsuitable for such functions. Conventional elastomeric materials used in soft robot frames (e.g. silicone elastomers) are extremely difficult to decompose, and on-demand transient materials so far (e.g. cyclic poly(phthalaldehyde)) possess insufficient mechanical properties to be used in soft robots. Here we introduce an on-demand transient silicone elastomer composite as an answer to allow this on-demand death of soft robots. A composite of silicone elastomer matrix and photo-fluoride generator additives degrades into a liquid state via a two-step trigger of ultraviolet radiation and heating at ~100 °C. The composite shows comparable mechanical properties to commercial silicone elastomers, retaining a modulus of 40 kPa and a fracture strain up to ~500%. On-demand transient soft robots and electronics which can degrade under the same stimuli are demonstrated using this material. A robot that self-diagnoses its destruction using embedded ultraviolet and temperature sensors was fabricated. The robot avoids the risk of destruction and carries out the desired function in peacetime, but rapidly degrades upon request by taking sufficient ultraviolet light and heat. Transformable robots that change their functions and actuation aspects upon selective on-demand transience are presented too, showing that on-demand transience can empower novel functions to soft robots. The concept reported in this research presents the domain of on-demand transient soft robotics, which is expected to bring innovation in environmental, medical, and security-related issues.

Researchers have continued to develop strategies for environmental adaptation and autonomy in soft robots with ever-increasing capabilities. Historically, robots have achieved the ability to interact with their external environment via feedback loops involving multiple sensors, microcontrollers, and actuators constructed from rigid electronic components that can be bulky, complex, and mechanically incompatible with soft materials. In this work, we create a strategy for achieving autonomous path-planning in soft robots, effectively embodying the control system and its feedback loop in the compositional and structural features of the soft robot body. By including responsive materials such as liquid crystal elastomers and distributing these suitably throughout the soft body of the robot, the system can sense and respond to different external stimuli (light, heat, solvent) to produce versatile and autonomous changes to locomotion in response to different environments. Distinct control strategies can be achieved in the same robot by rearrangement of modular structural features. Finally, we explore strategies for inverse design. This work provides a new strategy for achieving fully autonomous robots for operation in uncertain environments.

An ultrasmall soft gripper is a device that allows small objects, organisms, and cells to be held firmly or moved to a desired location. Furthermore, the measurement of signals arising from a grasped object is highly critical for identifying or predicting the object's past, present, and future states. Inspired by the human hand, we now report a new type of miniaturized soft gripper with variable stiffness consisting of silver nanowires and a shape-memory polymer. The developed soft gripper can reproduce the selective moved fingers of a human hand by ablation of silver nanowires and is designed to make strong conformal contact without any damage by using the variable stiffness of the shape-memory polymer to match the gripper stiffness to that of organisms being held. The demonstration includes gripping a snail egg with appropriate temperature stimulation for the hatching process and monitoring the heart rate (heart size: 300 ~ 600 μm) of a newly hatched snail. The results display the soft gripper has potential in a closed-loop system as a medical device.

Natural living systems such as wood frogs develop tissues composed of active hydrogels with cryoprotectants to survive in cold environments. Recently, hydrogels have been intensively studied to develop stretchable electronics for wearables and soft robots. However, regular hydrogels are inevitably frozen at the subzero temperature and easily dehydrated, and have weak surface adhesion. Herein, a novel hydrogel-based ionic skin (iSkin) capable of strain sensing is demonstrated with high toughness, high stretchability, excellent ambient stability, superior anti-freezing capability, and strong surface adhesion. The iSkin consists of a piece of ionically and covalently cross-linked tough hydrogel with a thin bioadhesive layer. With the addition of biocompatible cryoprotectant and electrolyte, the iSkin shows good conductivity in wide ranges of relative humidity (15–90%) and temperature (−95–25 °C). In addition, the iSkin can adhere firmly to diverse material surfaces under different conditions, including cloth fabric, skin, and elastomers, in both dry and wet conditions, at subzero temperature, and/or with dynamic movement. The iSkin is demonstrated for applications including strain sensing on both human body and winter coat, human–machine interaction, motion/deformation sensing on a soft gripper and a soft robot at extremely cold conditions. This work provides a new paradigm for developing high-performance artificial skins for wearable sensing and soft robotics.

Torque and rotation have been fundamental to robotics since its inception because of the ubiquity and utility of motors as actuators. Recent advances in architected auxetic materials have demonstrated how torque and rotation can drive transformations in the shape of materials. These transforming materials have formed the basis of robot bodies, arms and grippers. We have discovered new families of auxetic material architectures based on axial point groups. These new families of structures have properties mediated through rotation such as bistable volumetric expansion and torque transmission. Through composition into larger spatial symmetries, we can accentuate or suppress these properties. The results are materials with switchable stiffness and volume, stretchable soft transmissions that can transfer multiple coaxial torques, and non-reciprocal torque responsive materials. Each of these structures are complaint by nature and can be fabricated from a wide variety of processes. These new material architectures provide critical components in future robotic systems and point to a future in which we bridge the worlds of traditional and materials robotics

Biological organisms are rich with the ability to reconfigure their shape and properties to perform diverse tasks. These capabilities are inspiring emerging soft robots that have the potential to display biological capabilities through advanced machine intelligence. However, soft machines struggle to achieve multiple complex configurations, quickly lock into shape to support loads, and go between multiple states reversibly. In this presentation we will introduce a multifunctional shape morphing material with reversible, rapid, and lockable polymorphic reconfigurability. We couple elastomeric kirigami with an unconventional reversible plasticity mechanism in metal alloys to rapidly (< 0.1 s) lock-in complex, load bearing shapes, with reversibility and self-healing through phase-change. Large deformations and complex morphologies are demonstrated where we show the ability to exhibit positive, zero, and negative Gaussian curvatures and conformally wrap complex surfaces. This approach overcomes trade-offs in deformability and load bearing capacity while eliminating power requirements to lock in shapes. This combination of properties is enabling for creating multifunctional soft robots. We demonstrate this material through integration with onboard control, motors, and power, to create a soft robotic morphing drone which autonomously transforms from a ground to air vehicle and an underwater morphing machine which can be reversibly deployed to collect cargo.

Ideally soft machines could instantly take on any arbitrary shape, and then hold that shape with zero power consumption and with a high blocking force. Soft materials however often imply low load-bearing abilities, and multi-DOF shape reconfiguration is generally achieved using many actuators, complicating design and control.
I present three solutions developed in my lab towards reconfigurable actuators, combining soft actuators and phase change polymers. In the first two device, the shape holding, but also the direction of the shape change, is dynamically set using a film of shape memory polymer (SMP) covered with addressable heaters, that enable local reversible change of stiffness. By layering this SMP film with a single large dielectric elastomer actuator, we can obtain complex shapes and much higher blocking force than what the DEA can deliver. By combining the addressable SMP film with liquid metal coils in an elastomer substrate, we created a multi-segment finger where each joint can be independently bent or twisted using electromagnetic forces, and then latched into place, requiring no power to hold any given position.
The SMP devices offer high holding forces, but require up to tens of seconds to cool. The third type of soft actuator is based on electroadhesion, allowing for millisecond scale response, shear forces of over 100 kPa, and simple adaptation to complex shapes. I will contrast the performance limitations of the above-mentioned soft actuators, and illustrate use cases in grippers and manipulation.

