Nothing Special   »   [go: up one dir, main page]

skip to main content
10.1145/3613904.3642528acmconferencesArticle/Chapter ViewFull TextPublication PageschiConference Proceedingsconference-collections
research-article
Open access

Loopsense: low-scale, unobtrusive, and minimally invasive knitted force sensors for multi-modal input, enabled by selective loop-meshing

Published: 11 May 2024 Publication History

Abstract

Integrating sensors into knitted input devices traditionally comes with considerable constraints for textile and UI design freedom. In this work, we demonstrate a novel, minimally invasive method for fabricating knitted sensors that overcomes this limitation. We integrate copper wire with piezoresistive enamel directly into the fabric using weft knitting to establish strain and pressure sensing cells that consist only of single pairs of intermeshed loops. The result is unobtrusive and potentially invisible, which provides tremendous latitude for visual and haptic design. Furthermore, we present several variations of stitch compositions, resulting in loop meshes that feature distinct response with respect to direction of exerting force. Utilizing this property, we are able to infer actuation modalities and considerably expand the device’s input space. In particular, we discern strain directions and surface pressure. Moreover, we provide an in-depth description of our fabrication method, and demonstrate our solution’s versatility on three exemplary use cases.
Figure 1:
Figure 1: By knitting a litz wire with piezoresistive enamel, we establish minimal and unobtrusive textile force sensor cells, consisting of single loop pairs. We combine stitch type variations to provide multiple input modalities, enabling us to discern between orthogonal directions of strain, as well as surface pressure. Our sensor cells are highly sensitive, can be easily integrated into a variety of designs, and are potentially invisible at a fabric’s front face.

1 Introduction

With growing interest in unobtrusive human-machine interfaces, garments and textiles in general bear great potential to become ubiquitous user interfaces of the future. Areas of application range from wearables in the consumer and sports field [33, 41, 51] and in healthcare [5, 21, 34, 50], to smart environments in automotive [9], office [17], and home interior [7], and interaction with robots [44].
Researchers have tackled the field of textile-based user interfaces for decades. Most prominently, Project Jacquard [41] showed capacitive sensors that are well integrated into a woven fabric in a way that made them virtually unrecognizable. Challenges with woven fabrics however include their lack of stretchabilty, and – when compared to weft knitting – inflexibility of the fabrication process. This commonly requires woven fabrics to be tailored later on, which may be challenging when electronics are integrated that must not be cut.
Due to their composition, knitted fabrics on the other hand are inherently elastic and therefore ideal for stretching and draping. In the past, knitted sensors have been used as input devices for HCI, e.g., for gesture recognition [27] or for implementing haptic textile button elements for versatile scenarios [1], but also in medicine and healthcare, e.g., for tracking bio-features [34], gait-monitoring scenarios [21], or grasp classification in rehabilitation [5]. Even for space suits, 3D knitting has been discovered as a powerful means of engineering functional and complex multi-component gear with integrated sensors [19, 36].
However, traditional approaches of knitting touch or strain sensors are usually realized as an areal insert of conductive yarn which can be relatively large and distinct [5, 6, 21, 31, 36]. While such solutions work well within laboratory use, they can only sense exerted force, hence they are unable to discern between actuation types, such as strain and pressure. This makes them vulnerable to accidental activation or false positives, e.g., when a strain sensor is accidentally pressed. Furthermore, they come with a deficit in appeal and with a considerable restriction to visual and haptic textile design. Even integration into widespread patterns such as Milano, Cardigan, or Cable Designs can be limited since they may interfere with the sensor’s performance. This has also implications beyond appeal, since a knitted fabric’s properties are often designed to serve functional purposes, e.g., to provide a required degree of stability, recovery, or anisotropic extensibility.
With Loopsense, we present a solution for knitting resistive sensor cells that unite multiple materials and therefore provide exceptional design latitude. Our knitted sensors are constructed from a special copper core litz wire enamelled with piezoresistive material [35, 38]. By using this material in our knitting process, we are able to scale down the actual sensor cell to a single intersection of two wires, which can be accomplished by a range of basic stitch types that common weft or v-bed knitting machines provide. In this paper, we present the results of our exploration into physical sensor cell construction, i.e., loop meshing, as well as integration into knit designs. By utilizing different types of stitches, we are able to fabricate sensor geometries exhibiting distinct responsiveness with respect to actuation (cf. Figure 1). As a result, our novel textile devices can be used to discern strain along two directions as well as orthogonal pressure. This is a fundamental advantage over state-of-the-art knitted sensors, as it not only extends the fabric device’s input space and expressiveness, but along its small scale and potential unobtrusiveness it also promises innovative utilization in a variety of scenarios of interaction with textiles. Moreover, our method does not require manual postprocessing or stacking of multiple fabric layers and can be applied to virtually any fully-fashioned or 3D knitted fabric design, since the sensor cells are established entirely throughout the knitting process.
Summarizing, the five major contributions of this paper are:
basic concept and fabrication method for a wire-based knitted force sensor that is scaled down to its absolute minimum,
an investigation into loop meshing options for both single- and double-faced knits, spawning sensors that show distinct response with respect to actuation type, including results of a formal characterization,
exemplary combinations of those sensors in order to recognize and discern actuation modes (pressure and strain along orthogonal axes),
a demonstration of use-case scenarios as well as integration into three exemplary fabric designs, establishing unobtrusive sensing fabrics with vast design latitude,
a list of essential insights for replicating our work, including details on handling of material and machinery, as well as a summary of guidelines to avoid common pitfalls.

2 Related Work

First advances into the field of sensing textiles and garments have been made more than two decades ago. An early example is the work of DeRossi et al. [8], which coated conventional fabrics with a conductive polymer to sense strain and temperature. Post at al. [40] used embroidery to augment existing fabrics with conductive yarns to embed electronic components and establish recognizable user interfaces operated by capacitive touch. More recently, Strohmeier et al. [46] demonstrated a multi-layer patch for hybrid resistive/capacitive input. Honnet et al. [14] showed a method of functionalizing a wider range of base material by polymerization, equipping them with piezoresistive properties. In contrast to these works, we target manufacturing-time integration, instead of relying on modification of existing material.
Related to this objective, there has been considerable progress in recent years: Poupyrev et al. [41] showed a conductive yarn that was fit for weaving processes which was later integrated into the Levi’s Trucker Jacket available for purchase. Olwal et al. [33] demonstrated conductive yarn integrated into braids, both used capacitive sensing techniques to detect finger hover and touch. Devendorf et al. [11] demonstrated tools for supporting design and fabrication of woven user interfaces. While some of these solutions are hard to customize, we focus on weft knitting, which enables to design fully-fashioned fabrics in high structural detail and a combination of numerous raw materials.
Driven also by the increasing accessibility of knitting technology in recent years [26, 32], there has been growing interest in this field. Wicaksono et al. [49] showed a multi-layered knitted musical keyboard that combined capacitive and resistive sensing methods. Vallett et al. [47] showed a flexible knitted keyboard based on a capacitive sensing and swept frequency Bode analysis. McDonald et al. [28] built on-top of this and integrated the material into the knit using a spacer knit structure for better haptic appeal and discoverability. In [27], they investigated user interface scenarios based on this technology, focusing on touch and gesture input. Aigner et al. [1] demonstrated a haptic spacer knit element based on resistive sensing, that could be used for continuous input of pressure. Yu et al. [52] showed a knitted scarf for gesture input using passive electrical impedance tomography. In contrast to these works, the purpose of our work is to enable detecting strain as well as pressure. Furthermore, we aim at maximum compatibility with a wide variety of knitting structures, while giving the designer the opportunity to completely hide away the sensing parts.
Leong et al. [20] presented a multi-layered stretchable sensor matrix for sensing pressure on prosthetic limbs. Similarly, Wicaksono et al. [49, 50] knitted custom pressure matrices and permanently fused layers together with melting yarn for additional stability. They demonstrated sensing mats and socks for posture tracking, gait analysis, and to drive reactive audio systems during dance performances. Similarly, Luo et al. [25] showed a tool for knitting building blocks for multi-layered knitted sensors, to realize custom knitted UI layouts. We add to this body of work by presenting a method for a fully-integrated manufacturing of potentially numerous sensors in a single fabric, which can be fabricated ready-made and therefore does not require stacking, aligning, or fusing of multiple layers, since we insert a multi-material litz wire into the knitting structure that combines conductive and piezoresistive properties.
Parzer et al. [35] showed a wire with similar properties, however they did not go beyond sewing and weaving. Luo et al. [24] presented a knitted multi-layer sensor matrix for sensing pressure distribution. While the functional fibre was of a conductive core and a piezoresistive sheath, it was comparable to ours in terms of composition, it was unsuitable for knitting with a reported diameter of 0.6 mm and therefore inlaid in a straight trace, supported by the surrounding loops. Pointner et al. [38] demonstrated a knitted sensor using similar material, however the integrated wire was clearly visible on both sides and structurally disruptive particularly on the back side of the fabric.
However, for those solutions to be attractive for most everyday scenarios, they have to provide appropriate integration methods that enable textile and UI designers to be in control of the visual and haptic appearance. This is important in particular in the field of textiles, where sociocultural aspects have a big significance. Devendorf et al. [10] performed artist residencies where HCI practitioner teamed up with engineers and craftspeople to bridge this gap in the domain of woven user interfaces. More recent projects that consider design aspects in textile based technology include Project Brookdale [43] and Skill-Sleeves [18], where researchers team up with designers and artists to explore collaborative design and development. Further notable references that investigate traditional textile crafting in the context of technology are Irene Posch’s works, e.g., "Knitted Radio" [39], where she explored alternative production procedures. Finally, the research of Mlakar et al. [29, 30] showed that design aspects are essential for designing affordances without being limited by the sensing components. Our solution provides considerable latitude, so knit structures and compositions can be designed with little constraints.

