Smart textiles using fluid-driven artificial muscle fibers

Thank you for visiting Nature.com. You are using a browser version with limited CSS support. For the best experience, we recommend that you use an updated browser (or disable Compatibility Mode in Internet Explorer). In addition, to ensure ongoing support, we show the site without styles and JavaScript.
Displays a carousel of three slides at once. Use the Previous and Next buttons to move through three slides at a time, or use the slider buttons at the end to move through three slides at a time.
Combining textiles and artificial muscles to create smart textiles is attracting a lot of attention from both the scientific and industrial communities. Smart textiles offer many benefits, including adaptive comfort and a high degree of conformance to objects while providing active actuation for desired movement and strength. This article presents a new class of programmable smart fabrics made using various methods of weaving, weaving and gluing fluid-driven artificial muscle fibers. A mathematical model was developed to describe the ratio of the elongation force of knitted and woven textile sheets, and then its validity was tested experimentally. The new “smart” textile features high flexibility, conformality, and mechanical programming, enabling multi-modal movement and deformation capabilities for a wider range of applications. Various smart textile prototypes have been created through experimental verification, including various shape change cases such as elongation (up to 65%), area expansion (108%), radial expansion (25%), and bending motion. The concept of reconfiguration of passive traditional tissues into active structures for biomimetic shaping structures is also being explored. The proposed smart textiles are expected to facilitate the development of smart wearables, haptic systems, biomimetic soft robots, and wearable electronics.
Rigid robots are effective when working in structured environments, but have problems with the unknown context of changing environments, which limits their use in search or exploration. Nature continues to surprise us with many inventive strategies to deal with external factors and diversity. For example, the tendrils of climbing plants perform multimodal movements, such as bending and spiraling, to explore an unknown environment in search of a suitable support1. The Venus flytrap (Dionaea muscipula) has sensitive hairs on its leaves that, when triggered, snap into place to catch prey2. In recent years, the deformation or deformation of bodies from two-dimensional (2D) surfaces to three-dimensional (3D) shapes that mimic biological structures has become an interesting research topic3,4. These soft robotic configurations change shape to adapt to changing environments, enable multimodal locomotion, and apply forces to perform mechanical work. Their reach has extended to a wide range of robotics applications, including deployables5, reconfigurable and self-folding robots6,7, biomedical devices8, vehicles9,10 and expandable electronics11.
Much research has been done to develop programmable flat plates that, when activated, transform into complex three-dimensional structures3. A simple idea for creating deformable structures is to combine layers of different materials that flex and wrinkle when exposed to stimuli12,13. Janbaz et al. 14 and Li et al. 15 have implemented this concept to create heat-sensitive multimodal deformable robots. Origami-based structures incorporating stimulus-responsive elements have been used to create complex three-dimensional structures16,17,18. Inspired by the morphogenesis of biological structures, Emmanuel et al. Shape-deformable elastomers are created by organizing air channels within a rubber surface that, under pressure, transform into complex, arbitrary three-dimensional shapes.
The integration of textiles or fabrics into deformable soft robots is another new concept project that has generated widespread interest. Textiles are soft and elastic materials made from yarn by weaving techniques such as knitting, weaving, braiding, or knot weaving. The amazing properties of fabrics, including flexibility, fit, elasticity and breathability, make them very popular in everything from clothing to medical applications20. There are three broad approaches to incorporating textiles into robotics21. The first approach is to use the textile as a passive backing or base for other components. In this case, passive textiles provide a comfortable fit for the user when carrying rigid components (motors, sensors, power supply). Most soft wearable robots or soft exoskeletons fall under this approach. For example, soft wearable exoskeletons for walking aids 22 and elbow aids 23, 24, 25, soft wearable gloves 26 for hand and finger aids, and bionic soft robots 27.