Liquid crystal elastomers (LCEs) have attracted significant attention in soft robotics due to their salient features of programmable and reversible shape morphing with large strain. To enable untethered soft robotic actuators, optical stimuli are particularly promising for wireless and spatiotemporal operation capabilities. To consider practical performances and operating environments for soft robotics, actuators should be capable of linear actuations from dry to wet conditions. Here, we present a light-driven soft actuator by structural designing of azobenzene-functionalized LCEs (A-LCEs). The A-LCEs demonstrate a large reversible axial stoke up to 50% under wet environment, thus, provide a promising approach in underwater soft robotic applications.

Liquid crystalline elastomers (LCEs) are commonly used as soft actuators for applications such as robotics but are currently limited by a weak coupling of stimuli and response that restricts these materials for functional applications. This is evident in the extended deformation-temperature response of LCEs that is an inefficient use of energy. Fundamentally, LCEs exhibit continuous (2^nd order) phase transitions between nematic and isotropic (paranematic). Motivated by this fundamental limitation of these materials, this effort attempts to approach quasi-1^st order (subcritical) phase transitions in LCEs. Here, we introduce supramolecular hydrogen bonds in the mesogens of the polymer network that dissociate upon heating. The efficiency of phase transitions in LCEs is currently limited by the retention of order in the paranematic phase due to the retainment of crosslinks. By introducing non-covalent bonds using benzoic acid precursors as dimers within crosslinks, the polymer network can be disrupted more rapidly than seen in prior reports. Here, we also introduce liquid crystalline mesogens with weaker intermolecular interactions to combine lower, more accessible transition temperatures with sharper transitions. Overall, the purpose of this effort is to lead to improved performance of these functional materials as rapid soft actuators for robotics applications.

Ingestible capsule robots with pneumatic actuations show safe tissue interaction for their environment adaptability. However, conventional pneumatic actuators need bulky air reservoirs and air valves to realize programmed actuation. And wires that provide power and actuation limit its mobility and accessibility in confined space, and further restrict its in vivo application. Liquid-vapor phase transition provides temperature-controlled volume change, which is a promising solution for the reversible and wireless pneumatic actuation.
Inspired by pufferfish, here, we design a wireless capsule robot integrated with a soft balloon actuated by liquid-vapor phase transition. Specifically, the capsule robot composes of receiver coils with ferrite core, and a soft balloon loaded with liquid–vapor phase change materials (LVPCMs). During operation, a wireless power transmitter generates alternative magnetic fields at the resonant frequency of receiver coils, and the receiver coils convert the induced current into Joule heating. When heated up above boiling temperature, the LVPCM changes from the liquid phase into the vapor phase. Then the soft balloon made of a kind of soft elastomer (Ecoflex) is inflated due to the large volume change.
The performance of wireless pneumatic actuation depends on the Joule heat generated by wireless magnetic resonant power transfer and the boiling temperature of the LVPCMs. Firstly, power is the bottle neck for the wireless actuation. The wireless power transfer energy of receiver coils with different configurations are measured. And the corresponding Joule heat is recorded. Furthermore, typical solvents with various boiling temperature are tested. The inflation behavior and temperature response are characterized. In addition, the relationship between force with volume change is evaluated. Finally, the capsule robot is integrated with a 3D printed microneedle to demonstrate its ability of drug delivery in an in vitro experiment.

Acknowledgement:
The authors would like to thank the support from IDEATION programme and Multi-scale Medical Robotics Center (MRC), InnoHK at the Hong Kong Science Park.

In this work the material parameter identification, simulations and experimental verification of the new smart material Vacuum Packed Particles is presented. The considered material is a structure composed of granular material inside the plastomer coating. When the pressure inside is equal or higher than atmospheric the system behaves like a liquid, otherwise like an elasto-plastic solid with a temperature and strain rate dependence. By changing the underpressure inside the coating it is possible to control the mechanical properties of the structure in a real time. Addinionally materials allow to change the shape. In this work, the machine learning approach to model parameter estimation is presented. Additionally, the proposed model is implemented into Finite Element Method code. In order to verify the simulation results a corresponding experimental tests were conducted using optical strain measurement system. The comparison of the outcomes showed a good correlation.

Soft robotics is a growing field which offers new avenues to exploit material properties. Soft robots composed of materials with the intrinsic ability to self-heal can extend the functional lifetime by fully or partially repairing damage with little or no external assistance. Self-healing materials also open the possibility to reconfigure a soft robot with minimal effort during its operational lifetime. There are many chemical pathways for self-healing to occur. We have investigated mechanical properties of several 3D printable, photopolymerizable resins exhibiting these self-healing chemistry pathways. Our elastomeric hydrogels with self-healing ability make resins which are well suited to be 3D DLP printed into pneumatic actuators and incorporated into touch sensors.

Two of the most crucial topics in modern medicine are drug administration routes and their relative pharmacokinetics inside human body. Virtually all conventional delivery methodologies are based on the non-selective distribution of the active substance in the whole body, mediated by blood circulation. Consequently, a relevant fraction of the drug reaches non-target parts of the body, lowering thus administration efficacy and amplifying negative side-effects. To overcome these problematics, advanced administration strategies based on targeted delivery have been recently developed [1]. In these approaches, drugs are released in controlled amounts by specifically tailored smart materials. A promising class of such materials is constituted by functionalized hydrogels. Release from these biocompatible materials can be controlled through a variety of approaches [2]. For example, smart hydrogels can be tailored to release drugs only under specific conditions of pH, temperature ... In this way, release rate and timing can be efficiently tuned by controlling the structure and the chemistry of the hydrogels.

On the other side, a possible key to improve spatial control over release can be the use of magnetically controlled microrobots [3]. These remotely controlled devices are able to perform different tasks in-vivo, including for example cell transport [4] and medicine delivery applications [5]. For the latter, they can be coated with drug releasing materials and wirelessly guided inside human body to perform administration only in close proximity of the target organ. Moreover, magnetic field is harmless for humans, allowing a limited invasivity of the microrobots in conjunction with a great manipulation precision.

In this context, we describe the realization of magnetically guidable microdevices integrating smart hydrogel layers specifically tailored to perform controlled drug release. The microdevices are obtained employing additive manufacturing, specifically microstereolithography, in combination with wet metallization. This approach is highly scalable and flexible and can yield micrometric sized objects at relatively low cost. To allow actuation, a magnetic layer is applied by means of wet metallization on the 3D printed devices. The same technique is employed also to deposit a gold layer to make the surface biocompatible. Finally, the surface is coated with the hydrogels and drug release performances are evaluated in-vitro. Following this manufacturing strategy, we are able to obtain hard-soft hybrid microrobots able to combine the high magnetic properties typical of ferromagnetic alloys with the shape customizability typical of 3D printing and the drug releasing properties of hydrogels.

Two different strategies are investigated to control the drug release from the hydrogel. In the first, an alginate hydrogel is modified with click chemistry, binding the drug to the biopolymer chains by mean of a pH cleavable bond. In this way, release takes place only when the device reaches the part of the body presenting the correct pH range. An example of possible application of this method may be drug release in well-defined zones of the gastrointestinal apparatus, which is characterized by different pH levels according to the tract considered. In the second approach, two different hydrogel couples (alginate and chitosan or alginate and poly(allylamine) hydrochloride) are sequentially deposited following a layer-by-layer technique. In this way, drug release can be tuned by increasing the diffusion path in the material. This approach, even though not targeted as the first, is less complicate and more applicable to different types of drugs.

