Journal of Fashion Technology & Textile EngineeringISSN: 2329-9568

All submissions of the EM system will be redirected to Online Manuscript Submission System. Authors are requested to submit articles directly to Online Manuscript Submission System of respective journal.

Review Article, J Fashion Technol Textile Eng Vol: 3 Issue: 3

Piezoresistive Sensors Fabricated with Conductive Textiles for Monitoring the Step Rate with Read-Out Electronics and Wireless Connection to a Smart Watch

Lorenzo Capineri*
Department of Information Engineering, University of Florence, Via S. Marta 3, 50139, Firenze, Italy
Corresponding author : Lorenzo Capineri
Ultrasound and Non Destructive Testing Laboratory, Department of Information Engineering, University of Florence, Via S. Marta 3, 50139, Firenze, Italy
Tel: +39 055 2758627; Fax: +39 055 2758570
E-mail: lorenzo.capineri@unifi.it
Received: March 24, 2015 Accepted: June 08, 2015 Published: June 11, 2015
Citation: Capineri L (2015) Piezoresistive Sensors Fabricated with Conductive Textiles for Monitoring the Step Rate with Read-Out Electronics and Wireless Connection to a Smart Watch. J Fashion Technol Textile Eng 3:3. doi:10.4172/2329-9568.1000124

Abstract

Piezoresistive Sensors Fabricated With Conductive Textiles For Monitoring The Step Rate With Read-Out Electronics And Wireless Connection To A Smart Watch

Soft sensors made with conductive smart textiles have been developed for connecting with the environment, increasing operator safety, entertainment and physiological parameter monitoring. This paper describes the fabrication processes for smart textiles with conductive fibers made with different types of metal filaments and explains the design choice for pressure measurements with soft sensors. This new type of materials is investigated for the design of a removable insole sensor for step rate monitoring. The main components integrated into the wearable device are a piezoresistive sensor, a low power analog front-end and a wireless connection with a smart watch device. The dynamic behavior of the piezoresistive is characterized and the output voltage waveform is processed to estimate the step rate per second. The advantage of a customize graphical user interface on the display of the smart watch allowed easy monitoring of the step rate from theaverage pressure variation on the sensorized insole during walking and running. The paper demonstrates that this sensor technology can be used to design different types of piezoresistive sensors with multiple elements to record dynamically the spatial distribution of foot pressure.

Keywords: Smart textiles; Piezoresistive sensors; Wearable technologies; Smart watch; Training shoe; Analog front end; Foot pressure; Sensorized insole

Keywords

Smart textiles; Piezoresistive sensors; Wearable technologies; Smart watch; Training shoe; Analog front end; Foot pressure; Sensorized insole

Fabrication Technologies of Conductive Textiles

For smart soft sensor design, conductive fibers are the key element to build smart fabrics with defined electrical properties (resistance, capacitance etc.) [1,2]. The current flow in fabrics depends on the conductive material used percentage of conductive fibers, the fabric structure, and the conductive fiber contact surface. Fabrication technologies are described that use metal fibers only, and also those including a mixture with textile fibers. These yarns are produced using textile production technologies. Advanced processes of metallization of polyamide fibers with silver coating are also developed because polyamide gives the yarn strength and elasticity, while thin compliant silver coating guarantees electrical conductivity and biocompatibility for skin electrodes [3]. Carbon fibers can also be used to produce smart fabrics, form the fact that carbon is a conductive material [4]. They are about 0.005-0.010 mm in diameter, produced from a precursor polymer; the precursor is first spun into filaments and after spinning, the polymer fibers are then heated to drive off non-carbon atoms (carbonization). Thanks to carbon’s electrical properties, a very low temperature coefficient of resistivity -0.0005 [1/°C] is achieved. Electrons flow along different directions in the fabric depending on the thread: In woven fabrics, the current flows in the orthogonal directions of the filaments with almost same resistance while in knitted fabrics the resistance offered in the two orthogonal directions is different (Figure 1). The functionality of these conductive fabrics is exploited to build new resistive soft sensors [5,6]. Wearable soft sensors can be produced by adapting knitting machines to include the conductive filaments. There are two fabrication technologies for conductive knitting fabrics: weft and warp knitting. The weft knitting is the most versatile and promising technology for three-dimensional textiles and textile sensors. In Figure 2 is shown a machine for weft knitting with a detail of the pattern created with metal yarns. One important issue for the designers and integrators of sensors with conductive smart textiles is the electrical conductivity characterization and the results of the investigation on a type of conductive textile are described in the next section.
Figure 1: Current density and directions: (a) in a woven fabric, (b) knitted fabric.
Figure 2: Weft knitting technology for conductive textiles: (a) machine, (b) knitted conductive fabric with metal filament patterns.