The second approach is to use textiles as passive and limited components of soft robotic devices. Textile based actuators fall into this category, where the fabric is usually constructed as an outer container to contain the inner hose or chamber, forming a soft fiber reinforced actuator. When subjected to an external pneumatic or hydraulic source, these soft actuators undergo changes in shape, including elongation, bending or twisting, depending on their original composition and configuration. For example, Talman et al. Orthopedic ankle clothing, consisting of a series of fabric pockets, has been introduced to facilitate plantar flexion to restore gait28. Textile layers with different extensibility can be combined to create anisotropic movement 29 . OmniSkins – soft robotic skins made from a variety of soft actuators and substrate materials can transform passive objects into multifunctional active robots that can perform multi-modal movements and deformations for various applications. Zhu et al. have developed a liquid tissue muscle sheet31 that can generate elongation, bending, and various deformation motions. Buckner et al. Integrate functional fibers into conventional tissues to create robotic tissues with multiple functions such as actuation, sensing, and variable stiffness32. Other methods in this category can be found in these papers 21, 33, 34, 35.
A recent approach to harnessing the superior properties of textiles in the field of soft robotics is to use reactive or stimulus-responsive filaments to create smart textiles using traditional textile manufacturing methods such as weaving, knitting and weaving methods21,36,37. Depending on the composition of the material, reactive yarn causes a change in shape when subjected to electrical, thermal or pressure action, which leads to deformation of the fabric. In this approach, where traditional textiles are integrated into a soft robotic system, the reshaping of the textile occurs on the inner layer (yarn) rather than the outer layer. As such, smart textiles offer excellent handling in terms of multimodal movement, programmable deformation, stretchability, and the ability to adjust stiffness. For example, shape memory alloys (SMAs) and shape memory polymers (SMPs) can be incorporated into fabrics to actively control their shape through thermal stimulation, such as hemming38, wrinkle removal36,39, tactile and tactile feedback40,41, as well as adaptive wearable clothing. devices 42 . However, the use of thermal energy for heating and cooling results in slow response and difficult cooling and control. More recently, Hiramitsu et al. McKibben’s fine muscles43,44, pneumatic artificial muscles, are used as warp yarns to create various forms of active textiles by changing the weave structure45. Although this approach provides high forces, due to the nature of the McKibben muscle, its rate of expansion is limited (< 50%) and small size cannot be achieved (diameter < 0.9 mm). In addition, it has been difficult to form smart textile patterns from weaving methods that require sharp corners. To form a wider range of smart textiles, Maziz et al. Electroactive wearable textiles have been developed by knitting and weaving electrosensitive polymer threads46.
In recent years, a new type of thermosensitive artificial muscle has emerged, constructed from highly twisted, inexpensive polymer fibers47,48. These fibers are commercially available and are easily incorporated into weaving or weaving to produce affordable smart clothes. Despite the advances, these new heat-sensitive textiles have limited response times due to the need for heating and cooling (e.g. temperature-controlled textiles) or the difficulty of making complex knitted and woven patterns that can be programmed to generate the desired deformations and movements. Examples include radial expansion, 2D to 3D shape transformation, or bi-directional expansion, which we offer here.
To overcome these aforementioned problems, this article presents a new fluid-driven smart textile made from our recently introduced soft artificial muscle fibers (AMF)49,50,51. AMFs are highly flexible, scalable and can be reduced to a diameter of 0.8 mm and large lengths (at least 5000 mm), offering a high aspect ratio (length to diameter) as well as high elongation (at least 245%), high energy efficiency, less than 20Hz fast response). To create smart textiles, we use AMF as an active yarn to form 2D active muscle layers through knitting and weaving techniques. We have quantitatively studied the expansion rate and contraction force of these “smart” tissues in terms of fluid volume and pressure delivered. Analytical models have been developed to establish the elongation force relationship for knitted and woven sheets. We also describe several mechanical programming techniques for smart textiles for multimodal movement, including bi-directional extension, bending, radial expansion, and the ability to transition from 2D to 3D. To demonstrate the strength of our approach, we will also integrate AMF into commercial fabrics or textiles to change their configuration from passive to active structures that cause various deformations. We have also demonstrated this concept on several experimental test benches, including programmable bending of threads to produce desired letters and shape-shifting biological structures into the shape of objects such as butterflies, quadrupedal structures and flowers.