[1] M.W. Tibbitt et al., J. Am. Chem. Soc. 138, 704 (2016)
[2] K.S. Soppimath et al., Drug Dev. Ind. Pharm. 28(8), 957 (2002)
[3] J.J. Abbott, IEEE Robot. Autom. Mag. 14, 92 (2007)
[4] R. Bernasconi et al., Mater. Horiz. 5, 699 (2018)
[5] S. Fusco, Adv. Healthcare Mater. 2(7), 1037 (2013)

The cholesteric liquid crystalline (CLC) phase self-organizes into a helicoidal structure. The periodic nature of the hierarchical organization of these materials produces a 1-D photonic bandgap where periodicity in refractive index causes the CLC phase to exhibit selective reflection. Although the CLC phase has been prepared in elastomers, the synthesis procedures resulted in poor consistency and significant haze. Previously, we have reported on the use of surface-enforced alignment to prepare CLCEs¹ with high optical quality (haze < 0.5%) and large thermochromic response (> 200 nm tuning). Here, we explore electrical reconfiguration of CLCEs as dielectric elastomer actuators (DEAs). Application of an electric field causes significant change in the reflection wavelength of fully solid CLCE. The electromechanical response (Maxwell stresses) are reversible. Various chemistries to prepare the CLCEs were explored in our research to determine composition-property relationships in these materials. Electrical control of optical properties be utilized as responsive skins or logic gates in implementations in soft robotics.

[1] Brannum, M., et al., Adv. Optical Mater., 2019, 7, 1801683.

In this study, we developed an artificial intelligence (AI) exo-glove that changes the knuckle stiffnesses quickly and accurately in harmony with user’s intention, and thus, assists or controls hand gestures. Electrostiction phenomenon based on Maxwell stresses in a high dielectric film (HDF) can produce high attractive force between two HDF/electrode/substrate layers. The high attractive force in turn can induce high frictional force between the two contacting layers, and thus, can change the linear and rotational stiffnesses between two linked layers. When a poly(vinylidene fluoride-trifluoroethylene-chlorotrifluoroethylene) [P(VDF-TrFE-CTFE)] thin film of thickness 10 μm is used as the high dielectric films, a frictional shear stress of 27.92 N/cm² is induced under applied voltage of 250 V. Based on electrostiction mechanism, an exo-glove composed of multiple multilayered links are developed in this study, which fit the outer shape of five fingers and freely move together with hand gesture.
In order to perceive human’s intention and cooperate the exo-glove with human’s hand gesture, 8-channel fabric-type electromyography (EMG) sensors are used. EMG signals collected from a user in real-time through the EMG sensors are analyzed and classified into several gesture patterns with the support vector machine (SVM) algorithm. This classification (decision) is converted into voltage signals that control the multiple multilayered links in the exo-glove and therefore control the hand gesture by changing the stiffnesses at high speed in harmony with one’s intention. As a demonstration of the AI exo-glove developed in this study, two persons (one (A) wears the AI exo-glove, and the other (b) wears the EMG sensors) perform rock-paper-scissors games, and the one (A) wearing the AI exo-glove can mostly win the rock-paper-scissors games by aid of the machine learning algorithm and exo-glove. This demonstration shows how an exo-glove and also exo-suit operated by electrostiction mechanism can cooperate with a user to assist his or her gesture and motion by using a machine learning algorithm.

Human–robot collaboration has recently attracted considerable interest in manufacturing engineering to increase productivity and efficiency [1]. However, workspace sharing between humans and robots requires novel types of safety systems to minimize injuries resulting from human–robot collision. To ensure the protection of human workers, various types of robot safety skins have been developed for collision detection and shock absorption [2–4]. Although many researchers have designed innovative robot safety skins based on the piezoresistive or capacitive change in viscoelastic materials and the pressure change in air bag structures, the limited sensitivity and stretchability of these skins are far from complete [2–4].
In this paper, we propose a sensitive and stretchable robot safety skin based on the charge accumulation characteristics of polyvinyl chloride (PVC) gel to resolve the existing limitations. The PVC gel has distinctive mechanical and electrical characteristics. It has high stretchability and remarkable damping property due to its plasticized polymer chain structure [5–6]. Thus, the proposed robot safety skin can be applied to robot joints, which typically sustain considerable deformation, with an excellent shock absorption ability. In addition, when an electric field is applied to the PVC gel, the plasticizer inside the gel accumulates toward the anode to form a solvent-rich (SR) layer. Hence, the space charge is concentrated in the SR layer. [5–6]. The mechanical pressure exerted on the sensor leads to the structural deformation of the SR layer, resulting in a substantial change in the voltage distribution inside the PVC gel. By measuring this voltage change using signal electrodes, sensitive pressure detection can be achieved.
Liquid metal (LM) electrode is utilized to apply and measure the electrical signal to the stretchable PVC gel. The electrical resistance of the LM electrode is constant under a considerable strain, compared to the solid-phase conductive nanomaterials [7]. Since the resistance change of electrode affects the output of the sensor, the LM is the appropriate material for the stretchable skin sensor. To enhance the wettability to the PVC gel, the micro/nano droplets of LM are fabricated and dispersed in water [8-9]. During the ink preparation, we apply oxide-removing and coating-enhancing agents to make a sintering-free LM ink. The sintering-free LM electrode will be printed on the PVC gel without any post-procesing to retain own characteristics of the PVC gel.
The newly-developed robot safety skin has a wide pressure sensing range up to 122 kPa to estimate a large impact force. Especially, the proposed skin sensor exhibits a sensitivity of 0.031 kPa^-1 in the range of pressure level, 0 to 4.5 kPa. It is 9.4 times higher than the average sensitivity in other sensing ranges. It means that the skin can sensitively detect the very initial contact of collision between humans and robots. In addition, the skin sensor can endure a large strain more than 50% due to the high stretchability of PVC gel and LM. We believe that the robot safety skins will be used in real industrial robots soon.

This work has been supported by the Industrial Strategic Technology Development Program (No. 20008952) funded by the Ministry of Trade, Industry, & Energy (MOTIE, Korea), the R&D Program (No.2019R1A2C2006362) of National Research Foundation (NRF), Korea, and the Basic Science Research Program of National Research Foundation (NRF), Korea (NRF-2021R1A2C2007835) funded by the Ministry of Science.

[1] S. Robla-Gomez et al., IEEE Access 5, (2017).
[2] T. Kim et al., IEEE Robot. Autom. Lett. 3, (2018).
[3] S. Tsuji et al., IEEE Sens. J. 20, (2020).
[4] R. Weitschat et al, IEEE Int. Conf. Robot. Autom, (2017)
[5] Y. Li et al., J. Mater. Chem. C 7, (2019).
[6] T. Hwang et al., Sci. Rep. 9, (2019).
[7] K. Park et al., Bull. APS, DFD, (2021).
[8] J. Xu et al., ACS Appl. Mater. Interfaces 13, (2021).
[9] Y. Liu et al., npj Flex. 5, (2021).