Resistive Conductive Textiles for Soft Sensors

Conductivity is the main parameter that needs to be well characterized in a resistive conductive textile. The conductivity obtained in different ways (metal filaments in conductive fabrics or with screen printed methods) can be measured in laboratory and often the manufacturers apply their own protocol. The increasing interest for commercialization of these types of textiles has stimulated the formation of a working group for the definition of guidelines for measuring the resistive behavior of a textile [7]. A work in progress for these guidelines is focused on the definition of the electrodes configuration that is in contact with the same side of textile under test, the textile preconditioning and, for non-homogeneous fabrics, even from the orientation of the specimen. In the conductive textile sample with dimension 60 mm×60 mm the linear resistance measurement along the two orthogonal directions shown in Figure 1 provide 1.8 Ω and 1.3M Ω.
Among several applications of conductive textile for soft sensors (mechanical pressure, strain, position with a soft circuit potentiometer, temperature) the paper is focused on piezoresistive sensors. For electronics applications, soft piezoresistive sensors can be used as an electrical switch element that can also be arranged as a matrix of switches. Depending on the application the electronic front-end can be designed to obtain a digital switch (voltage amplitude variation between two assigned thresholds detected with a comparator) or analog switch by processing the digital sampling of the dynamic voltage signals. These types of soft switches used with digital output need de-bouncing analog circuits with hysteretic comparator like Schmitt trigger or de-bouncing routines when are interfaced directly to microcontrollers such as Arduino [8,9]. The analog solution of the de-bouncing circuit has the advantages to be low cost, little size and low power consumption respect to the microcontroller solution that is more flexible to be adapted and reprogrammed for different kind of sensors.
Piezoresistive sensors made with pressure sensitive fabrics have also other advantages when used as soft switches (Figure 3a), no need of further production steps, low cost, transpiring, semi-transparent, flexible, different activating pressures, matrix switches, large area (up to 50 cm×50 cm) switches (Figure 3b), skin compatible conductive materials. The piezoresistive materials (carbon or polypyrrole) are also used for analog pressure sensors and they are attractive respect to piezoelectric materials like PVDF fibers inserted between two conductive textiles [10,11] or fabric with piezoelectric fibers [12].
Figure 3: (a) Pressure sensitive conductive fabrics used as electronic switch. The colour of the fabric can be changed to be used in different fashion clothes or professional garments. (b) Piezoresistive matrix of 8x8 sensors with connectors for column-row readout electronics [9].
Two conductive fabrics layer and an inexpensive pressure sensitive layer placed in between build the proposed piezoresistive sensor. The intermediate layer is made of Velostat™; that is a packaging material made of a polymeric foil impregnated with carbon black to make it electrically conductive and its conductivity varies according to the applied mechanical pressure (Figure 4).
Figure 4: (a) zoom of the weft knitted fabric with a steel filament. (b) Assembly of the piezoresistive sensor with two conductive sides of the textile with orthogonal orientation (see arrows).
Details of the metal pattern obtained via weft knitting, with a stainless steel filament is also shown in Figure 4. The two conductive sides are made with stainless steel knitted fabrics and assembled with orthogonal orientation to obtain an improved homogeneity in the resistance variation. As explained in section, Fabrication technologies of conductive textiles, because of the anisotropic characteristic of the current flow, we found the maximum piezoresistive sensitivity when the conductive patterns on the external fabrics are orthogonal (Figure 5a). Thanks to the versatility of the weft knitting and the described assembly process, matrix sensors can be built for measuring the spatial distribution of pressure and interfaced to a multichannel read out electronics. Array of piezoresistive sensors fabricated with nanotechnology and screen printing technology have been already proposed in the literature [12]. In this respect our system is provided of an electronic interface and the max step rate has been characterized experimentally and by simulation of the electronic circuit. A review of different sensor technologies for monitoring the foot plantar pressure has been published [13,14] and silicon based piezoresistive sensors are mentioned, in the later published work it is also emphasized the importance of wireless sensor systems that allow an easy use of the information derived from this type of sensors; following this trend we presented a simple/low power consumption piezoresistive sensors interface for step rate monitoring on a smart watch.
Figure 5: (a) Structure of a pressure sensitive fabric made of two outer layers with orthogonal conductive patterns and a Velostat (3MTM) intermediate layer. (b) Two piezoresistive sensors based on knitted fabric process with different metal wires: steel (diameter 0.12 mm) and tinned copper (diameter 0.07 mm). The sample on the left shows the connection to the readout electronics are made by two flexible PCB stripes obtained with copper metalized Kapton soldered on the conductive fabrics. This prototype is stitched on the edges to provide a robust packaging.