3 Implementation

3.1 Resistive Sensing on Knitted Fabrics

Figure 2:
Figure 2: (a) Traditional knitted sensor structure (left) vs. our approach (right). (b) FSRs are subject to the interface effect [48]: applied stress increases the contact area and therefore conductivity; our wires utilize the same effect, i.e., every intersection of wire pairs represents an FSR cell.
The principle of resistive sensing is commonly applied, e.g., to knitted strain sensors [4, 37], however, in traditional approaches, areas of conductive yarn are inserted using tubular structures [22], Intarsia techniques [23], or plating [21], allowing the current to pass along the loops’ intermeshing points. The pressure between yarn loops at those points effects a piezoresistive property of the structure, according to Holm’s Theory [13]. This is corresponding to the interface effect [48] in common force sensitive resistors (FSR), where external stress varies the surface contact and thus alters conductivity (cf. Figure 2 b). Equivalent to metallic electrodes of an FSR, connector traces are occasionally realized with yarn of considerably higher conductivity than the yarn used for the sensing part [6] (cf. Figure 2 a). This is beneficial to minimize their contribution to the sensor readings [1, 2], since the connector loops are just as well subject to the sensing principle, which would introduce undesired sensor response. However, this solution comes with several downsides: it (i) requires yet another type of functional yarn which (ii) needs to be properly integrated and is therefore structurally constraining. This in turn (iii) consumes additional space/loops along the sensor edges.
To overcome this, we use a wire with metal core and piezoresistive coating [35], which allows us to unite conductive and resistive parts into one thread. This furthermore enables us to realize both sensing and connecting parts using with a single material. Moreover, the actual sensor is scaled down to the coating volume between the metal cores of two intersecting threads (cf. Figure 2 b). This way, we are able to implement a minimal FSR sensor cell on loop level. Since apart from these force-sensitive wire intersections current is flowing along the copper wire, we entirely eliminate sensor noise coming from the connector parts. Our sensor therefore drastically reduces structural complexity and space constraints, providing superior latitude also for aesthetics, i.e., visual and haptic textile design. Furthermore, the freedom in intersection formation on a loop-level that is provided in particular by V-bed knitting opens up a whole range of potential functional arrangements and specifically targeted sensor engineering (cf. Section 4).
Figure 3:
Figure 3: (a) Various parts of a V-bed knitting machine. (b) Storage feeder used for achieving consistent wire tension. (c) Top: closeups of crimps used to connect TextileWire to copper wire leading to electronics (reaching out to the right). Bottom: crimp covered by heat shrink tube to avoid accidental shorts with nearby crimps. (d) Repeat of our substrate patterns that we used for evaluating our sensor candidates.

3.2 Material

Our sensor yarn is a TextileWire TW-F1 (TW) from Elektrisola GmbH & Co KG2 with a core of electrolytic copper (99.95% purity), coated by a Polyamide-imide enamel, which provides excellent thermal, chemical, and mechanical durability, according to vendor data sheet. The wire was previously presented in [35], including an in-depth characterization and essential fabrication details. The enamel is enriched with particles of Carbon Black, which makes it electrically conductive with a sheet resistance of 1.7  kΩ /sq. Note however that the enamel’s resistivity is nonessential for this work, since we operate with relative change of resistive and our results are therefore translatable to other materials with similar piezoresistive characteristics. For knitting, we used a litz wire consisting of 12 ends à ø 48 µm (nominal; outer diameters 58.5 ±8.5 µm, corresponds to dtex 152), twisted with a twist pitch of 4 mm in S direction. The resulting litz provides a tensile strength of 7  N, where it breaks at an elongation of 17%.
For the surrounding knits, we used Polyamide (PA) yarn from TWD fibres, with dimensions of dtex 78/23/2 (yellow and black), dtex 78/17/2 (red and gray).

3.3 Fabrication

We knitted all of our samples using a 15-gauge Shima Seiki SWG061N2 V-bed machine. In order to speed up our prototyping workflow, we implemented a pipeline for generating our knitting programs using JavaScript, building upon the Knitout high-level knit description language [26]. Generated Knitout files were transpiled into Shima Seiki’s proprietary DAT format, and then compiled to machine executable format in APEX4 suite version S-04.C. This streamlined process enabled us to iterate quickly and experiment efficiently with numerous structures, materials, and settings.
Adequate unwinding of non-elastic material is highly challenging mostly due to non-linear yarn demand during the knitting process, where yarn carriers pull and relax the material frequently and jerkily; the machine’s spring-based tensing system proved insufficient to compensate for this. We therefore equipped our machine with an MSF 3 storage feeder from MEMMINGER-IRO GmbH3 with magnetic yarn tensioner (cf. Figure 3 b) to better control the feeding conditions: the winding body buffers 8 meters of TW closely to the yarn feeder so it does not have to be not pulled from the spool directly. Note that the device is not coupled to the knitting machine and therefore the knitting process, however it operates using optical sensors for inlet, feed rate and yarn speed, which we found sufficient to greatly reduce complications such as wire breaks caused by kinks and slack wire (cf. Section 8.2). Additionally, we knitted the TW with a carriage speed of 0.02 m/s, which furthermore reduced complications such as needles missing some of the wire strands, which occasionally occured otherwise.

3.4 Providing Connections to Sampling Electronics

Beyond the sensor cell, the wire has to be guided through the substrate by what we term connector traces, bridging to a location for convenient connecting of electronics, e.g., at the swatch’s edge. The most straightforward implementation in weft knitting is to lay them out along courses, which is the primary knitting direction. It is however crucial to have an adequate number of courses of regular (i.e., non-conductive) yarn in between, which act as an insulation, since the two strands of litz wire must only make contact at the sensor location4. For fixation, there are three options of combining it with the regular yarn: the wire can be (i) knitted along the regular yarn, it can be (ii) tucked into the yarn’s loops, or in case of two-sided knits it can be (iii) tunnelled between front and back faces using miss stitches. We experimented with those three options and found the use of knit-stitches performed best, however these additional loops could be disruptive to the appearance of the knit, so the yarn’s repeat at the corresponding course may have to be adjusted to compensate for this.
Since the piezoresistive coating is thermally resistant and therefore not solderable, we found that crimping is a viable solution for attaching, e.g., enamelled copper wires as connectors (cf. Figure 3 c). The physical pressure causes the crimp terminal to crack the TW’s coating and establish exceptional and durable contact with the copper core.

4 Utilizing Knit Mesh Topologies

Since our sensor design requires merely a pair of intersecting wires, we were interested in the implications of different sensor cell structures that may show vastly different mechanical features. Given the potential geometric complexity of knitted fabrics, there are numerous possibilities to establish these intersections, or to embed them in the surrounding knit. We observed that in certain structures, threads tend to either compress or dodge, depending on which direction the fabric is strained. One common limitation of resistive knitted sensors is while the amount of stress can be inferred, the cause is unknown. We hypothesized that, especially for double-faced fabrics, the mesh geometry can be designed in a way that threads have little to no contact except if compressed by an external force, making them ideal for sensing pressure exclusively, and vice versa. Other structures may instead be most affected by strain forces. We see these features closely related to structural metamaterials [15, 16], and believe they can be purposefully engineered with the use of mechanics and rope statics. However, the physical complexity of a knit topology as well as non-trivial material properties render this a highly challenging task and we therefore argue at this point a reliable simulation by computational means may be impractical. Hence, to advance into this direction, we restricted to an empirical analysis of a selected number of sensor compositions that we manufactured and inspected. As in V-bed knitting, there are numerous options to intermesh yarn, we focused on the most basic ones that we judged least specific to particular pattern repeats and therefore most portable to other solutions.
Figure 4:
Figure 4: Variations of sensor cells developed for Plain (i.e., single-faced) knits that we fabricated for evaluation. Knit: both wires form loops that provide firm and consistent contact between head and feet. Front-Miss: the upper wire is floating in front of the legs of the lower, securing actual contact between the wires, which is not guaranteed when using a standard miss stitch. Tuck: wires are only making contact at the lower loop’s head, however the upper yarn is able to move, shift, and slip, which may cause inconsistencies.

4.1 Single-Faced Knits

As a substrate pattern for an exploration of possibilities within the domain of single-faced fabrics, we decided for the most basic one possible, with is arguably a Plain knit (aka. Single Jersey, cf. Figure 3 d). Again, we integrated two courses of wire, which were separated by two courses of yarn, to avoid accidental contact between upper and lower connector traces loops, which were knitted once every three wales. Knitout and DAT files of all our samples are included in the supplementary material for more details.

4.1.1 Structures.

We started with variations based on the common stitch types to form intersections: knit, miss, and tuck (cf. Figure 4, note however that for sake of simplicity, we omitted wire-separating courses in the dot notation diagrams). We knitted our samples with three ends of red PA (dtex 78/17/2), with stitch settings 34/24 for PA and 30/0 for the wire. To prevent the fabric from curling, we surrounded it with a rib framing.

4.1.2 Evaluation Apparatus.

We characterized our sensor swatches using a custom-built tensile tester (cf. Figure 5) that we equipped with a Sauter CP 10-3P1 single-point load-cell with nominal load of 10 kg for measuring force. To use the machine for both strain and pressure tests, we built replaceable actuators that could be moved along three axes via Art-Soft Mach4 CNC Control Software, which we used to run test procedures that were scripted in G-Code. Out of Mach4, we continuously sent axes coordinates, timestamp, and procedure state (actuate, release, park) via UDP using custom LUA scripts. In parallel, we sampled the load cell as well as the sensor signal using an ADS 1231 (24 bit) as well as an ADS 1115 (16 bit) ΔΣ ADC, both operated by a ESP32 boards. We measured the sensor resistance using voltage divider with a reference resistor of 150 Ω. Data including timestamps was sent from the ESPs to a PC via USB and was written to CSV files, along with the actuator data out of Mach4. Syncing these different data sources resulted in an effective sampling rate of \raise.17ex\(\scriptstyle \mathtt {\sim }\)40 Hz.
Figure 5:
Figure 5: (a) Our custom-built three-axis tensile and strain tester with load cell. (b) For recording wale-directional and (c) course-directional strain data, we attached the fabrics using mounting clamps and moved the machine tool back and forth along X. (d) For recording pressure data, we attached a stamp-actuator and moved the machine tool up and down along Z.
Figure 6:
Figure 6: Plots of captured data from the plain sensor variants: for wale-directional strain (left) and course-directional strain (center), force F (dashed) and strain e (dotted) are superimposed. For press (right), we plotted pressure P (dashed) and actuator deflection z (dotted). Solid lines show the 200-cycle average, areas represent the respective SDs.