In recent years, developing robotic systems constituted by soft materials (e.g., elastomers, hydrogels, liquid crystals, etc.) has been recognized as an opportunity for pushing the boundaries of robotic technologies and bridging the gap between humans and machines [1]. Due to the intrinsic safety in physical contact, low cost, and the capability to deal with uncertainty, soft robots exhibit greater potential in achieving unprecedented adaptation, sensitivity and agility compared to the rigid counterparts [2].
Nature often acts as a source of inspiration for robotic design, and various types of soft robots have been developed, some of which are even commercialized and benefit the biomedical areas, such as surgery and endoscopy, rehabilitation and assistance, prosthetics, and so on [3]. When soft robots are downscaled to small scales, i.e., millimeter scale and below, they are expected to enter the hard-to-access regions inside human bodies, and then perform a variety of tasks in vivo, thus revolutionizing extensive biomedical fields [4, 5]. However, the design and actuation strategies of soft robots at small scales should be carefully considered. Most strategies adopted for macroscale soft robots, such as pneumatic actuation that takes effects by inflating the elaborately designed chambers inside elastic polymers, become improper at such scale due to the challenges lying in the miniaturization of devices. Alternatively, several methodologies have been developed to actuate small-scale robots, among which, magnetic field-driven soft robots with programmable morphologies are exciting candidates [6] due to the high degree-of-freedom and simple modulation of magnetic field parameters, as well as the excellent compatibility of magnetic field with many application scenarios. However, how to guide and program the magnetization profiles following theoretically promising designs remains a challenge.
Herein, we present a 4D printing-assisted approach to design and program the magnetization profiles inside small-scale magnetic soft robots. The orientations of magnetic components (i.e., NdFeB) are encoded in three dimensions with ease, which may overcome the fundamental constraints suffered by existing methods (e.g., mold casting, chemical assembly, or the customization of complex equipment). Our approach realizes transformations between a variety of 3D structures and Gaussian surfaces under the guidance of inverse design strategies, distinguishing itself from the previous intuitive methodologies. Therefore, the present work enhances the design freedom substantially, which facilitates the constructions of miniature soft machines with demanded complex geometries and functionalities.
ACKNOWLEDGMENT
This work is financially supported by the Hong Kong Research Grants Council with project No. JLFS/E-402/18, the ITF project with Project No. MRP/036/18X funded by the HKSAR Innovation and Technology Commission (ITC) and CUHK internal grants. We thank the support from Multi-scale Medical Robotics Centre (MRC), InnoHK, at the Hong Kong Science Park; and The Chinese University of Hong Kong (CUHK) - The Shenzhen Institutes of Advanced Technology (SIAT) of the Chinese Academy of Sciences (CAS) Joint Research Laboratory on Robotics and Intelligent Systems.
REFERENCES
[1] Rus D., Tolley M. T., Nature 2015, 521, 467-475.
[2] Laschi C. et al., Sci. Robot. 2016, 1, eaah3690.
[3] Cianchetti M et al., Nat. Rev. Mater. 2018, 3, 143-153.
[4] Wang B. et al., Sci. Robot. 2021, 6, eabd2813.
[5] Jin D., Zhang L., Nat. Mach. Intell. 2020, 2, 663-664.
[6] Hu W. et al., Nature 2018, 554, 81-85.

Nature inspires diverse designs and locomotion strategies of untethered robots. Herein, we present lightweight yet strong nanocomposite robots inspired by human musculoskeletal systems for agile and magnetically organizable swimming at air-water interfaces. The nanocomposite robot consists of carbon nanotube yarn (CNTY) framework and magnetic polymer composite surrounding it, functioning as a bone and skeletal muscle. When a pulsed rotating magnetic field is induced underneath the robots, the nanocomposite robot swims with on-demand rectilinear translational motion or rotational motion according to magnetic frequency, achieving agile swimming velocity of up to 180 body lengths per second. In the manipulation of multiple robots, magnetic attraction provides soft jointed locomotion among the nanocomposite robots, enabling reconfigurable trimodal swimming: assembled rectilinear translational swimming, assembled rotational swimming, and fluctuating rotational swimming. Finally, we will discuss vorticity control of water and vortex-induced transportation of numerous cargos through the reconfigurable collective swimming of multiple nanocomposite robots.

Collective behaviors represent a wide range of phenomena that exist both in nature and artificial systems. Simple interactions among many constituting members in a system can result in complex structures and motions. Here, we report an innovative colloidal system made of dense Si nanorods that exhibit agilely reconfigurable swarming behaviors that can be readily switched by visible light in an AC rotating field. We observed two distinct self-organizing phenomena, i.e., nanowires form into interconnected 2D networks and self-assembled multi-droplets. Theoretical modeling considering light-controlled electrostatic interactions well explains the observation. This research presents an emerging swarm system that opens new opportunities for fundamental research and applications.

Recent work has demonstrated the potential of actuators consisting of bulk elastomers with phase-changing inclusions for generating high forces and large volumetric expansions. Simultaneously, granular assemblies have been shown to enable tunable properties via different packings, dynamic moduli via jamming, and compatibility with various printing methods via suspension in carrier fluids. In this talk, we will introduce granular actuators, a class of soft actuators made of discrete grains. The soft grains consist of a hyperelastic shell and multiple solvent cores. Upon heating, the encapsulated solvent cores undergo a liquid-to-gas phase change, inducing rapid and strong volumetric expansion of the hyperelastic shell up to 700%. The grains can be used independently for micro-actuation or in granular agglomerates for macro-scale actuation, demonstrating the scalability of the granular actuators. Furthermore, the grains can be suspended in a carrier resin or solvent to enable printable soft actuators via established granular material processing techniques. By combining the advantages of phase-change soft actuation and granularity, we present the opportunity to realize soft actuators with tunable bulk properties, compatibility with self-assembly techniques, and emergent intelligence.

The majority of soft robotic approaches utilize just a handful of material chemistries, such as silicone rubbers or hydrogels, and often use non-scalable materials processing techniques such as molding. Very few materials chemistries are designed with both a processing method (such as additive manufacturing) and particular soft robotic application in mind. To address these challenges, we looked to develop better materials and manufacturing methods for soft actuators, with a particular interest towards soft, yet tough, robotic devices for burrowing applications. Photopolymerization can be utilized to enable new materials and fabrication methods towards these ends. In one such example, we investigated the 3D printing of an underutilized material in soft robotics, polyurethane elastomers, by employing an ultraviolent light-assisted direct ink writing (UV-DIW) approach. Printing is enabled using a dual curable ink composed of two orthogonal chemistries: a photocurable acrylate and a thermally curable step growth polyurethane. By photopolymerizing the ink during the printing process, the light-induced crosslinking locks in the initial structure of the printed part, allowing high aspect ratio prints by preventing sagging during printing and thermal post-curing. This technique affords high spatial resolution (down to 250 microns) and can produce highly extensible elastomers (ε >100%) with a range of mechanical properties (1<E<20MPa). By tuning this dual cure chemistry, multi-material parts can be fabricated with gradient properties through the addition of a second ink reservoir. Such a materials chemistry offers potential for manufacture of entire soft devices in one single step.

Animals are semi-discretized. Systems of organs that perform multiple functions and are spatially discrete from each other, yet interconnected chemically and electrically. The complexity of animals such as vertebrates allow for adaptation that permits survival within single generations upon significant environmental change. In the search for generally adaptive robots, we believe that the complex, multifunctional, and interconnected organ systems of animals is a model that should be embraced, rather than avoided. Fortuitously, additive manufacturing enables “complexity for free.” Of course, it is not yet that simple to be complex, but we will present additive manufacturing approaches we have used to distribute sensing, actuation, energy, and computation in soft robots. The framework we use for guiding our design evolution is Autonomous Materials, where we push the manufacturing of robots towards forming processes, rather than assembly.