Textiles are flexible two-dimensional structures formed from interwoven one-dimensional threads such as yarns, threads and fibers. Textile is one of mankind’s oldest technologies and is widely used in all aspects of life due to its comfort, adaptability, breathability, aesthetics and protection. Smart textiles (also known as smart clothes or robotic fabrics) are increasingly being used in research due to their great potential in robotic applications20,52. Smart textiles promise to improve the human experience of interacting with soft objects, ushering in a paradigm shift in the field where the movement and forces of thin, flexible fabric can be controlled to perform specific tasks. In this paper, we explore two approaches to the production of smart textiles based on our recent AMF49: (1) use AMF as an active yarn to create smart textiles using traditional textile manufacturing technologies; (2) insert AMF directly into traditional fabrics to stimulate the desired movement and deformation.
The AMF consists of an internal silicone tube to supply hydraulic power and an external helical coil to limit its radial expansion. Thus, AMFs elongate longitudinally when pressure is applied and subsequently exhibit contractile forces to return to their original length when pressure is released. They have properties similar to traditional fibers, including flexibility, small diameter and long length. However, the AMF is more active and controlled in terms of movement and strength than its conventional counterparts. Inspired by recent rapid advances in smart textiles, here we present four major approaches to producing smart textiles by applying AMF to a long-established fabric manufacturing technology (Figure 1).
The first way is weaving. We use weft knitting technology to produce a reactive knitted fabric that unfolds in one direction when hydraulically actuated. Knitted sheets are very stretchy and stretchable but tend to unravel more easily than woven sheets. Depending on the control method, AMF can form individual rows or complete products. In addition to flat sheets, tubular knitting patterns are also suitable for the manufacture of AMF hollow structures. The second method is weaving, where we use two AMFs as warp and weft to form a rectangular woven sheet that can expand independently in two directions. Woven sheets provide more control (in both directions) than knitted sheets. We also wove AMF from traditional yarn to make a simpler woven sheet that can only be unwound in one direction. The third method – radial expansion – is a variant of the weaving technique, in which the AMPs are located not in a rectangle, but in a spiral, and the threads provide radial constraint. In this case, the braid expands radially under the inlet pressure. A fourth approach is to stick the AMF onto a sheet of passive fabric to create a bending motion in the desired direction. We have reconfigured the passive breakout board into an active breakout board by running the AMF around its edge. This programmable nature of AMF opens up countless possibilities for bio-inspired shape-transforming soft structures where we can turn passive objects into active ones. This method is simple, easy, and fast, but can compromise the longevity of the prototype. The reader is referred to other approaches in the literature that detail the strengths and weaknesses of each tissue property21,33,34,35.
Most threads or yarns used to make traditional fabrics contain passive structures. In this work, we use our previously developed AMF, which can reach meter lengths and submillimeter diameters, to replace traditional passive textile yarns with AFM to create intelligent and active fabrics for a wider range of applications. The following sections describe detailed methods for making smart textile prototypes and present their main functions and behaviors.
We handcrafted three AMF jerseys using the weft knitting technique (Fig. 2A). Material selection and detailed specifications for AMFs and prototypes can be found in the Methods section. Each AMF follows a winding path (also called a route) that forms a symmetrical loop. The loops of each row are fixed with loops of the rows above and below them. The rings of one column perpendicular to the course are combined into a shaft. Our knitted prototype consists of three rows of seven stitches (or seven stitches) in each row. The top and bottom rings are not fixed, so we can attach them to the corresponding metal rods. Knitted prototypes unraveled more easily than conventional knitted fabrics due to the higher stiffness of AMF compared to conventional yarns. Therefore, we tied the loops of adjacent rows with thin elastic cords.
Various smart textile prototypes are being implemented with different AMF configurations. (A) Knitted sheet made from three AMFs. (B) Bidirectional woven sheet of two AMFs. (C) A unidirectional woven sheet made from AMF and acrylic yarn can bear a load of 500g, which is 192 times its weight (2.6g). (D) Radially expanding structure with one AMF and cotton yarn as radial constraint. Detailed specifications can be found in the Methods section.
Although the zigzag loops of a knit can stretch in different directions, our prototype knit expands primarily in the direction of the loop under pressure due to limitations in the direction of travel. The lengthening of each AMF contributes to the expansion of the total area of ​​the knitted sheet. Depending on specific requirements, we can control three AMFs independently from three different fluid sources (Figure 2A) or simultaneously from one fluid source via a 1-to-3 fluid distributor. On fig. 2A shows an example of a knitted prototype, the initial area of ​​which increased by 35% while applying pressure to three AMPs (1.2 MPa). Notably, AMF achieves a high elongation of at least 250% of its original length49 so knitted sheets can stretch even more than current versions.