4.1.3 Procedure.

For recording strain data, we mounted our swatches in clamps that were equipped with pin needles for pinching through the fabric so it would not slip during straining. We marked areas of 2 cm × 2 cm with the sensors centered and mounted the swatches accordingly (cf. Figure 5). We pre-strained to 25% and then moved the actuator repeatedly along the X-axis with a jog rate of 1.6 mm/s, straining the textile between 25% and 50% without dwelling in between, which resulted in average peak forces of \raise.17ex\(\scriptstyle \mathtt {\sim }\)8 N for our Plain knit swatches. Pre-straining was required since knitted fabrics are subject to wear-out and would therefore bulge when returning the actuator back to X = 0, introducing inconsistencies in our recordings. We repeated the actuation 200 times along wales and 200 times along courses.
For pressure tests, we placed the textiles on rubber sheets, secured them with tape so they could not shift, and repeatedly pressed with a circular PMMA actuator with a diameter of 10 mm. Despite the tester’s low jog rate of 0.8 mm/s along Z, we taped a piece of Polyethylene (PE) foam (diameter: 10 mm, thickness: 10 mm) to the actuator’s face, to get a more smooth transition in pressure amplitude and therefore better quantization in our recorded data. Otherwise a slight change in Z induced a huge change in pressure which degraded the data’s quality. The PE foam provides exceptional stability and showed no signs of material fatigue even after numerous repetitions, nonetheless we used a new one for each swatch. We moved the Z-axis along 6 mm, which exerted an average peak pressure of \raise.17ex\(\scriptstyle \mathtt {\sim }\)200 kPa, which was again repeated 200 times.

4.1.4 Results.

In Figure 6, the raw captured data is plotted on a timeline along force (left and center for wale and course directional strain), as well as pressure (right), with strain/pressure (dashed line) and actuator trajectory (dotted) overlaid. Solid lines show the 200-cycle average. Low-opacity areas represent SDs, which hints towards some inconsistencies, however the signals were largely similar along all 200 repetitions, apart from reasonable drift and settling effects. Most interestingly, we can discover different behavior of sensor types regarding actuation mode, as hypothesized: while knit sensor is most responsive to strain, the front-miss variant shows higher sensitivity to pressure. The tuck sensor slightly responds strain, but hardly to pressure. Even though there are inconsistencies and a more extensive evaluation could provide a precise characterization of each variation, the most interesting outcome of this study is that the sensor geometry indeed seems to have a large impact on the responsiveness with respect to actuation type. We explored this property further by constructing more complex knitting structures, which we show in the following. The ultimate objective is the utilization of this feature as a valuable asset for increasing the sensor’s expressiveness and increasing the fabrics’ DOF, as will be demonstrated in Section 5.

4.2 Double-Faced Knits

Arguably, a Double Jersey structure is the double-faced counterpart of a Plain knit and would be an obvious candidate for our exploration. However, due to its extraordinary elasticity and tendency to wear out quickly, we found it unsuitable and used an Interlock instead (cf. Figure 3 d), which is a reasonably simple and common structure and provides better stability and recovery. We fabricated our samples using three ends of yellow PA (dtex 78/23/2) with stitch settings of 35/30 for PA and again 30/0 for the wire.

4.2.1 Structures.

In contrast to a single-faced knit such as a Plain fabric, a double-faced knit provides numerous options of engineering wire intersections. Utilizing the possibilities of loop transfers, merging with loops on the opposite beds, or using split stitches are just a few examples. We expected that each of these compositions comes with slightly different geometric behavior on loop level, and therefore of the sensor cell’s response signal to different actuation types. In an explorative ex-ante study, we manufactured a large number of different sensor designs mostly for visual microscopic inspection, however, investigating the extensive possibilities of a two-faced fabric with all combinations is beyond what can be included in this paper, so we focus on a low number of sensor cell compositions. Our criterion for selecting them was simplicity and therefore transferability to a larger variety of base patterns, i.e., beyond the Interlock that we used. We furthermore removed structures that did not result in operational sensors.
Figure 7:
Figure 7: Variations of sensor cells developed for Interlock (i.e., double-faced) knits, including simplified knitting instructions using dot notation.
We report four variations of type knit, one of type miss, two of type front-miss, and two of type tuck. Figure 7 provides a overview of the compositions, including a simplified description using dot notation, where we again omitted wire-separating courses for a more compact illustration. To get extensive details about the knitting programs, we refer to the Knitout and DAT files in the supplementary material.

4.2.2 Procedure.

We again tested our swatches using the procedure described above for the Plain knits, however due to the different physical features of the Interlock, especially the relation between force and strain, as well as fabric thickness, the force and pressure ranges were slightly different: straining by 50% resulted in average peak values of \raise.17ex\(\scriptstyle \mathtt {\sim }\)10 N, while pressure tests applied up to \raise.17ex\(\scriptstyle \mathtt {\sim }\)250 kPa.

4.2.3 Results.

Plots in Figure 8 again show the 200-cycle average (solid lines) and SD (areas) data on timelines. Most knit types respond well to strain and reasonably to pressure. Most notably, types A and D show antipodal response with respect to train direction, suggesting they could be valuable to discern direction of actuation. Regarding pressure, type A shows highest sensitivity, while all remaining types perform similarly.
All miss variations showed good response to pressure, while they showed relatively low sensitivity to strain, in particular along wale direction, where front-miss type B shows no response at all. This suggests that elaborate combinations with knit sensors can support discerning between strain and pressure actuation, by mutual exclusion based on the sensor readings. Note that our Interlock miss variants exhibit an electronic break (G=0) when they are at rest, which can be utilized as a revealing characteristic. Plots in Figure 8 only include relevant measurements with G > 0.
Finally, tuck variations feature singular properties: while none responded to strain along wale direction, type A shows sensitivity to pressure only, and type B shows sensitivity to course-directional strain exclusively.
Overall, we observe low relative SD, i.e., good consistency throughout many iterations. We noticed long-term drift and settling, which we expect to be mitigated by draping the textile, i.e., by keeping the knit slightly strained permanently, which would be the case in many potential use case scenarios. While more effort into solid fabrication and more extensive evaluation of every variation could yield a more precise characterization of each of the sensors, our main interest is in harnessing the lessons learned: the sensors’ largely varying characteristics suggest that well-considered combinations of different types can be a valuable asset for inferring actuation mode, e.g., by utilizing sensor fusion methods. For this work, we see this as the most valuable outcome.
Figure 8:
Figure 8: Data captured from our Interlock sensors: again, force F (dashed) and strain e (dotted) are superimposed for wale- and course-directional straining tests (left and center column). Pressure P (dashed) as well as actuator deflection z (dotted) are plotted for press tests (right column). For better readability of the diagrams, we grouped into stitch type categories, i.e., we present knit (top), miss (center), and tuck (bottom) variants in individual rows. It is clearly visible that our sensor variations exhibit distinct profiles with respect to actuation, suggesting a combination of those could be valuable for discerning actuation modality, and/or to learn about the fabric’s current geometric state.