Origami patterns have gained interest within the soft robotic and active materials communities for their programmable and tunable mechanical properties. This presentation will discuss two recent innovations in origami-patterned soft materials, in fabrication processes and in origami-patterned sensors. The applications of origami to mechanical sensors (such as in soft robotics) will be presented first. The sensors are cast in the desired pattern and foil electrodes are applied to the pattern faces to create a capacitive strain sensor. Definition of the as-fabricated pattern angles modifies the sensor’s compliance and strain sensitivity.

Fabrication innovations will also be discussed. Precisely defining pattern angles in compliant or stretchable materials is a challenge. Approaches to origami soft-material patterning using both elastomer casting and direct additive manufacturing processes will be presented. In the direct additive manufactured process, PLA filament is underextruded as the structure is fabricated [1], which creates a structure that is porous, strong, mechanically robust— and flexible. The performance and tunability of origami-patterned structures fabricated in this process will be presented.

[1] J. Forman, M.D. Dogan, H. Forsythe, and H. Ishii, "DefeXtiles: 3D Printing Quasi-Woven Fabric via Under-Extrusion," in Proceedings of the 33rd Annual ACM Symposium on User Interface Software and Technology (UIST '20), 2020.

Hydrogels are well-established as actuators, as chemoresponsive materials, and as drug delivery devices. Insulin therapy has proven its effectiveness in the treatment of diabetes, but there are still several limitations, including insulin pump failure, infusion set blockage, infusion site problems, insulin stability issues, and user error, among others. Continuous glucose monitoring also poses some problems, including the need for constant calibration, the adhesion of proteins to sensor surfaces, and a combination of other factors. In this study, we examine the synthesis and characterization of a layered hydrogel system that combines a microfluidic device with both sensing and drug delivery mechanisms that allows for the continuous monitoring of the glucose concentration in a solution and delivers a concentration of drug that is proportional to the concentration of analyte detected in solution. While still in the early stages, this hydrogel system could prove beneficial to the monitoring and treatment of diabetes, while removing several of the limitations of current technologies, by addressing the need for real-time blood glucose monitoring, providing for the accurate dosing of insulin through a solution that integrates the sensor with a drug delivery system, and limits the human error that may exist. We report some early-stage data on hydrogel-based microfluidics and the resulting drug release kinetics of the hydrogel system, including quartz crystal microbalance data.

Work presents an electromagnetically-controlled shape morphing composite (e-morph) as a new material in soft robotics area. The structure is composed of silicon with liquid metal channels and granular core. Composite can change its shape in a controlled way by electromagnetic force, while the jamming mechanism allows the granules used in the core to maintain the desired shape. This paper provides a detailed description of material properties. Additionally the modeling procedure and potential application is presented.

Aerial robotics has become a key enabler of industrial automation and environmental interaction tasks with applications ranging from environmental sensing in forests, over water sampling in lakes to industrial manufacturing and repair. Most aerial robots however are rigid in structure and rely on accurate state estimation and controlled sensor-actuator loops to perform the tasks and navigate through the environment. Natural flyers on the other hand rely not just on flight control but also on smart and adaptive materials to expand their operational flight envelope and to increase the functionality of their wings allowing them to display an impressive range of skills. Recently, the robotics community has started to employ adaptive morphologies, smart materials and soft robotics methods in aerial robotics with the promise of achieving higher levels of system resilience and adaptation to complex or changing environments. In this talk I will present how animal-inspired soft robot design methods can greatly enhance aerial robotics capabilities and how adaptive morphologies and functional materials can be co-evolved in tandem with aerial robotic embodiments. The talk will also include application examples of such soft aerial robots placing sensors in forests, collecting water samples autonomously or landing on wind turbines to perform construction and repair tasks.

In robotics, autonomy is achieved by feedback mechanisms that include the sensing of the environment and the response to the environment with proper feedback mechanisms. In conventional robots, this is usually accomplished with electronics and programming. However, in soft robots, common electronics infrastructures cannot be used since they do not compile with the robot's low weight and high agility. Especially for robots made up of only soft materials like polymers and hydrogels, the incorporation of such electronics is cumbersome. In living organisms, on the other hand, such autonomy is provided by internal and external sensing and self-regulation. In animals and plants, numerous biochemical pathways operate to track the environment actively and respond to it continuously. Although it is impossible to design a synthetic system that copies these complicated pathways, it is possible to use the concept of self-regulation to achieve autonomous movement in soft robots.
Here we will show examples of the self-regulation concept applied to soft plant robots. In the first system, the robot is operated by the material feedback supplied by the proper geometrical arrangement of shape memory alloys, which respond to sunlight by reversible thermal shrinking. In the second one, the feedback is achieved by the reversible hydration/dehydration of water from hydrogels. In both systems, the action is autonomous; there is no requirement of external energy input or control other than the sun and its elevation in the sky. The efficiency for tracking is measured by stretchable solar panels attached to the 'leaves' and compared to the cases where the robots cannot track the sun because of the absence of the feedback elements. The robot movement has a complete angle-tracking nature rather than the previous solar trackers' on/off response. The simple systems, especially the latter one, have tailorable geometries and chemistries. We envisage that such interplay of material science, bioinspiration, and chemistry can help integrate autonomous mechanisms in soft robots.

Understanding natural ecosystems is fundamental to preserve the health of our planet, since natural resources are being consumed at an unsustainable rate now more than ever. Plants have an irreplaceable role in planet’s health; they account for 80 percent of the total biomass, represent the base of every food chain and produce energy from sun. In this scenario, there is an increasing need for new sustainable and smart tools capable of precisely operating over plant organs for in situ plant applications, including sensing and delivery of molecules or micro/nano particles. However, standard technologies in agriculture and plant bioengineering suffer from low spatial and temporal resolution or require extensive spraying of agrochemicals to plants (thus, increasing pollution).
Bioinspired soft- and miniature machines can offer innovative solutions for operating in dynamic real-world environments by adapting their morphology to plant organs like leaves in forests or agricultural fields^1,2. In this perspective, we mimicked the leaf-leaf anchoring mechanism of G. aparine to propose miniature adhesive machines that enable in situ anchoring to leaf tissues via multifunctional microhooks^3,4. The hook climber Galium aparine has exploited a fascinating parasitic ratchet-like climbing mechanism to strongly anchor over host plant leaves surfaces, using its leaves covered with microscopic hooks for mechanical interlocking⁵. Starting from this observation, we built microhook-based biodegradable and compliant miniaturized systems by using high-resolution 3D micro-manufacturing techniques in combination with molding and casting of recyclable materials. Then, we demonstrated and characterized the strong, yet reversible, anchoring forces of the microhooks to leaves (up to F_shear = 12N/cm²), according to leaf microstructures and mechanics. Finally, we embedded our microhooks with electronics and actuators to create and demonstrate different microhooks-based machines for in situ leaf applications, including a flexible wireless multi sensors device to monitor the leaf proximal parameters; degradable hooks for in-plant drug delivery; and an on-leaf-locomoting light-driven robot⁴.
Our research highlights the great potential to use plant-like anchoring strategies in miniature machines for real-world scenarios, such as on natural leaf surfaces. We aim to further develop an integrated microhook-based machine capable of autonomously and simultaneously performing sensing and delivery tasks on plant leaves, increasing the multifunctionality and complexity of the system. These integrated systems can be spread everywhere in a real field “case study”, like a vineyard or a forest, providing capillary information on plant needs, and supplying them with drugs, nutrients or notifying the irradiation system to give water when needed, thus preserving resources while reducing costs and pollution.