We also created bidirectional weave sheets formed from two AMFs using the plain weave technique (Figure 2B). AMF warp and weft are intertwined at right angles, forming a simple criss-cross pattern. Our prototype weave was classified as a balanced plain weave because both the warp and weft yarns were made from the same yarn size (see Methods section for details). Unlike ordinary threads that can form sharp folds, the applied AMF requires a certain bending radius when returning to another thread of the weaving pattern. Therefore, woven sheets made from AMP have a lower density compared to conventional woven textiles. AMF-type S (outer diameter 1.49 mm) has a minimum bending radius of 1.5 mm. For example, the prototype weave we present in this article has a 7×7 thread pattern where each intersection is stabilized with a knot of thin elastic cord. Using the same weaving technique, you can get more strands.
When the corresponding AMF receives fluid pressure, the woven sheet expands its area in the warp or weft direction. Therefore, we controlled the dimensions of the braided sheet (length and width) by independently changing the amount of inlet pressure applied to the two AMPs. On fig. 2B shows a woven prototype that expanded to 44% of its original area while applying pressure to one AMP (1.3 MPa). With the simultaneous action of pressure on two AMFs, the area increased by 108%.
We also made a unidirectional woven sheet from a single AMF with warp and acrylic yarns as weft (Figure 2C). The AMFs are arranged in seven zigzag rows and the threads weave these rows of AMFs together to form a rectangular sheet of fabric. This woven prototype was denser than in Fig. 2B, thanks to soft acrylic threads that easily filled the entire sheet. Because we only use one AMF as the warp, the woven sheet can only expand towards the warp under pressure. Figure 2C shows an example of a woven prototype whose initial area increases by 65% ​​with increasing pressure (1.3 MPa). In addition, this braided piece (weighing 2.6 grams) can lift a load of 500 grams, which is 192 times its mass.
Instead of arranging the AMF in a zigzag pattern to create a rectangular woven sheet, we fabricated a flat spiral shape of the AMF, which was then radially constrained with cotton yarn to create a round woven sheet (Figure 2D). The high rigidity of AMF limits its filling of the very central region of the plate. However, this padding can be made from elastic yarns or elastic fabrics. Upon receiving hydraulic pressure, the AMP converts its longitudinal elongation into a radial expansion of the sheet. It is also worth noting that both the outer and inner diameters of the spiral shape are increased due to the radial limitation of the filaments. Figure 2D shows that with an applied hydraulic pressure of 1 MPa, the shape of a round sheet expands to 25% of its original area.
We present here a second approach to making smart textiles where we glue an AMF to a flat piece of fabric and reconfigure it from a passive to an actively controlled structure. The design diagram of the bending drive is shown in fig. 3A, where the AMP is folded down the middle and glued to a strip of inextensible fabric (cotton muslin fabric) using double-sided tape as an adhesive. Once sealed, the top of the AMF is free to extend, while the bottom is limited by the tape and fabric, causing the strip to bend towards the fabric. We can deactivate any part of the bend actuator anywhere by simply sticking a strip of tape on it. A deactivated segment cannot move and becomes a passive segment.
Fabrics are reconfigured by sticking AMF onto traditional fabrics. (A) Design concept for a bending drive made by gluing a folded AMF onto an inextensible fabric. (B) Bending of the actuator prototype. (C) Reconfiguration of a rectangular cloth into an active four-legged robot. Inelastic fabric: cotton jersey. Stretch fabric: polyester. Detailed specifications can be found in the Methods section.
We made several prototype bending actuators of different lengths and pressurized them with hydraulics to create a bending motion (Figure 3B). Importantly, the AMF can be laid out in a straight line or folded to form multiple threads and then glued to fabric to create a bending drive with the appropriate number of threads. We also converted the passive tissue sheet into an active tetrapod structure (Figure 3C), where we used AMF to route the borders of a rectangular inextensible tissue (cotton muslin fabric). AMP is attached to the fabric with a piece of double-sided tape. The middle of each edge is taped to become passive, while the four corners remain active. Stretch fabric top cover (polyester) is optional. The four corners of the fabric bend (looks like legs) when pressed.