5 Expanding Input Expressiveness Via Sensor Fusion

As discussed, our evaluation’s results suggest that it is possible to combine multiple sensor types and utilize their response profiles’ diversity to infer high-level actuation and deformation states. Since the sensors are exceptionally low-scale, it should be viable to do so on a relatively small area of 2 to 3 cm2, or even smaller, depending on knitting machine’s gauge. For sake of mere demonstration, we present a straightforward and prototypical approach of harnessing the sensors’ features: we implemented sensor fusion using simple data processing methods, such as arithmetic combination of sensor readings, as well as smoothing and thresholding. Our method is built to operate on live data, hence it only utilizes past sensor readings. Due to its low complexity, it provides output at an interactive rate with negligible processing load and can therefore be run on a small-scale embedded device with little computational power. A detailed description of the involved processing steps can be found in the supplementary material.
Table 1:
sf 1walecourse
kD11
tB01
kD × (1 – tB)10
kD × tB01
Table 1: Sensors of type knit D (kD) and tuck B (tB) show largely different response with respect to actuation modes, therefore basic metrics for discerning those modes can be engineered by simple algebraic operations.
Table 2:
sf 2 & 3walecoursepress
kA111
mfA011
tA001
kA × (1 – mfA) × (1 – tA)100
kA × mfA × (1 – tA)010
kA × mfA × tA001
Table 2: By joining three sensor types that provide adequately oppositional response profiles, all three actuation modes can be inferred. In our example, we combined knit A (kA), front-miss A (mfA), and tuck A (tA).
We fabricated three sensing fabric prototypes, using an Interlock substrate, with a combination of integrated knit, front-miss, and tuck sensors (cf. Figure 9).
Figure 9:
Figure 9: We chose three combinations of sensors for exemplary sensor fusion swatches: sf 1 combines knit D with tuck B, with the objective to detect strain direction. Moreover, we implemented sf 2 to also detect pressure next to strain, with knit A and tuck A, and added front-miss A. In a third implementation sf 3, we combined the upper two sensor cells by sharing one wire, to save space and simplify fabrication and wiring. Pictures on the right show that with this sensor combination, we are able to discern between wale-directional (top) and course-directional strain (center), as well as surface pressure (bottom).
For our first swatch, for reasons of visual communication of functionality, we chose to have our sensors visible on the A-side of the fabric. In the first step, our objective was to discern between strain along course and wale directions; we therefore integrated a knit D and a tuck B, since we found the response signals of those a good combination of mutual exclusiveness to operate activation functions for both actuation modes (cf. Figure 8). In a second variation, our objective was to additionally detect pressure, therefore we added a front-miss A as a third sensor to get an additional input feature. In order to have the sensors hidden from the A-side of the fabric, we combined it with a knit A and tuck A. In a third swatch we successfully united the upper two sensors (tuck and knit) to a single dual-mode sensor (cf. Figure 9), since the upper loop of the knit A can simultaneously serve as the lower loop for tuck A. This lowers complexity in programming, knitting, and wiring, and furthermore reduces the space required as well as interference with the knitting pattern. We refrained from merging all three sensors together at this point, since we expected the behavior of a combined front-miss and knit could deviate too much from the evaluated versions and as a result our reference data could be rendered invalid.
We fed the sensor features into signal processing pipelines that we tuned specifically for each of the swatches. The underlying principle is simple and relies on mutual masking and exclusion. Table 1 shows a simplified process: assuming ideal and normalized input for the sf 1 swatch, multiplying the signals of knit D (kD) and tuck B (tB) yields a measure representing wale-directional strain, since both sensors respond to this actuation. Conversely, multiplying with a flipped signal of tuck B (1 – tB) yields the opposite, which must represent course-directional strain, since tB does not respond to this modality and therefore stays at 0. Table 2 illustrates how this concept can be extended to a third mode, by combining three sensors with distinct response profiles. Since in practice, the sensor readings are not as distinct and ideal, we applied scaling, clamping, and thresholding to achieve or estimate the required masks. Most essential is however to chose sensor types that provide response features that are adequately distinct in order to minimize false-positives. Serving as an example implementation, processing details for the sensor fusion swatches presented in this paper can be found in the supplement. Figure 9 (right) shows a demonstration of all three actuation modes.
In this section, we presented swatches that joined two sensors to infer two modes, as well as three sensors to infer three modes. However, we expect that by join an ideally complementary pair of sensors, it should be possible to infer all three actuation types from just two sensors. We can see in Figure 8 that knit B and tuck B could be promising candidates for such a combination, since kB responds to strain along wales, tB responds to strain along courses, and both respond to pressure. This would further downscale the required sensing area and reduce manufacturing and wiring complexity. We did implement this prototypically, and already gained first promising results, however, they were not robust enough the be included in the paper.

6 Use Cases

6.1 Motion-Tracking Glove

We show the potential of subtle integration of strain sensors into a common garment on the example of a knitted glove, with two sensors of type Plain knit, which were placed on the thumb and index finger for tracking finger posture (cf. Figure 10 a). The sensors were positioned at the far end of the proximal phalanges, in order to sense strain. Note that this concept can be easily extended to other fingers, and furthermore to multiple phalanges per finger, in order to sense individual joints if more accurate mapping is required. For simplicity, the Textile Wire was crimped and soldered in-situ. However, traces can be guided to connectors at the glove’s wrist by more sophisticated methods, e.g., using tunneling or advanced knitting techniques5. For illustrating the performance of our glove we created a basic visualization in Unity3D that showed a 3D model of a robotic hand with finger joint angles controlled live by continuous sensor data. In terms of signal processing, we slightly filtered our sensor signal using an exponential smoothing filter and calibrated to minimum and maximum values, for linear mapping sensor readings to the required value range.

6.2 2-axis analog controller

To demonstrate control of two continuous values by orthogonal strain, we implemented a controller for an application mimicking the popular game "Angry Birds" (cf. Figure 10 b), since its game mechanics are well-known and adequate for the given input modality. The goal is to operate a slingshot by specifying strength and angle in order to hit and knock over targets. We modified a Unity3D template by ggghostmaker6 so it can be controlled accordingly. We used the sf 1 swatch presented in Section 5 as a controller and mapped course-directional strain to strength and wale-directional strain to angle. We implemented a two-step input, letting the player specify strength and angle in this order. Both values can be adjusted using slight pulling on the fabric; to quit the respective input modes, the fabric has to be released quickly, which causes a steeply falling edge in the sensor signal that is easily detected.

6.3 Textile Musical Keyboard with Multi-Functional Keys

In order to present a use-case scenario that combines pressure and strain input, we created a device with three input areas, to mimic a minimalist musical keyboard. It features velocity sensitivity and pitch bend for each area individually (cf. Figure 10 c). We knitted a complex 4-yarn swatch with three Interlock areas inserted, which served as visually and haptically recognizable key areas. In each of these areas, we integrated a sensor of type knit A and a sensor of type tuck B, to discern strain and pressure. We created a patch in the visual programming multimedia tool Pure Data7 (Pd), which generated MIDI messages from our input using the makenote and noteout objects. Virtual MIDI driver LoopBe18 routed MIDI to Ableton Live 119, where notes were synthesized using an E-Piano Basic instrument. Adapting the method described in Section 5, we multiplied values of both sensors to get course-directional strain and pressure metrics. Data was previously normalized and smoothed using an exponential smoothing lowpass-filter. MIDI notes were triggered out of Pd whenever the pressure signal passed a manually adjusted threshold; the rising slope’s amplitude was mapped to note velocity. Duration of notes was hardcoded in our implementation, however, the Pd patch could be easily extended to send MIDI note on and note off according to sensor press and release events. Moveover, we implemented a pitch bend feature using the bendout object of Pd, which we fed with the accordingly scaled strain sensor signal.
Figure 10:
Figure 10: (a) We integrated sensors into thumb and index finger of a knitted glove and used them as strain sensors to track continuous finger bend. (b) We demonstrate the ability of discerning strain direction for operating two parameters (strength and angle) of a slingshot in a computer game. (c) We knitted a minimal musical keyboard for controlling key velocity (pressure) and pitch bend (strain).

6.4 Further Example Use Cases

Apart from the basic use case scenarios we demonstrated, we believe several more complex ones can be easily imagined, for example sensors could be integrated at key positions of draped fabrics, e.g., for tracking activity and posture, or geometry-aware fabrics that track their deformation state for 2.5D surface reconstruction. Sensors could further be integrated into comfortable knitted motion-tracking garments for VR, tracking joint bend as well as contact pressure. Integrating a high number of sensors into a larger 3D knit could be used to track geometry deformation and touch or collision in real time. Furthermore, such fabrics could be used in the field of human-robot-interaction as a form of artificial skin or to control robots by touch [24]. Moreover, our sensors can be valuable in scenarios where strain and pressure has to be discerned in order to prevent false positives, i.e., accidental actuation by motion of and in the textile.

7 Integration into Universal Knitting Patterns

One of our objectives was to increase textile design latitude by providing sensor systems that can be concealed within knitted structures and are therefore unobtrusive or even entirely unrecognizable. The focus is in increasing visual and haptic appeal while providing full functionality for preserving UI expressiveness. As discussed, we achieve this by reducing the component count, since all required materials are merged into one wire. In the following, we present a basic demonstration of integration, using sensor made merely of knit stitches (corresponding to Plain knit or Interlock knit A of Section 4), which is arguably the variant most portable among numerous knitting patterns.

7.1 Design Scope

In terms of textile design, the following limitations have to be kept in mind: first, the physical behavior of the wire differs substantially from common textile fibers, in particular in elasticity and rigidity. This may introduce slightly noticeable irregularities in the resulting knit, therefore finding the wire’s adequate pattern repeat can require exploration by trial-and-error for the particular knitting structure at hand. Second, for reasons of durability, it may be beneficial to hide away the coated wire from the fabric’s side that is most exposed to abrasion, even though the coating used in this work is reasonably resilient. This is generally possible in two-sided knits, where one side completely covers the opposite.

7.2 Multi-Point Sensing Fabrics

For demonstrating our method’s versatility, we designed a line of three fabrics that integrate four force sensors in a 2 × 2 arrangement in rectangular swatches of ca. 12  cm × 12  cm. All patterns are double-faced and vary in complexity and number of required yarn carriers (cf. Figure 11). Our objective was that sensors must only be visible at the fabrics’ B-sides and entirely unrecognizable on the A-sides. We implemented the tracing parts using knit-stitches along the entire course, using knit stitches at every second needle to be visually compatible with the patterns: sensor courses marked "S" are knitted at needles I, III, V, etc.).
Figure 11:
Figure 11: Repeats of the three designs we chose for demonstration. For sensor cell areas, local adjustments are required.
Figure 12:
Figure 12: (a) Closeups show that sensor cells are entirely unrecognizable at respective A-sides. (b) Schematic of our 2  ×  2 multi-sensor fabrics developed from those patterns. (c) A prototypical demo visualizes sensor readings, in this case the readings from diagonal strain.
To create a sensor cell, we require two courses of textile wire, which need to be reasonably separated, so the tracing stitches do not accidentally touch off-sensor. To establish the intermeshed wire loops required for the sensor, we skip all knit and tuck stitches on the according needle between those two wire courses (cf. Figure 11, needle III). This way, the lower wire’s loop remains on the needle until the upper wire is knitted, providing the required contact at this wale. All other loops are cleared by regular yarn that separates the two wires along the entire courses, establishing the required insulation. We included Knitout and DAT files of all our knits in the supplementary material for further details and clarification.
For simplicity, we laid out the connector traces along single courses. Note that, since a single sensor cell requires a pair of individual wires, it is not straightforward to fabricate multiple sensors side-by-side using this layout. We decided to bypass this by offsetting our sensors slightly along wale-direction (i.e., "vertically", cf. Figure 12b). To create two sensors each on the upper and lower half of the swatch, this would require four wire courses each, however, we spare one by merging the inner two, thus we require only three for each half. We connect the center wire (pin 2) to ground, and obtain resistance values of sensors RA and RB by sequentially measuring voltage between pins 1-2, and 3-2, respectively. We repeat this for sensors RC and RD of the upper half of the swatch. We operate our sensors using an ESP32 board, using the on-board 12-bit ADCs for sampling voltage dividers with Rref=150 Ω, connected to 3.3 V via individual GPIOs of the ESP32 which were set in the firmware. We set the respective GPIO to HIGH and let all others float, to avoid potential leak currents between ADC channels and to prevent potential power supply voltage drops. The firmware code we used can be found in the supplementary material.