1 Mazzolai, B. & Laschi, C. A vision for future bioinspired and biohybrid robots. Science robotics 5, eaba6893 (2020).
2 Fiorello, I., Del Dottore, E., Tramacere, F. & Mazzolai, B. Taking inspiration from climbing plants: Methodologies and benchmarks–A review. Bioinspiration & Biomimetics (2020).
3 Fiorello, I. et al. Climbing Plant Inspired Micropatterned Devices for Reversible Attachment. Advanced Functional Materials 30, 2003380 (2020).
4 Fiorello, I. et al. Plant-like hooked miniature machines for on-leaf sensing and delivery. Communications Materials 2, 1-11 (2021).
5 Bauer, G. et al. Always on the bright side: the climbing mechanism of Galium aparine. Proc Biol Sci 278, 2233-2239, doi:10.1098/rspb.2010.2038 (2011).

For a long time, movements of plants were a phenomenon that received little attention in the field of (soft) robotics, and only in recent years plants have been recognized as interesting model systems. This is partly because plant movements are either too slow (e.g. growth processes) or too fast (e.g. motile traps of carnivorous plants) to be perceived as movements by the human eye. Plant movements are typically effected by hydraulic mechanisms (e.g. reversible shrinking and swelling of cells and irreversible growth) and can be sped up by the release of stored energy, initially caused by continuously built up mechanical pre-stresses. Of special importance for the transfer to plant-inspired technical motile systems is the finding that plants typically perform movements by elastic deformation of larger regions of their tissues or organs, i.e. without localized hinges. This set-up avoids stress concentrations, which are typical for pivot points in conventional technical joints and reduces failure caused by overload and/or pronounced wear. Localized hinges are not only found in many technical joints but also are characteristic for many mobile animals such as arthropods and vertebrates, in which movement is based on a complex interplay of muscles (actuators), stiff members (bones or cutinized leg elements), and localized joints with gliding parts.

With the development of soft robots and soft machines motion based on elastic deformation became relevant. Examples include soft robots inspired by octopus arms and plants and other motile systems, in which the motion pattern, the type of actuation, and the general appearance became increasingly similar to their biological models. In these structures not only the functionality of the biological models was successfully transferred, but increasingly also the underlying hierarchical structuring, the actuation principles and also the esthetics of motion patterns of the biological concept generators. Recent attempts aim at transferring additional technically interesting functions of the biological models as for example improved sensing-(re)action-coupling, (self-)adaptability, self-repair, and energy-autonomy in the next generations of bio-inspired soft robots and motile systems in general. Plants have proven to be very suitable concept generators for this task as they lack a central control unit (i.e. a brain) and most sensing and (re)acting take place decentralized in an integrated manner in the various organs or tissues.

Plant-inspired developments in the fields of soft machines and soft robotics are demonstrated by research projects currently carried out in the Plant Biomechanics Group Freiburg and the Cluster of Excellence livMatS. Examples include liana-inspired soft robots, leaf- and flower-inspired façade shading systems and artificial Venus flytraps. In collaboration with the ICD and the IntCDC Cluster of Excellence at the University of Stuttgart, demonstrators were developed for self-adaptive building envelopes inspired by pinecones, soft machines with sequential motion steps based on the movement of silver thistle bracts, and - as an example of a medical application - an adaptive wrist forearm splint. A particular focus of current research is on embodied energy and intelligence found in moving plant organs, which offer a huge potential for a new generation of materials systems for soft robots, bioinspired architecture and technical applications in general.

Ionoelastomers, consisting of low glass-transition polyelectrolyte networks, represent soft and stretchable ionic analogs of electronic semiconductors. We have recently studied two different types of electro-mechanical responses of heterojunctions formed by contacting ionoelastomer layers of opposite polarity, both with potential applicability in soft robotic contexts. In the first case, we demonstrated that such junctions provide reversibly switchable electro-adhesion. Thanks to the nanometer-scale thickness of the ionic double layer, large Maxwell stresses and corresponding enhancements in adhesion can be generated using applied potentials of only 1 – 3 V, in contrast to the kV levels generally required for established dielectric electro-adhesives. This leads to superior performance in terms of adhesive force per electrostatic energy, as well as damage resistance. In the second case, we found that mechanical deformation of a bilayer sample provides transient electrical signals associated with rearrangement of polymer strands and counterions within the ionic double layer. This effect provides a new mechanism for electrical strain sensing and energy harvesting.

Stimulus-response system and conscious response enable humans to respond effectively to environmental changes and external stimuli. Inspired by human conscious response, this paper presents an artificial stimulus-response system capable of conscious response. The system is composed of an artificial visual receptor, artificial synapse, artificial neuron circuit, and an actuator. Vertically stacked graphene/semiconductor heterostructure is employed as a base layer for the artificial visual receptor and artificial synapse to allow high integration density. The artificial visual receptor with InP quantum dot layer generates the electrical pulse under light illumination. The artificial synapse with retentive electric double layer processes the electrical pulse signals to strengthen synaptic connections. The artificial neuron circuit is designed with CMOS circuit which integrates synaptic output signal to activate the actuator. By incorporating these artificial nervous components, a series of conscious response process is demonstrated which drastically reduces response time as a result of learning from repeated stimuli. The demonstrated artificial stimulus-response system offers a promising research field providing an alternative to patients with neurological disorders.

The electronic skin (e-Skin) in robotics is expected to provide the sensation and perception functionalities just like our own skin. To this end, the e-Skin should have distributed touch sensors and electronics to attain local computation. To this end, we have developed sensors, synaptic transistors and neural-like multi-gate transistors that can mimic the functionality of biological mechanoreceptors, synapse, and neurons, respectively. The devices are realised with low-dimensional materials (i.e., graphene, Si nanowire, ZnO nanowire) using the printing method onto soft substrates. These devices show excellent mechanical flexibility and can be wrapped onto various curvature surfaces. The developed distributed devices emulate several principles discovered in biological neural system, such as long-term potentiation (LTP), long-term depression (LTD), spiking-rate dependent plasticity (SRDP), spatial-temporal integration, all-or-none rule, etc. Finally, based on the demonstrated emerging devices, a computational e-Skin prototype has been demonstrated.

Various insects can entrap and stabilize air plastrons and bubbles underwater. When these bubbles interact with surfaces underwater they create air capillary bridges that de-wet surfaces and even allow underwater reversible adhesion. In this study, a robotic arm with an interchangeable 3D- printed bubble-stabilizing unit is used to create air capillary bridges underwater for manipulation of small objects. Particles of various sizes and shapes, thin sheets and substrates of diverse surface tensions, from hydrophilic to superhydrophobic, can be lifted, transported, placed and oriented using one or two-dimensional arrays of bubbles. Underwater adhesion, derived from the air capillary bridges, is quantified depending on the number, arrangement and size of bubbles, and the contact angle of the surface. This includes a variety of commercially available materials and chemically modified surfaces. Overall, it is possible to manipulate millimeter to sub-millimeter scale objects underwater. This includes cleaning submerged surfaces from colloids and arbitrary contaminations, folding thin sheets to create three-dimensional structures and precise placement and alignment of objects of various geometries. The robotic underwater manipulator can be used for automation and control in cell culture experiments, lab-on-chip devices and manipulation of objects underwater. It offers the ability to control transport and release of small objects without the need for chemical adhesives, suction based adhesion, anchoring devices or grabbers.