We built a test bench to quantitatively study the properties of the developed smart textiles (see the Methods section and Supplementary Figure S1). Since all samples were made of AMF, the general trend of the experimental results (Fig. 4) is consistent with the main characteristics of AMF, namely, the inlet pressure is directly proportional to the outlet elongation and inversely proportional to the compression force. However, these smart fabrics have unique characteristics that reflect their specific configurations.
Features smart textile configurations. (A, B) Hysteresis curves for inlet pressure and outlet elongation and force for woven sheets. (C) Expansion of the area of ​​the woven sheet. (D,E) Relationship between input pressure and output elongation and force for knitwear. (F) Area expansion of radially expanding structures. (G) Bending angles of three different lengths of bending drives.
Each AMF of the woven sheet was subjected to an inlet pressure of 1 MPa to generate approximately 30% elongation (Fig. 4A). We chose this threshold for the entire experiment for several reasons: (1) to create a significant elongation (approximately 30%) to emphasize their hysteresis curves, (2) to prevent cycling from different experiments and reusable prototypes resulting in accidental damage or failure. . under high fluid pressure. The dead zone is clearly visible, and the braid remains motionless until the inlet pressure reaches 0.3 MPa. The pressure elongation hysteresis plot shows a large gap between the pumping and releasing phases, indicating that there is a significant loss of energy when the woven sheet changes its motion from expansion to contraction. (Fig. 4A). After obtaining an inlet pressure of 1 MPa, the woven sheet could exert a contraction force of 5.6 N (Fig. 4B). The pressure-force hysteresis plot also shows that the reset curve almost overlaps with the pressure build-up curve. The area expansion of the woven sheet depended on the amount of pressure applied to each of the two AMFs, as shown in the 3D surface plot (Figure 4C). Experiments also show that a woven sheet can produce an area expansion of 66% when its warp and weft AMFs are simultaneously subjected to a hydraulic pressure of 1 MPa.
The experimental results for the knitted sheet show a similar pattern to the woven sheet, including a wide hysteresis gap in the tension-pressure diagram and overlapping pressure-force curves. The knitted sheet showed an elongation of 30%, after which the compression force was 9 N at an inlet pressure of 1 MPa (Fig. 4D, E).
In the case of a round woven sheet, its initial area increased by 25% compared to the initial area after exposure to a liquid pressure of 1 MPa (Fig. 4F). Before the sample begins to expand, there is a large inlet pressure dead zone up to 0.7 MPa. This large dead zone was expected as the samples were made from larger AMFs which required higher pressures to overcome their initial stress. On fig. 4F also shows that the release curve almost coincides with the pressure increase curve, indicating little energy loss when the disc movement is switched.
Experimental results for the three bending actuators (tissue reconfiguration) show that their hysteresis curves have a similar pattern (Figure 4G), where they experience an inlet pressure dead zone of up to 0.2 MPa before lifting. We applied the same volume of liquid (0.035 ml) to three bending drives (L20, L30 and L50 mm). However, each actuator experienced different pressure peaks and developed different bending angles. The L20 and L30 mm actuators experienced an inlet pressure of 0.72 and 0.67 MPa, reaching bending angles of 167° and 194° respectively. The longest bending drive (length 50 mm) withstood a pressure of 0.61 MPa and reached a maximum bending angle of 236°. The pressure angle hysteresis plots also revealed relatively large gaps between the pressurization and release curves for all three bending drives.
The relationship between input volume and output properties (elongation, force, area expansion, bending angle) for the above smart textile configurations can be found in Supplementary Figure S2.
The experimental results in the previous section clearly demonstrate the proportional relationship between applied inlet pressure and outlet elongation of AMF samples. The stronger the AMB is strained, the greater the elongation it develops and the more elastic energy it accumulates. Hence, the greater the compressive force it exerts. The results also showed that the specimens reached their maximum compression force when the inlet pressure was completely removed. This section aims to establish a direct relationship between elongation and maximum shrinkage force of knitted and woven sheets through analytical modeling and experimental verification.