7.2.1 Single-colored: Swiss Weave-Knit Pattern.

Our first fabric is a Swiss Weave-Knit, knitted with three ends of black PA in a single yarn carrier with stitch settings 28/18 for the PA and 30/0 for the wire. Since the repeat contains several miss stitches, the fabric shows subjectively good stability along course-direction. We integrated the wire on the front bed, by adding one extra course to the repeat. Courses 2 and 4 had to be slightly modified (2’ and 4’), to omit knitting on front needle III. At wales I and V, there are four loops of regular yarn knitted between lower and upper wire, providing adequate insulation between traces.

7.2.2 Two-colored: Honeycomb Pattern.

The second fabric is based on a Honeycomb structure, knitted with two carriers (two ends of gray PA and two ends of black PA) with stitch settings 36/26 for PA and again 30/0 for the wire. As the pattern includes more tucks than misses, it shows good elasticity in particular in course-direction. We preferred the aesthetics of the front face and therefore decided to use it as side A, i.e., we integrated the sensor on the back bed, in contrast to the other two fabrics. We replaced course 2 entirely, since it consisted of an alternation of knit and miss on the back bed, as was required for the wire course. Courses 4 and 7 were again modified (4’ and 7’), to keep the sensor loop on needle III. The result showed again no irregularities on the A-side. Using this pattern, lower and upper traces were separated by one knit and two tuck stitches, which provided enough separation.

7.2.3 Three-colored: Milanolock.

Our third fabric is a custom creation that we termed "Milanolock" since it merges properties of Milano and Interlock. We knitted with three carriers (two ends of black PA, two ends of gray PA, and two ends of yellow PA) with stitch settings 38/28 for the PA and again 30/0 for the wire. Since the pattern consists solely of knit stitches, it shows balanced elasticity. For this pattern, we again inserted the wire on the front bed, adding an extra course that is only visible on the B-side. Three loops provided adequate insulation between connector trace stitches.
Figure 13:
Figure 13: Sides A (top) and B (bottom) of our resulting fabrics, with wire traces and positions of sensor cells clearly recognizable on the B-sides but concealed on the A-sides.
As can be seen in Figure 13, we successfully integrated our basic sensors in all three patterns without noticeable irregularities on the A-sides of the fabric. Especially in the tighter knit Swiss Weave-Knit, the structure barely even affected the B-side. We implemented a prototypical demo for visualization (cf. Figure 12c) that visualized sensor readings and showed that all four sensors of all three swatches provided outstanding responsiveness.

7.2.4 Washing.

Since interactive garments are an obvious use case, we performed washing tests using a Honeycomb sample, to assess the effect on sensor performance (i.e., on the coating’s durability), as well as wire resilience. Based on the method proposed by Rotzler et al. [42], we washed 10 times with 1 kg of additional fabric using an AEG ProTex L73471FL. We placed our fabric in a laundry net for protection, used 10 ml of neutral detergent and washed for 20 Minutes at 40°C with spin speed of 800 rpm. The fabric was subsequently laid out flat and air-dried for 6+ hours at 24°C and 45% relative humidity. After each washing cycle, we captured data for three iterations of straining it to 14 cm as well as 18 cm (i.e., strain e = 1/6 and e = 1/3) along course-direction and calculated mean and SD values.
Figure 14:
Figure 14: Washing tests showed a conductivity drop of \raise.17ex\(\scriptstyle \mathtt {\sim }\)50% after the first cycle. The values appear to settle around 25%.
Figure 14 shows that after the first washing cycle, conductivity drops by \raise.17ex\(\scriptstyle \mathtt {\sim }\)50% and further to about one third after the second cycle, before it appears to settle between 20% and 30%, meaning absolute values are affected, however relative measured range was still excellent, which for our strain of e = 1/6 we calculated as \raise.17ex\(\scriptstyle \mathtt {\sim }\)86% after the 10th cycle (as against 96% before washing). These results are reassuring since there seem to be only minor effects after the 3rd cycle, indicating long-term functionality. We did not experience any wire breaks; sensors were still intact after the 10th washing cycle.
Note that technical details like sensor range, saturation, and absolute conductivity, as well as the effects of washing are subject to the material ultimately used for implementation. As the concepts proposed in this work are not limited to a specific product, the base materials can be replaced by others with similar features. We therefore refrain from an in-depth technical evaluation of these metrics since they would not be representative for other materials and therefore be of limited value. Instead, we investigated utilization of loop meshing, i.e., the effect of stitch geometry, as presented in the following section.

8 Discussion

8.1 Limitations

Overall, we learned from our study that the choice of the substrate, i.e., the pattern repeat, has a huge impact on loop movement and therefore the yarn interaction in the micro-structure. This has great implications on the sensor’s properties, such as amplitude and range, but most importantly to its sensitivity with respect to actuation, which can be a valuable asset. However, the trend of force in Figures 6 and 8 shows there is structural hysteresis between force and strain, which may have implications for sensor performance. One way to mitigate this is to ply Spandex to the base yarn, which can be used to fine-tune elastic recoil and therefore reduces this structural hysteresis, as shown in [2]. For this paper, we deliberately did not optimize our samples using such specialized materials, as our objective was to provide baseline learnings that can be transferred to arbitrary scenarios easily, however, a logical next step would be to investigate into this direction. Furthermore, problematic effects of knitted sensors (offset and hysteresis, but most notably long-term drift) can be overcome using small-scale, computationally low-cost machine learning techniques, as demonstrated in [3].
We stated that our proposed low-scale coated wire based sensor cells are transferable not only to different structures but also different base materials, i.e., wires with similar features. Ultimately, the concept of harnessing the geometric mesh topology and utilizing it to engineer sensors of dedicated functionality as presented in Section 4 is not limited to wires with piezoresistive coating. Using yarn that is sheathed with resistive polymers10, or silver-coated wire with relatively high conductivity11 could be alternatives that are less challenging in textile processing due to their advantageous mechanical properties. This track is worth investigating, however, the result would be close to traditional knitted strain sensors [37] and would therefore lack the advantages of a solution that combines conductivity and piezoresistivity into one, as shown in Section 3.1.
Regarding the signal processing pipeline we implemented for our sensor fusion demos, which combined sensor signals by arithmetic operations, we see potential improvement in utilization of state-of-the-art Machine Learning techniques. This could greatly improve robustness and versatility, in particular we invision the possibility to infer elaborate geometric states (e.g., surface deformation and shapes) of fabrics that involve a high number or our sensors.

8.2 Lessons Learned for Handling Wires in Knitting Processes

Figure 15:
Figure 15: Closeups of issues that may arise when handling wires in a knitting machine: (a, b, c) If stitch numbers, wire tension, and carriage speed are not adjusted well, blank spots, kinks, and wire breaks can occur easily. (d) Kinks emerge frequently when the wire tension is not consistent during feeding. (e) When loops do not sit tight or wire is not fixated using adjacent tracing loops, the loops may lift from the intermeshing points, causing inconsistent sensor readings. (f, g) Yarn mechanics are sometimes unexpected and diagrams give false clues to actual yarn traversal in the final knit.

8.2.1 Preventing Wire Damage.

The most essential lesson learned is that, being nonelastic, the wire is highly sensitive to small variations in manufacturing parameters. As the knitting process is highly demanding to the material, sheared blank spots and wire kinks or even broken strands can easily emerge (cf. Figure 15 a,b,c). We managed to reduce those issues by knitting the wire courses at extra-low speed (0.02 m/s) and by tuning the stitch number to each fabric. The ideal setting for stitch number however depends largely on the substrate material and structure and has to be found individually.

8.2.2 Preventing Needle Damage.

Due to the wire’s inelasticity, needles can break easily if stitch numbers are chosen too low (i.e., loops are knitted too tight). Whenever a new pattern is first knitted, it is advisable to start with high stitch numbers (such as 40/30 for our 15 gauge machine) to fabricate first trials, and then tweak and converge to an optimal setting.

8.2.3 Enamel and Litz Construction.

We experimented with alternative enamel, most notably with a Polyurethane (TextileWire TW-A) which is easier to remove thermally and therefore simplifies tin-soldering of the copper core. Unfortunately, due to the inferior mechanical durability, it was sheared off more easily during fabrication and the results showed blank spots frequently (cf. Figure 15 a), risking shorts and highly inconsistent sensor readings. We also tried variations of litz wires with 4 and 8 ends and ø 50 µm, nominal (dtex 186), which were otherwise identical in terms of twist pitch and core material, however they turned out to break more frequently during the knitting procedure, regardless of the stitch settings.

8.2.4 Loop Tension for Firm Wire Loop Contacts.

Early on, we tried to reinforce wire loop contacts by using a different stitch number for the lower wire loop. While this seems reasonable and indeed resulted in tenser connections at the sensors, it caused frequent wire breaks: we observed that this was due to the carriage moving out for readjusting the cam, which causes jerky pull on the rather delicate wire.

8.2.5 Consistent Wire Tension.

Initially, we experienced frequent yarn breaks and identified highly problematic wire kinks that form during unwinding (cf. Figure 15 d). We noticed that unlike a regular yarn, the strands of a copper wire add up spin when they are pulled upwards off the cone. Tangential unwinding the wire could resolve this effect, it is however non-trivial to accomplish due to the cone’s inertia and the non-linear yarn demand in weft knitting. A feasible solution is to provide consistent wire tension in between feeding mechanism and yarn carrier, which can be accomplished with specialized active yarn feeding systems. We failed with a Shima Seiki i-DSCS+DTC active control yarn-feed system, since it could not handle the pull-back of the non-elastic wire in adequate speed when it went slack. As mentioned in Section 3.3, we managed to resolve most issues with a MSF 3 storage feeder from MEMMINGER-IRO with magnetic tensioner.