The octopus's arm has a constant volume muscular hydrostat system composed of longitudinal musculature, transverse musculature, oblique musculature, and a thin layer of connective tissues. These morphological features enable the octopus arm to achieve multimodal locomotion modes: elongation, shortening, bending, and stiffening. The existing octopus-like robot arms do not fully reproduce the muscular hydrostat system, especially, how the constant volume feature of the muscular hydrostats affects the fast deformation and stiffening is less understood. We designed a biomimetic prototype that aims at reproducing the muscular hydrostat of the octopus arm. The transverse musculature is actuated by a piled three-channel hydraulically amplified electrostatic soft actuator with constant volume. We used cables to drive the longitudinal musculature. The transverse musculature and the longitudinal musculature are arranged based on the biological structure of the octopus arm system, which can achieve multimodal motion via a coupled actuation of electrostatic effect and cables. The proposed prototype may help to understand the effect of the octopus muscular hydrostats system's constant volume feature on fast deformation and stiffness tuning. This work may also extend the potential applications of the hydraulically amplified electrostatic soft actuator in the biomimetic soft robots.

Reversible programming of 3D soft mesostructures is desired for many applications including soft robotics. The large, reversible shape changes of liquid crystal elastomers (LCEs), which result from the coupling between the alignment of liquid crystal (LC) molecules and the macroscopic deformation of polymer networks, have attracted much attention for such applications. In this talk, I will discuss our recent effort in developing a facile and versatile strategy to create reconfigurable, freestanding 3D mesostructures of LCEs and magnetic LCE composites for soft robotics via spatially programming LC molecules. Experimental and theoretical results of more than 20 reconfigurable 3D LCE mesostructures of diverse configurations and their large, reversible shape-switching behaviors over multiple cycles will be demonstrated. In addition, an LCE gripper and a robot of ferromagnetic LCE composites that simultaneously responds to magnetic and thermal stimuli for diverse biomimetic behaviors, especially crawling underneath a narrow crack, will be used to illustrate the integration of other functional materials to LCEs for multifunctional systems.

Introduction: The gastrointestinal (GI) tract provides a convenient, non-invasive access point for medical therapeutics and diagnostics. Ingestible devices have traditionally been limited to transient operation and rigid, inorganic structures, largely due to a lack of robust, biodegradable materials. As shown in this work, genipin-crosslinked gelatin (Gelapin) biopolymers offer a high degree of mechanical tunability and can be fabricated into a diverse range of functional geometries. Leveraging auxetic origami design and laser processing techniques, we demonstrate the potential for edible, self-deployable Gelapin actuators. Such structures can be fabricated into gastrorentitive interfaces capable of packing up to ~10cm² of surface area into a standard “000” pill for future bioelectronic and drug-delivery platforms.
Methods: All raw materials were acquired from Millipore-Sigma (Milwaukee, WI, USA) and used as-received unless otherwise noted. For this study, genipin dissolved in ethanol and water was used to crosslink 300-Bloom, Type A porcine. Approximate crosslinking percentages were calculated based on the occupation of free amine sites using a ninhydrin assay. Varying amounts of glycerol plasticizer were incorporated into solution and the resulting biogels were dehydrated for 24 hours at 70°C. Auxetic origami designs were etched using a Rabbit CO₂ laser (Middletown, Ohio) to create hill and valley crease lines. For initiating the origami folds, a 3D-printed polymer mold was used to compress the etched substrates by 10% from their initial height. Mechanical and viscoelastic properties of biogel and bioplastic samples were evaluated using a tabletop RSA-G2 Solids Analyzer and Discovery HR 20 Rheometer by TA Instruments (New Castle, DE) along with a Instron Universal Testing System (Norwood, MA).
Results: In order to survey the mechanical tunability of gelapin as a function of crosslinking density, a variety of mechanical tests were done both in the gel and plastic state. Storage modulus of the hydrated biogel was measured in the linear viscoelastic region at a frequency of 1Hz as the sample crosslinked in-situ between parallel plates. After 36 hours, the un-crosslinked sample exhibited a storage modulus of 1.95kPa, while the fully crosslinked sample measured at 11.09kPa, an increase of 469%. A similar degree of mechanical tunability was demonstrated via uniaxial tensile tests of Gelapin bioplastic dog bones. These samples were prepared using glycerol as a plasticizer at constant loading (35% mass relative to gelatin) and laser-cut for testing. The resulting stress-strain plots show Young’s modulus increasing from 32MPa to 417MPa and strain at failure decreasing from 963% to 108% between 0% and 100% crosslinking, respectively. This baseline mechanical data for Gelapin can be used to describe deployment due to elastic strain energy from a confined pill volume. Similarly, swelling kinetics can be used to identify deployment forces resulting from osmotic actuation in GI fluid. Gelapin biogels showed a nearly exponential decay in swelling capacity a function of crosslinking with equilibrium mass changes shifting from +55% all the way to -4% (syneresis) over an extended 60hr period.
Conclusion: Here we have described thorough mapping of the Gelapin system’s mechanical, rheological and swelling properties. Additionally, Gelapin bioplastics are demonstrated to be compatible with a highly scalable CO₂ etching process for the production of detailed, expandable origami structures. In future research, we aim to leverage this material knowledge in order to construct and evaluate a semi-empirical model for the bimodal deployment of gelapin origami actuators. This model will be applied to optimize geometric and material design parameters of an auxetic origami tube capable of dilating to a diameter of 25mm, while exerting sufficient radial pressure to be retained in the small intestine for prolonged periods of time (>8hrs).

Biological nervous systems are spotlighted recently to be mimicked as a neuromorphic system for the detection of environmental stimuli such as touch, force, temperature, light, sound, taste, or smell, etc. To fully mimic biological performance, bio-inspired neuromorphic systems have beenreported. However, few neuromorphic systems have been demonstrated with functions close to the actual biological nervous system.
We present here a neuromorphic system integrated with a 3D printed humanoid hand for the demonstration of artificial unconscious stimulus response. For the realization of 3D conductive printing on the curved humanoid hands, multi-axis robot 3D printing technology was developed for the fabrication. Then, the fabricated portable neuromorphic device and 3D printed tactile sensor were integrated with the 3D structural humanoid hand as a neuromorphic humanoid hand. Three combined performance as bio-inspired detection of evironmental stimuli, signal conversion/ transmission, and signal procession were demonstrated through such a neuromorphic humanoid system. Bio-inspired signal perception and biomimetic reflex arc function for a quick and unconscious response were also demonstrated to hold an unknown object with an automatically increased gripping force. The proposed neuromorphic system explores a new method to realize unconscious response-ability for the next generation of artificially intelligent robot hands.