Piezoresistive Sensor Design and Characterization

The application of piezoresistive sensors for the sensorized removable insole has the mechanical requirement to be flexible, comfortable, and easily removable from the shoe, but also a good sensitivity of resistance variation during walking and running is needed. First, the comparison of two types of piezoresistive sensors fabricated with conductive layers made by copper wire and the other with steel wire was investigated. The weft knitted fabric made with tinned copper wire of 75 μm diameter is very flexible and easy for soldering to connecting cables while steel wire of 120 μm diameter is mechanically more robust but can wear out the intermediate layer of piezoresistive material due to mechanical friction.
Two samples have been fabricated with dimensions of 40 mm × 60 mm in order to be inserted in a training shoe on the top side of the existing insole. The resistance (R) variation has been characterized in the laboratory when the piezoresistive sensor is not subjected to any pressure (P=0) and when subjected to a uniform pressure P=4583 N/m2. The results obtained for the two piezoresistive (Figure 5b) sensors are summarized in Table 1. The advantage of a greater resistance of the steel wire is the lower power consumption when the piezoresistive sensors voltage is measured by a voltage divider with a series resistor and low voltage power supply such as 3.3V. The copper wire sample has demonstrated also the advantage of a better isotropic current flow being built with a thinner diameter wire than the steel one. However the lower resistance in the rest condition (P=0) requires an electronic interface with switched mode power supply and synchronous measurement of the output voltage in order to limit the power consumption. For these reasons the prototype system has been built with the piezoresistive sensor made with steel wire. The final prototype was made with a piezoresistive soft sensor with area 40 mm × 90 mm connected to a 3.3V battery pack with a series resistor. The dynamic voltage for different values of the series resistor is shown in Figure 6a. It can be observed that the maximum output voltage is 2.61V for a 3.9 kΩ resistor and this is a good trade off also for current consumption that is about 0.8 mA considering a minimum value of R=100 Ω. However, this value is still too high for a continuous operation with button type batteries. The solution to this problem comes from the electronic front end that can activate the power supply only for the time necessary to measure the resistance value and so the average current flowing in the piezoresistive sensor becomes proportional to the duty cycle. The power consumption is an important aspect of the electronic design and will be explained in more detail in the next section. Once the sample has been inserted firmly in the shoe the resistance variation has been measured with an adult jumping on the spot and the range found from about R=500 Ω to R=38 kΩ. The output dynamic voltage from the resistor divider is shown in Figure 6 and a quite regular pulsed waveform can be observed. However, some signal processing is needed to extract the basic information that is the step rate. A challenging problem for the sensors fabricated with smart textiles are the connections between the electronics and the sensors and at present a viable solution are thin flexible circuit based on Kapton® polyimide film. An example of this type of connections is shown in Figure 5, with two flat connections soldered on the knitted wire mesh. The choice of making the connections with flexible Kapton® provides also a possible solution for sealing the electrical paths from harsh environmental conditions (moisture, dirt, mud etc) as the shoes are normally used. The contacts between the insole and the flat Kapton® stripes are made by soldering.
Figure 6: (a) Dynamic voltage output for different series resistor value for piezoresistive sensor supplied with a 3.3V battery (b) Dynamic voltage output with the piezoresistive sensors inserted in a training shoe while an adult is jumping on the spot.
Table 1: Resistance variation with Pressure.