8.2.6 Multi-loop Sensor Cells for Redundancy.

Wire tension is crucial and has to be well adjusted, otherwise slack wire is pulled into the sensor loops, causing the loops losing contact randomly (cf. Figure 15 e), rendering the sensor readings inconsistent. For redundancy, it can be beneficial to connect multiple loop pairs side-by-side, instead of just a single pair like in this paper.

8.2.7 Optimization of Wire Insertion in the Knitting Process.

To provide enough residue wire for attaching crimps or other connectors, it is beneficial to not cut the wire immediately after it was last used, but to hold it on a few needles some centimeters off the fabric, knitting at least two more courses of substrate, and only then cut the wire and drop the held loops. Otherwise, the wire end may get pulled into the fabric and trying to get it out manually later without damaging the knit may be challenging.

8.2.8 Integration into Knitting Patterns.

In terms of integration into complex designs with the objective of making the modifications imperceptible, we noticed that required adaption of the pattern repeat is highly depending on the case, i.e., what yarn material, fabric structure, tightness settings are used. For example, we expected that when integrating an additional course of wire, we should leave out according loops on the base repeat, e.g., when inserting a wire trace with a half-gauge plain course ("knit-miss") into a plain course of PA (i.e., fully knit), we would for this row leave out the according needles for the PA (i.e., resulting in a "miss-knit") so the total number of loops adds up to the same number. This turned out to be not the case, since the wire’s interaction with the fabric is vastly different and at times, replacing PA courses by wire courses (cf. Honeycomb swatch) or even simply adding wire courses to the structure (cf. Swiss Weave-Knit) with no modifications whatsoever turned out to be least disruptive to the pattern’s look and feel. As a result, general rules can hardly be inferred and a pleasing solution has to be worked out by trial and error in conjunction with experience of a knitting engineer.

8.2.9 Non-obvious yarn traversal.

In terms of designing knitted geometry on loop-level, with the goal of engineering wire intersection, a key takeaway is that in the final mesh, the yarn will hardly traverse as anticipated based on diagrams such as Figure 16 c. This is due to yarn elasticity and complex structure meshing. It is crucial to keep this in mind, in particular when using more sophisticated structures, such as our example designs of Section 7.2, with numerous loop interactions. This may affect sensor functionality, as yarn may shift between wires in the final knit, unexpectedly (cf. Figure 15 f). Most notably, this led to our omitting of the basic miss variant for the Plain pattern, where due to fabric shrinkage, yarn is pulled between the wires, preventing the required contact. Moreover, the upper wire trace deflects on the back side of the fabric, due to its inelasticity (cf. Figure 15 g).

9 Conclusion and Future Work

We showed a novel method of fabricating low-scale knitted force sensors using a copper wire with special piezoresistive enamel. By harnessing knit mesh geometry, we are able to distinguish between different types of actuation and to infer continuous values representing wale- and course-directional strain, as well as orthogonal pressure. This ability to track multiple degrees of freedom on a fabric surface’s state makes them ideal for integration in fabric-based human interface devices, such as gloves, garments, seats, etc. Moreover, by merging all materials that are required for FSRs into one and therefore reducing complexity in knit structure and material composition, we are able to conceal our sensors in numerous appealing knitting patterns, since our solution requires only a single intersection of a pair of wires to produce functional force-sensing cells. Being fabricated on a computerized V-bed weft knitting machine, and in combination with our streamlined knit generation pipeline, our devices can be created quickly and ready-made, without requiring noteworthy manual postprocessing, finishing, or layering of multiple fabrics.
For next steps, we plan to closer investigate additional intersection types and approach the idea with an additional theoretical angle, considering loop-design based on rope mechanics and estimations based on computational physics simulation. Our goal is to provide a number of atomic and universal functional building-blocks, similar to those of [12] that could be easily placed as required. However, due to the versatility of V-bed knitting and the variety and individuality of knitted fabrics, we expect those will still have to be specially adapted for each scenario.
In terms of utilization, we plan to integrate our sensors into functional fully-fashioned 3D-knits, e.g., for specialized furnishing and garments, such as seats, gloves, etc. We will optimize the physical features of the substrate knits by complementing with additional material such as Spandex, Nylon, and melting yarn to achieve targeted improvement in elasticity and stability. For data processing, we plan to utilize ML techniques for more reliable tracking of the surfaces’ geometric state.

Acknowledgments

This research is part of the COMET project TextileUX (No. 865791, which is funded within the framework of COMET – Competence Centers for Excellent Technologies by BMVIT, BMDW, and the State of Upper Austria. The COMET program is handled by the FFG.

A Knitting Basics

In yarn-based textiles, one distinction is to be made in particular between weaves and knits that is relevant to this work: in weaves, the warp and weft strands are going straight and the resulting fabric is usually cut and sewn. Knitted textiles on the other hand consist of yarn that is forming loops and traverses throughout the structure (cf. Figure 16 a). These loops provide slack and therefore inherent stretchability, a property that renders knits the ideal textile candidates for sensing strain, however they can also be used for sensing contact pressure.
Weft knitting, and in particular V-bed knitting, provides many opportunities for thoroughly designing and engineering geometries and yarn interaction locally and on loop-level. This can be a powerful feature for sensor design, especially since in V-bed knitting, two knit faces can be combined arbitrarily and loops can be transferred between needle beds, both of which are features we extensively utilize in our work. Note that this also implies that most of the knits presented in this paper cannot be easily reproduced using hand-knitting or single-bed machines.
In the following, we explain knitting-related terms that are required to comprehend the concepts discussed in this paper. However, due to space limitations we confine this list to the essentials. For more details, we refer to literature covering the field of knitting technology comprehensively, such as [45].
Figure 16:
Figure 16: (a) Courses and wales in a knit are similar to rows and columns of a matrix; a knit is composed of one or more yarn ends that traverse through the fabric. (b) Anatomy of a knit loop with intermeshing points. (c) Illustration of basic stitch types used in weft knitting (c).
courses, wales:
simplified, "horizontal" and "vertical" dimensions of a knit (cf. Figure 16 a), similar to rows and columns to a matrix, or to weft and warp on a woven fabric.
intermeshing point:
contact points between loops along a wale (cf. Figure 16 b), where heads and feet of loops meet along a wale.
knit, purl, tuck, miss:
different stitch types (cf. Figure 16 c). To form a new loop, yarn is pulled through the loop that is currently held by the needle (from back to front to form a knit stitch, the other way around for a purl stitch, i.e., on a two-bed machine, a front bed needle creates knit stitches, while a back bed needle creates purl stitches). A tuck stitch adds the new yarn to a needle, without pulling it through the loop. Simply bypassing the yarn along a needle is called miss or float.
needle beds, racking, transfer, split:
a weft knitting machine may consist of multiple arrays of needles, called beds (cf. Figure 3 a), that can be racked, i.e., offset along each other. In combination with the possibility of transferring loops, i.e., handing loops over to the vis-á-vis needle of the opposite bed, this enables to fabricate intricate structures. A split stitch is combining this process with the formation of a new loop: new yarn is knitted at the respective needle, while the currently held loop is transferred over to the opposite bed.
carriage:
a component that moves left and right along the needle beds (cf. Figure 3 a), in order to extend and retract needles via a programmable cam system. Generally, a single carriage pass introduces a new course.
stitch number:
programmable value for controlling needle retracting distance, therefore controlling the resulting loop length. In this paper, we denote stitch number with two values S/YG, where S represents loop size for front and back stitches, and YG is the yarn guide output value.
yarn carrier:
component used to insert yarn into the needles’ hooks when they are extended by the traversing carriage (cf. Figure 3 a). Weft-knitting machines usually provide several, so complex multi-material knits can be created.
repeat:
sequence of stitches that specify a pattern, e.g., a sequence of 1 knit and 1 purl (which is referred to as a 1 × 1 rib) represents a 2-needle repeat. Frequently, these operations are repeated across the entire course. Furthermore, multi-course repeats specify sequences of consecutive courses that are again repeated over and over.

Footnotes

4
It is also possible to knit more complex trace layouts, e.g., by using techniques like short-rowing.

Supplemental Material

MP4 File - Video Preview
Video Preview
Transcript for: Video Preview
MP4 File - Video Presentation
Video Presentation
Transcript for: Video Presentation
MP4 File - Video Figure
Video Figure
Transcript for: Video Figure
ZIP File - Supplemental Material
This archive contains Knitout and Shima Seiki DAT files used for the evaluation and demo applications, as well as an extensive description of the signal processing pipeline used in our sensor fusion demos.