Endovascular surgery has become a popular minimally invasive approach for vascular diseases diagnosis and therapy. However, the conventional procedure still has limitations. For one thing, most commercial guidewires/catheters are operated at the remote end, in which case the successful operation requires the surgeon’s high expertise and prudent operation. For another, most existing continuum devices have fixed properties, which cannot simultaneously satisfy the demand for navigating through a tortuous path and supporting the subsequent force application.
This study proposes an adaptive catheter that can be magnetically steered and electrically softened, featured by improved manipulability and easy fabrication. An electrically conductive polymer (ECP) is first prepared by dispersing the carbon black (CB) filler into the polycaprolactone (PCL) matrix, which is further processed into small-scale U-shape electrode cores through 3D printing. Then, a magnetic catheter is fabricated by the molding process. The main components include the electrode core, the polydimethylsiloxane (PDMS) channel, the silicone jacket, and the tip magnet. The deformation can be controlled by external magnetic fields through applied torque, and the stiffness can be changed by powering electrical currents through joule heat-induced phase transition. Furthermore, based on the positive temperature coefficient effect of the ECP, the temperature-related stiffness can be estimated and maintained by monitoring and regulating the real-time current and voltage. This is helpful in dynamic environments, considering the temperature of the electrode core depends on both heat generation and dissipation.
Various tests are conducted to calibrate the characteristics of the proposed ECP. Simulations are conducted to analyze temperature distribution to ensure body safety during heating. Magnetic catheter prototypes with single and multiple segments are fabricated for demonstration. The proposed catheter has the potential to improve operational safety and clinical outcomes of the interventional procedure.
Acknowledgments
The authors would like to thank the support from the Croucher Foundation Grant with Ref. No. CAS20403, and Multi-scale Medical Robotics Center (MRC), InnoHK at the Hong Kong Science Park.

A textile actuator has high adaptivity and flexible structural characteristics, suggesting applicability to various fields requiring highly adaptable properties and safety. Many textile-based sensors are also being developed due to their high compatibility with soft robots. Meanwhile, existing textile sensors are available in single filament or layered composites or in connected form. We propose a novel type of textile end-effector, a fully integrated textile actuator and sensor system.

A fully textile end-effector consists of a monofilament where a shape memory alloy (SMA) wire is wrapped in non-stretchable fibers, and conductive yarn coated by silver. An SMA wire was used to fabricate a smart textile actuator composed of the looping architecture and then manufactured into a cylindrical shape to induce volumetric deformation. It was designed to allow volume expansion and contraction of the cylindrical-shaped textiles by applying a current to the fabricated textile actuators. In order to monitor the actuator’s grip conditions, the conductive yarn was knitted together. The physical mechanism of the textile sensor is variation of resistance, which changes in response to inverse proportion to the curvature of the actuator. In this research, the sensor is used as a strain sensor to monitor the end-effector's gripping force and state. It allows feedback control of the entire soft actuating system.

Most of the textile end-effectors keep their gripping state by continuously flowing current. However, using the textile sensor, we can determine whether the end-effector has gripped an object stably. This enables energy-efficient control with minimal power consumption through feedback control. This integrated end effector is commercially available as it can be made simply by knitting a conductive yarn with an SMA filament to enable machine automation. They also can be used for human-friendly systems, nature mimicking robots, and wearable devices.

This work was supported by the National Research Foundation of Korea(NRF) grant funded by the Korea government(MSIT) (No.2018R1A5A7023490),(NRF-2021R1F1A1063485) by the Korea Institute of Machinery and Materials (KIMM) funding (NK230C), and by the Ministry of Trade, Industry, and Energy (MOTIE) and the Korea Institute for Advancement of Technology (KIAT) through the International Cooperative R&D program (Project No. P0016173).

Soft actuators that mimic creature's superior properties can easily be adapted to its environment by mimicking it with soft materials. Since they do not require advanced control techniques and are lightweight, these actuators are handy. Especially, these soft actuators can be expected to explore and operate in barren environments such as deep seas where various obstacles exist. Therefore, research on bioinspired soft actuators that can make safe interactions in various environments has recently been accelerated.

In this study, we present a soft actuator that simulates elephant’s trunk which are different from conventional joint actuators. Existing rigid actuators have to increase the number of joints and links for high degree of freedom, which makes it more difficult to operate. However, this actuator consists of soft silicon and one continuous joint, so the movement shows a soft and smooth curve. Therefore, an actuator that simulates an elephant nose has the advantage of helping workers to implement multi-freedom movements even within a limited space without damaging the surrounding environment.

To mimic the free and smooth movement of the elephant's nose, which consists solely of muscles, the soft actuator is manufactured by four shape memory alloy wire springs include cylindrical silicon. These wires have separate power sources, so when applied, the lower end of the column moves forward, backward, right, and left. It also shows more diverse movements by drawing straight lines and curves within three dimensions depending on the application of power, such as applying power to No. 1 and No. 2 SMA spring. We analyzed how much movement varies depending on which shape memory alloy wires are powered and the strength of the power. We also checked results of how far the end of the actuator is floated from the floor and which direction it went from the center. Furthermore, this actuator could be combined with different type of end effector and developed to conduct a finer control.

This work was supported by the National Research Foundation of Korea(NRF) grant funded by the Korea government(MSIT) (No.2018R1A5A7023490),(NRF-2021R1F1A1063485) by the Korea Institute of Machinery and Materials (KIMM) funding (NK230C), and by the Ministry of Trade, Industry, and Energy (MOTIE) and the Korea Institute for Advancement of Technology (KIAT) through the International Cooperative R&D program (Project No. P0016173).

The fabrication of soft robotic systems is at the forefront of creating devices which are lightweight, flexible, and feature increasing functionality making it more suitable for many applications such as artificial muscles and morphing surfaces. Specifically, thermally activated smart materials, which can be defined as materials that repeatably deform in a controlled manner by an external thermal stimulus, are particularly promising in expanding this design space. The integration of this type of smart materials into modern designs tackles demands such as increased multi-functionality and response speed. This research theme focuses on the development of thermally activated materials and the implementation of these active materials as Electrothermal Actuators (ETA) and Shape Memory Polymer (SMP) actuators. ETAs represent a new generation of materials that can produce different motions due to thermal expansion caused by Joule heating, while SMPs have the ability to be ‘programmed’ to a temporary position and actuate to its original position when a specific stimulus is applied. Furthermore, these topics can be applied to advanced Additive Manufacturing (AM) techniques to construct 4D printed structures allowing greater design and fabrication freedom. In this work, the fabrication of a soft ETA for a shape-morphing surface application is demonstrated. Variations can by made in the carbon nanotube-elastomer composite to tailor the shape-response and deformability. The various factors that influence the actuation performance such as geometrical parameters, electrical conductivity, and actuator fabrication methods were explored. In addition, a new SMP composite with 2-dimensional MXene (Ti₃C₂) is fabricated and demonstrates a significant improvement in the mechanical and shape memory performance. As the actuation performance of these thermally sensitive materials depends highly on their microstructure it is shown that via tailoring the amorphous and crystalline phases, it is possible to adjust the response behaviours. The overall performance of these developed materials shows great promise for improving soft robotics systems.

By laser scribing with a CO₂ laser onto a polymer precursor like polyimide, it is possible to create conductive tracks by laser-induced pyrolysis. So-called laser-induced graphene (LIG) is obtained in this process, which consists of a porous 3D structure with a high surface area and good conductivity. By embedding the LIG into a stretchable and soft polydimethylsiloxane matrix, a conductive and stretchable composite was created. The electro-mechanical properties of this LIG composite were investigated. We realized various kinds of soft actuators based on LIG composites. The embedded LIG is used as a joule heating element to trigger a thermoresponsive actuation in the composite. By coupling the LIG/PDMS composite with a thin film of a smart humidity-responsive hydrogel (poly-(N-vinycaprolactam), pNVCL) by means of initiated chemical vapor deposition (iCVD), we realized a multi-responsive soft actuator. The actuator performance was characterized and a demonstrator was fabricated. Thanks to the piezoresistive behaviour of the embedded LIG the bending angle of the soft actuating structures could be measured.