The maximum contractile force Fout (at inlet pressure P = 0) of a single AMF was given in ref 49 and reintroduced as follows:
Among them, α, E, and A0 are the stretching factor, Young’s modulus, and cross-sectional area of ​​the silicone tube, respectively; k is the stiffness coefficient of the spiral coil; x and li are offset and initial length. AMP, respectively.
the right equation. (1) Take knitted and woven sheets as an example (Fig. 5A, B). The shrinkage forces of the knitted product Fkv and the woven product Fwh are expressed by equation (2) and (3), respectively.
where mk is the number of loops, φp is the loop angle of the knitted fabric during injection (Fig. 5A), mh is the number of threads, θhp is the engagement angle of the knitted fabric during injection (Fig. 5B), εkv εwh is the knitted sheet and the deformation of the woven sheet, F0 is the initial tension of the spiral coil. Detailed derivation of the equation. (2) and (3) can be found in the supporting information.
Create an analytical model for the elongation-force relationship. (A,B) Analytical model illustrations for knitted and woven sheets, respectively. (C,D) Comparison of analytical models and experimental data for knitted and woven sheets. RMSE Root mean square error.
To test the developed model, we performed elongation experiments using the knitted patterns in Fig. 2A and braided samples in Fig. 2B. Contraction force was measured in 5% increments for each locked extension from 0% to 50%. The mean and standard deviation of the five trials are presented in Figure 5C (knit) and Figure 5D (knit). The curves of the analytical model are described by equations. Parameters (2) and (3) are given in Table. 1. The results show that the analytical model is in good agreement with the experimental data over the entire elongation range with a root mean square error (RMSE) of 0.34 N for knitwear, 0.21 N for woven AMF H (horizontal direction) and 0.17 N for woven AMF . V (vertical direction).
In addition to the basic movements, the proposed smart textiles can be mechanically programmed to provide more complex movements such as S-bend, radial contraction, and 2D to 3D deformation. We present here several methods for programming flat smart textiles into desired structures.
In addition to expanding the domain in the linear direction, unidirectional woven sheets can be mechanically programmed to create multimodal movement (Fig. 6A). We reconfigure the extension of the braided sheet as a bending motion, constraining one of its faces (top or bottom) with sewing thread. The sheets tend to bend towards the bounding surface under pressure. On fig. 6A shows two examples of woven panels that become S-shaped when one half is cramped on the top side and the other half is cramped on the bottom side. Alternatively, you can create a circular bending motion where only the entire face is constrained. A unidirectional braided sheet can also be made into a compression sleeve by connecting its two ends into a tubular structure (Fig. 6B). The sleeve is worn over a person’s index finger to provide compression, a form of massage therapy to relieve pain or improve circulation. It can be scaled to fit other body parts such as arms, hips, and legs.
Ability to weave sheets in one direction. (A) Creation of deformable structures due to the programmability of the shape of sewing threads. (B) Finger compression sleeve. (C) Another version of the braided sheet and its implementation as a forearm compression sleeve. (D) Another compression sleeve prototype made from AMF type M, acrylic yarn and Velcro straps. Detailed specifications can be found in the Methods section.
Figure 6C shows another example of a unidirectional woven sheet made from a single AMF and cotton yarn. The sheet can expand by 45% in area (at 1.2 MPa) or cause circular motion under pressure. We have also incorporated a sheet to create a forearm compression sleeve by attaching magnetic straps to the end of the sheet. Another prototype forearm compression sleeve is shown in Fig. 6D, in which unidirectional braided sheets were made from Type M AMF (see Methods) and acrylic yarns to generate stronger compression forces. We have equipped the ends of the sheets with Velcro straps for easy attachment and for different hand sizes.
The restraint technique, which converts linear extension into bending motion, is also applicable to bidirectional woven sheets. We weave the cotton threads on one side of the warp and weft woven sheets so that they do not expand (Fig. 7A). Thus, when two AMFs receive hydraulic pressure independently of each other, the sheet undergoes a bi-directional bending motion to form an arbitrary three-dimensional structure. In another approach, we use inextensible yarns to limit one direction of bidirectional woven sheets (Figure 7B). Thus, the sheet can make independent bending and stretching movements when the corresponding AMF is under pressure. On fig. 7B shows an example in which a bidirectional braided sheet is controlled to wrap around two-thirds of a human finger with a bending motion and then extend its length to cover the rest with a stretching motion. The two-way movement of sheets can be useful for fashion design or smart clothing development.