References

[1]
Roland Aigner, Mira Alida Haberfellner, and Michael Haller. 2022. spaceR: Knitting Ready-Made, Tactile, and Highly Responsive Spacer-Fabric Force Sensors for Continuous Input. In Proceedings of the 35th Annual ACM Symposium on User Interface Software and Technology (Bend, OR, USA) (UIST ’22). Association for Computing Machinery, New York, NY, USA, Article 68, 15 pages. https://doi.org/10.1145/3526113.3545694
[2]
Roland Aigner and Frank Hepper. 2024. An Evaluation of Multi-Component Weft-Knitted Twill Structures for Sensing Tensile Force. IEEE Sensors Journal (2024). https://doi.org/10.1109/JSEN.2024.3353821
[3]
Roland Aigner and Andreas Stöckl. 2023. Machine Learning Based Compensation for Inconsistencies in Knitted Force Sensors. IEEE Sensors Journal (2023). https://doi.org/10.1109/JSEN.2023.3346686
[4]
Ozgur Atalay, William Richard Kennon, and Erhan Demirok. 2015. Weft-Knitted Strain Sensor for Monitoring Respiratory Rate and Its Electro-Mechanical Modeling. IEEE Sensors Journal 15, 1 (2015), 110–122. https://doi.org/10.1109/JSEN.2014.2339739
[5]
Emmanuel Ayodele, Tianzhe Bao, Syed Ali Raza Zaidi, Ali M. A. Hayajneh, Jane Scott, Zhi-Qiang Zhang, and Des McLernon. 2021. Grasp Classification With Weft Knit Data Glove Using a Convolutional Neural Network. IEEE Sensors Journal 21, 9 (2021), 10824–10833. https://doi.org/10.1109/JSEN.2021.3059028
[6]
Natalija Baribina, Aleksander Oks, I. Baltina, and Peteris Eizentals. 2018. Comparative analysis of knitted pressure sensors. https://doi.org/10.22616/ERDev2018.17.N498
[7]
Hyung Choi, Dong-Wook Shin, Jiajie Yang, Sanghyo Lee, Cátia Figueiredo, Stefano Sinopoli, Kay Ullrich, P. Jovančić, Alessio Marrani, Roberto Momentè, João Gomes, Rita Branquinho, Umberto Emanuele, Hanleem Lee, Sang Bang, Sung-Min Jung, Soo Deok Han, Shijie Zhan, William Harden-Chaters, and Jong Kim. 2022. Smart textile lighting/display system with multifunctional fibre devices for large scale smart home and IoT applications. Nature Communications 13 (02 2022). https://doi.org/10.1038/s41467-022-28459-6
[8]
D De Rossi, A Della Santa, and A Mazzoldi. 1999. Dressware: wearable hardware. Materials Science and Engineering: C 7, 1 (1999), 31–35. https://doi.org/10.1016/S0928-4931(98)00069-1
[9]
TU Delft. [n. d.]. Textile pressure sensor for car seats. https://www.tudelft.nl/en/ide/research/research-labs/emerging-materials-lab/smart-textiles/textile-pressure-sensor-for-car-seats
[10]
Laura Devendorf, Katya Arquilla, Sandra Wirtanen, Allison Anderson, and Steven Frost. 2020. Craftspeople as Technical Collaborators: Lessons Learned through an Experimental Weaving Residency. In Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems (Honolulu, HI, USA) (CHI ’20). Association for Computing Machinery, New York, NY, USA, 1–13. https://doi.org/10.1145/3313831.3376820
[11]
Laura Devendorf, Sasha de Koninck, and Etta Sandry. 2022. An Introduction to Weave Structure for HCI: A How-to and Reflection on Modes of Exchange. In Proceedings of the 2022 ACM Designing Interactive Systems Conference (Virtual Event, Australia) (DIS ’22). Association for Computing Machinery, New York, NY, USA, 629–642. https://doi.org/10.1145/3532106.3534567
[12]
Jun Gong, Olivia Seow, Cedric Honnet, Jack Forman, and Stefanie Mueller. 2021. MetaSense: Integrating Sensing Capabilities into Mechanical Metamaterial. In The 34th Annual ACM Symposium on User Interface Software and Technology (Virtual Event, USA) (UIST ’21). Association for Computing Machinery, New York, NY, USA, 1063–1073. https://doi.org/10.1145/3472749.3474806
[13]
Ragnar Holm. 1967. "Electric Contacts: Theory and Application" (4 ed.). "Springer Berlin", "Heidelberg". https://doi.org/10.1007/978-3-662-06688-1
[14]
Cedric Honnet, Hannah Perner-Wilson, Marc Teyssier, Bruno Fruchard, Jürgen Steimle, Ana C. Baptista, and Paul Strohmeier. 2020. PolySense: Augmenting Textiles with Electrical Functionality Using In-Situ Polymerization. In Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems (Honolulu, HI, USA) (CHI ’20). Association for Computing Machinery, New York, NY, USA, 1–13. https://doi.org/10.1145/3313831.3376841
[15]
Alexandra Ion, David Lindlbauer, Philipp Herholz, Marc Alexa, and Patrick Baudisch. 2019. Understanding Metamaterial Mechanisms. In Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems (Glasgow, Scotland Uk) (CHI ’19). Association for Computing Machinery, New York, NY, USA, 1–14. https://doi.org/10.1145/3290605.3300877
[16]
Alexandra Ion, Ludwig Wall, Robert Kovacs, and Patrick Baudisch. 2017. Digital Mechanical Metamaterials. In Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems (Denver, Colorado, USA) (CHI ’17). Association for Computing Machinery, New York, NY, USA, 977–988. https://doi.org/10.1145/3025453.3025624
[17]
Karlos Ishac and Kenji Suzuki. 2018. LifeChair: A Conductive Fabric Sensor-Based Smart Cushion for Actively Shaping Sitting Posture. Sensors 18, 7 (2018). https://doi.org/10.3390/s18072261
[18]
Jarrod Knibbe, Rachel Freire, Marion Koelle, and Paul Strohmeier. 2021. Skill-Sleeves: Designing Electrode Garments for Wearability. In Proceedings of the Fifteenth International Conference on Tangible, Embedded, and Embodied Interaction (Salzburg, Austria) (TEI ’21). Association for Computing Machinery, New York, NY, USA, Article 33, 16 pages. https://doi.org/10.1145/3430524.3440652
[19]
MIT SELF-ASSEMBLY LAB. 2019. 3D Knit BioSuit. https://selfassemblylab.mit.edu/3d-knit-biosuit
[20]
Joanne Leong, Patrick Parzer, Florian Perteneder, Teo Babic, Christian Rendl, Anita Vogl, Hubert Egger, Alex Olwal, and Michael Haller. 2016. ProCover: Sensory Augmentation of Prosthetic Limbs Using Smart Textile Covers. In Proceedings of the 29th Annual Symposium on User Interface Software and Technology (Tokyo, Japan) (UIST ’16). Association for Computing Machinery, New York, NY, USA, 335–346. https://doi.org/10.1145/2984511.2984572
[21]
Yutian Li, Xuhong Miao, Jonathan Y. Chen, Gaoming Jiang, and Qing Liu. 2021. Sensing performance of knitted strain sensor on two-dimensional and three-dimensional surfaces. Materials & Design 197 (2021), 109273. https://doi.org/10.1016/j.matdes.2020.109273
[22]
Xinhua Liang, Honglian Cong, Zhijia Dong, and Gaoming Jiang. 2022. Size Prediction and Electrical Performance of Knitted Strain Sensors. Polymers 14, 12 (2022). https://doi.org/10.3390/polym14122354
[23]
Xiaoyou Lin and Boon-Chong Seet. 2015. A Linear Wide-Range Textile Pressure Sensor Integrally Embedded in Regular Fabric. IEEE Sensors Journal 15, 10 (2015), 5384–5385. https://doi.org/10.1109/JSEN.2015.2453214
[24]
Yiyue Luo, Yunzhu Li, Pratyusha Sharma, Wan Shou, Kui Wu, Michael Foshey, Beichen Li, Tomás Palacios, Antonio Torralba, and Wojciech Matusik. 2021. Learning human–environment interactions using conformal tactile textiles. Nature Electronics 4, 3 (March 2021), 193–201. https://doi.org/10.1038/s41928-021-00558-0
[25]
Yiyue Luo, Kui Wu, Tomás Palacios, and Wojciech Matusik. 2021. KnitUI: Fabricating Interactive and Sensing Textiles with Machine Knitting. In Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems (Yokohama, Japan) (CHI ’21). Association for Computing Machinery, New York, NY, USA, Article 668, 12 pages. https://doi.org/10.1145/3411764.3445780
[26]
James McCann, Lea Albaugh, Vidya Narayanan, April Grow, Wojciech Matusik, Jennifer Mankoff, and Jessica Hodgins. 2016. A Compiler for 3D Machine Knitting. ACM Trans. Graph. 35, 4, Article 49 (July 2016), 11 pages. https://doi.org/10.1145/2897824.2925940
[27]
Denisa Qori McDonald, Shruti Mahajan, Richard Vallett, Genevieve Dion, Ali Shokoufandeh, and Erin Solovey. 2022. Interaction with Touch-Sensitive Knitted Fabrics: User Perceptions and Everyday Use Experiments. In Proceedings of the 2022 CHI Conference on Human Factors in Computing Systems (New Orleans, LA, USA) (CHI ’22). Association for Computing Machinery, New York, NY, USA, Article 426, 20 pages. https://doi.org/10.1145/3491102.3502077
[28]
Denisa Qori McDonald, Richard Vallett, Erin Solovey, Geneviève Dion, and Ali Shokoufandeh. 2020. Knitted Sensors: Designs and Novel Approaches for Real-Time, Real-World Sensing. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 4, 4, Article 145 (dec 2020), 25 pages. https://doi.org/10.1145/3432201
[29]
Sara Mlakar, Mira Alida Haberfellner, Hans-Christian Jetter, and Michael Haller. 2021. Exploring Affordances of Surface Gestures on Textile User Interfaces. In Proceedings of the 2021 ACM Designing Interactive Systems Conference (Virtual Event, USA) (DIS ’21). Association for Computing Machinery, New York, NY, USA, 1159–1170. https://doi.org/10.1145/3461778.3462139
[30]
Sara Mlakar and Michael Haller. 2020. Design Investigation of Embroidered Interactive Elements on Non-Wearable Textile Interfaces. In Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems (Honolulu, HI, USA) (CHI ’20). Association for Computing Machinery, New York, NY, USA, 1–10. https://doi.org/10.1145/3313831.3376692
[31]
Bodhibrata Mukhopadhyay, Osheen Sharma, and Subrat Kar. 2018. IoT Based Wearable Knitted Fabric Respiratory Monitoring System. In 2018 IEEE SENSORS. 1–4. https://doi.org/10.1109/ICSENS.2018.8589540
[32]
Vidya Narayanan, Lea Albaugh, Jessica Hodgins, Stelian Coros, and James Mccann. 2018. Automatic Machine Knitting of 3D Meshes. ACM Trans. Graph. 37, 3, Article 35 (aug 2018), 15 pages. https://doi.org/10.1145/3186265
[33]
Alex Olwal, Jon Moeller, Greg Priest-Dorman, Thad Starner, and Ben Carroll. 2018. I/O Braid: Scalable Touch-Sensitive Lighted Cords Using Spiraling, Repeating Sensing Textiles and Fiber Optics. In Proceedings of the 31st Annual ACM Symposium on User Interface Software and Technology (Berlin, Germany) (UIST ’18). Association for Computing Machinery, New York, NY, USA, 485–497. https://doi.org/10.1145/3242587.3242638
[34]
R. Paradiso, G. Loriga, and N. Taccini. 2005. A wearable health care system based on knitted integrated sensors. IEEE Transactions on Information Technology in Biomedicine 9, 3 (2005), 337–344. https://doi.org/10.1109/TITB.2005.854512
[35]
Patrick Parzer, Florian Perteneder, Kathrin Probst, Christian Rendl, Joanne Leong, Sarah Schuetz, Anita Vogl, Reinhard Schwoediauer, Martin Kaltenbrunner, Siegfried Bauer, and Michael Haller. 2018. RESi: A Highly Flexible, Pressure-Sensitive, Imperceptible Textile Interface Based on Resistive Yarns. In Proceedings of the 31st Annual ACM Symposium on User Interface Software and Technology (Berlin, Germany) (UIST ’18). Association for Computing Machinery, New York, NY, USA, 745–756. https://doi.org/10.1145/3242587.3242664
[36]
Syamantak Payra, Irmandy Wicaksono, Juliana Cherston, Cedric Honnet, Valentina Sumini, and Joseph A. Paradiso. 2021. Feeling Through Spacesuits: Application of Space-Resilient E-Textiles to Enable Haptic Feedback on Pressurized Extravehicular Suits. In 2021 IEEE Aerospace Conference (50100). 1–12. https://doi.org/10.1109/AERO50100.2021.9438515
[37]
Andreas Pointner, Thomas Preindl, Sara Mlakar, Roland Aigner, Mira Alida Haberfellner, and Michael Haller. 2022. Knitted Force Sensors. In Adjunct Proceedings of the 35th Annual ACM Symposium on User Interface Software and Technology (Bend, OR, USA) (UIST ’22 Adjunct). Association for Computing Machinery, New York, NY, USA, Article 77, 3 pages. https://doi.org/10.1145/3526114.3558656
[38]
Andreas Pointner, Thomas Preindl, Sara Mlakar, Roland Aigner, and Michael Haller. 2020. Knitted RESi: A Highly Flexible, Force-Sensitive Knitted Textile Based on Resistive Yarns. In ACM SIGGRAPH 2020 Emerging Technologies (Virtual Event, USA) (SIGGRAPH ’20). Association for Computing Machinery, New York, NY, USA, Article 21, 2 pages. https://doi.org/10.1145/3388534.3407292
[39]
Irene Posch. 2019. The Knitted Radio. https://ireneposch.net/the-knitted-radio/
[40]
Ernest Rehmi Post, Margaret Orth, Peter Russo, and Neil Gershenfeld. 2000. E-broidery: Design and fabrication of textile-based computing. IBM Systems Journal 39, 3.4 (2000), 840–860. https://doi.org/10.1147/sj.393.0840
[41]
Ivan Poupyrev, Nan-Wei Gong, Shiho Fukuhara, Mustafa Emre Karagozler, Carsten Schwesig, and Karen E. Robinson. 2016. Project Jacquard: Interactive Digital Textiles at Scale. In Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems (San Jose, California, USA) (CHI ’16). Association for Computing Machinery, New York, NY, USA, 4216–4227. https://doi.org/10.1145/2858036.2858176
[42]
Sigrid Rotzler, Christine Kallmayer, Christian Dils, Malte von Krshiwoblozki, Ulrich Bauer, and Martin Schneider-Ramelow. 2020. Improving the washability of smart textiles: influence of different washing conditions on textile integrated conductor tracks. The Journal of The Textile Institute 111, 12 (2020), 1766–1777. https://doi.org/10.1080/00405000.2020.1729056 arXiv:https://doi.org/10.1080/00405000.2020.1729056
[43]
Teddy Seyed, James Devine, Joe Finney, Michal Moskal, Peli de Halleux, Steve Hodges, Thomas Ball, and Asta Roseway. 2021. Rethinking the Runway: Using Avant-Garde Fashion To Design a System for Wearables. In Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems (Yokohama, Japan) (CHI ’21). Association for Computing Machinery, New York, NY, USA, Article 45, 15 pages. https://doi.org/10.1145/3411764.3445643
[44]
Zilin Si, Tianhong Catherine Yu, Katrene Morozov, James McCann, and Wenzhen Yuan. 2023. RobotSweater: Scalable, Generalizable, and Customizable Machine-Knitted Tactile Skins for Robots. arXiv preprint arXiv:2303.02858 (2023).
[45]
David J. Spencer. 2001. Knitting Technology: A Comprehensive Handbook and Practical Guide (3 ed.). Woodhead Publishing.
[46]
Paul Strohmeier, Jarrod Knibbe, Sebastian Boring, and Kasper Hornbæk. 2018. zPatch. In Proceedings of the Twelfth International Conference on Tangible, Embedded, and Embodied Interaction (TEI ’18). ACM Press, New York, NY, USA, 188–198. https://doi.org/10.1145/3173225.3173242
[47]
Richard Vallett, Denisa Qori McDonald, Genevieve Dion, Youngmoo Kim, and Ali Shokoufandeh. 2020. Toward Accurate Sensing with Knitted Fabric: Applications and Technical Considerations. Proc. ACM Hum.-Comput. Interact. 4, EICS, Article 79 (jun 2020), 26 pages. https://doi.org/10.1145/3394981
[48]
K. Weiss and H. Worn. 2005. The working principle of resistive tactile sensor cells. In IEEE International Conference Mechatronics and Automation, 2005, Vol. 1. 471–476 Vol. 1. https://doi.org/10.1109/ICMA.2005.1626593
[49]
Irmandy Wicaksono, Don Derek Haddad, and Joseph Paradiso. 2022. Tapis Magique: Machine-Knitted Electronic Textile Carpet for Interactive Choreomusical Performance and Immersive Environments. In Proceedings of the 14th Conference on Creativity and Cognition (Venice, Italy) (C&C ’22). Association for Computing Machinery, New York, NY, USA, 262–274. https://doi.org/10.1145/3527927.3531451
[50]
Irmandy Wicaksono, Peter G. Hwang, Samir Droubi, Franny Xi Wu, Allison N. Serio, Wei Yan, and Joseph A. Paradiso. 2022. 3DKnITS: Three-dimensional Digital Knitting of Intelligent Textile Sensor for Activity Recognition and Biomechanical Monitoring. In 2022 44th Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC). 2403–2409. https://doi.org/10.1109/EMBC48229.2022.9871651
[51]
Irmandy Wicaksono, Carson I. Tucker, Tao Sun, Cesar A. Guerrero, Clare Liu, Wesley M. Woo, Eric J. Pence, and Canan Dagdeviren. 2020. A tailored, electronic textile conformable suit for large-scale spatiotemporal physiological sensing in vivo. npj Flexible Electronics (2020), 2397–4621. https://doi.org/10.1038/s41528-020-0068-y
[52]
Tianhong Catherine Yu, Riku Arakawa, James McCann, and Mayank Goel. 2023. UKnit: A Position-Aware Reconfigurable Machine-Knitted Wearable for Gestural Interaction and Passive Sensing Using Electrical Impedance Tomography. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems (Hamburg, Germany) (CHI ’23). Association for Computing Machinery, New York, NY, USA, Article 628, 17 pages. https://doi.org/10.1145/3544548.3580692