Bi-directional woven sheet, knitted sheet and radially expandable design capabilities. (A) Bi-directional bonded bi-directional wicker panels to create a bi-directional bend. (B) Unidirectionally constrained bidirectional wicker panels produce flex and elongation. (C) Highly elastic knitted sheet, which can conform to different surface curvature and even form tubular structures. (D) delimitation of the center line of a radially expanding structure forming a hyperbolic parabolic shape (potato chips).
We connected two adjacent loops of the upper and lower rows of the knitted part with sewing thread so that it would not unravel (Fig. 7C). Thus, the woven sheet is fully flexible and adapts well to various surface curves, such as the skin surface of human hands and arms. We also created a tubular structure (sleeve) by connecting the ends of the knitted part in the direction of travel. The sleeve wraps well around the person’s index finger (Fig. 7C). The sinuosity of the woven fabric provides excellent fit and deformability, making it easy to use in smart wear (gloves, compression sleeves), providing comfort (through fit) and therapeutic effect (through compression).
In addition to 2D radial expansion in multiple directions, circular woven sheets can also be programmed to form 3D structures. We limited the center line of the round braid with acrylic yarn to disrupt its uniform radial expansion. As a result, the original flat shape of the round woven sheet was transformed into a hyperbolic parabolic shape (or potato chips) after pressurization (Fig. 7D). This shape-shifting ability could be implemented as a lift mechanism, an optical lens, mobile robot legs, or could be useful in fashion design and bionic robots.
We have developed a simple technique for creating flexural drives by gluing AMF onto a strip of non-stretch fabric (Figure 3). We use this concept to create shape programmable threads where we can strategically distribute multiple active and passive sections in one AMF to create desired shapes. We fabricated and programmed four active filaments that could change their shape from straight to letter (UNSW) as pressure was increased (Supplementary Fig. S4). This simple method allows the deformability of the AMF to turn 1D lines into 2D shapes and possibly even 3D structures.
In a similar approach, we used a single AMF to reconfigure a piece of passive normal tissue into an active tetrapod (Fig. 8A). Routing and programming concepts are similar to those shown in Figure 3C. However, instead of rectangular sheets, they began to use fabrics with a quadrupedal pattern (turtle, cotton muslin). Therefore, the legs are longer and the structure can be raised higher. The height of the structure gradually increases under pressure until its legs are perpendicular to the ground. If the inlet pressure continues to rise, the legs will sag inwards, lowering the height of the structure. Tetrapods can perform locomotion if their legs are equipped with unidirectional patterns or use multiple AMFs with motion manipulation strategies. Soft locomotion robots are needed for a variety of tasks, including rescues from wildfires, collapsed buildings or hazardous environments, and medical drug delivery robots.
The fabric is reconfigured to create shape-shifting structures. (A) Glue the AMF to the border of the passive fabric sheet, turning it into a steerable four-legged structure. (BD) Two other examples of tissue reconfiguration, turning passive butterflies and flowers into active ones. Non-stretch fabric: plain cotton muslin.
We also take advantage of the simplicity and versatility of this tissue reconfiguration technique by introducing two additional bioinspired structures for reshaping (Figures 8B-D). With a routable AMF, these form-deformable structures are reconfigured from sheets of passive tissue to active and steerable structures. Inspired by the monarch butterfly, we made a transforming butterfly structure using a piece of butterfly-shaped fabric (cotton muslin) and a long piece of AMF stuck under its wings. When the AMF is under pressure, the wings fold up. Like the Monarch Butterfly, the Butterfly Robot’s left and right wings flap the same way because they are both controlled by the AMF. Butterfly flaps are for display purposes only. It cannot fly like Smart Bird (Festo Corp., USA). We also made a fabric flower (Figure 8D) consisting of two layers of five petals each. We placed the AMF below each layer after the outer edge of the petals. Initially, the flowers are in full bloom, with all petals fully open. Under pressure, the AMF causes a bending movement of the petals, causing them to close. The two AMFs independently control the movement of the two layers, while the five petals of one layer flex at the same time.