Cited By

View all
  • (2024)What's in a cable? Abstracting Knitting Design Elements with Blended Raster/Vector PrimitivesProceedings of the 37th Annual ACM Symposium on User Interface Software and Technology10.1145/3654777.3676351(1-20)Online publication date: 13-Oct-2024

Index Terms

  1. Loopsense: low-scale, unobtrusive, and minimally invasive knitted force sensors for multi-modal input, enabled by selective loop-meshing

      Recommendations

      Comments

      Please enable JavaScript to view thecomments powered by Disqus.

      Information & Contributors

      Information

      Published In

      cover image ACM Conferences
      CHI '24: Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems
      May 2024
      18961 pages
      ISBN:9798400703300
      DOI:10.1145/3613904
      This work is licensed under a Creative Commons Attribution-NoDerivatives International 4.0 License.

      Sponsors

      Publisher

      Association for Computing Machinery

      New York, NY, United States

      Publication History

      Published: 11 May 2024

      Check for updates

      Badges

      Author Tags

      1. e-textiles
      2. fabrication
      3. force sensor
      4. knitting
      5. resistive sensor
      6. smart textiles
      7. textile interface

      Qualifiers

      • Research-article
      • Research
      • Refereed limited

      Conference

      CHI '24

      Acceptance Rates

      Overall Acceptance Rate 6,199 of 26,314 submissions, 24%

      Upcoming Conference

      CHI 2025
      ACM CHI Conference on Human Factors in Computing Systems
      April 26 - May 1, 2025
      Yokohama , Japan

      Contributors

      Other Metrics

      Bibliometrics & Citations

      Bibliometrics

      Article Metrics

      • Downloads (Last 12 months)1,420
      • Downloads (Last 6 weeks)202
      Reflects downloads up to 08 Dec 2024

      Other Metrics

      Citations

      Cited By

      View all
      • (2024)What's in a cable? Abstracting Knitting Design Elements with Blended Raster/Vector PrimitivesProceedings of the 37th Annual ACM Symposium on User Interface Software and Technology10.1145/3654777.3676351(1-20)Online publication date: 13-Oct-2024

      View Options

      View options

      PDF

      View or Download as a PDF file.

      PDF

      eReader

      View online with eReader.

      eReader

      HTML Format

      View this article in HTML Format.

      HTML Format

      Login options

      Media

      Figures

      Other

      Tables

      Share

      Share

      Share this Publication link

      Share on social media