Wearable Read-Out Electronics and Wireless Connection to a Smart-Watch

The block scheme of the complete system is shown in Figure 7a and the picture of the prototype in Figure 7b, The system consists of a wearable piezoresistive sensor, read-out electronics and a wireless connection with a smart watch. On the smart watch the information of the step rate is displayed with a customized graphical user interface (GUI). This device can be used as an odometer, to count steps or strides, but the smart insole can also measure foot contact and lift durations, pressures, spent energy, etc. A shaping analog circuit generates pulses through a voltage comparator and then is processed by a low power microcontroller (Transceiver CC1110 by Texas Instruments).
Figure 7: (a) Block scheme of the read-out electronics for the smart insole with piezoresistive sensors, (b) prototype device.
Read-out electronics
The first stage of the analog read-out electronics is a low power voltage comparator circuit design with a Schmitt trigger with fixed hysteresis that converts the dynamic voltage response generated by the sensorized insole. A LMV7291 single low power comparator is chosen because it provides high precision with very low supply current consumption. The signal from the sensor during a normal walk or run is a quite periodic pulsed waveform (Figure 6), and the comparator converts it into a square wave that can be processed as a digital input by the CC1110 microcontroller. This read-out electronics is entirely powered by two 1.5 Volt AAA batteries necessary for power supply of the CC1110 mini development kit. The latter sends the information to a programmable watch to visualize if the step rate is within the expected range. In our design the selected smart watch is the Texas Instruments EZ430-Chronos because has an open platform for GUI design and very low power consumption. The complete read-out electronics is designed with minimum number of components in order to be mounted on the same Kapton® PCB flexible circuit used for the piezoresistive sensor connections. The drawing in Figure 8 describes the concept of the integration of the smart piezoresistive sensors in a training shoe.
Figure 8: Pictorial description of the removable sensorized insole. 1) Removable insole with piezoresistive soft sensor. 2) Two flexible conductive stripes with metalized Kapton film. 3) Battery operated read-out electronics with transceiver. Sealed packaging fixed to the shoe with removable Velcro fabric hook and loop fastener.
Firmware and software development
The main task of the firmware is to detect the rising edges of the square wave generated by the voltage comparator that trigger interrupts on CC1110; these interrupt events are counted by microprocessor on the development board. CC1110 sends the number of steps per second to TI-EZ430 Chronos, every counted second. Packet dispatching resets of the counter is variable to zero, so the update of the steps rate is ready for the following 1s interval. At present to limit the processing on the microcontroller, the data are not processed or stored but just sent to smart watch. The EZ430- Chronos smart watch receives packets every second. Each packet is saved in a buffer of adjustable length, so steps rate is estimated by calculating the mean on the last received 10 samples. The step rate estimation is carried out in two steps that work as a simple digital filter: first the mean of 10 samples in the buffer is calculated, the average between this result and the previous estimate is calculated. The update of the estimation on the display every 1s is considered sufficient for a dynamic monitoring of the step rate during normal walking or running at moderate speed. The firmware has also tackled the timing and duty cycle of the CC1110. The latter is equipped with a 16-bit timer and three 8-bit timers, so timing is not critical for the update rate of 1s but the main task is to keep the interrupts detection interval always equal to 1s. Because in CC1110 interrupt tasks have higher priority than timer tasks, the microcontrollers can delay the timer task for running the code associated to the interrupt task. This is the reason why the processing tasks assigned to CC1110 are minimal and good timing can be preserved. On the other side, EZ430-Chronos can keep precisely the 1s time interval because timers have the highest priority. Moreover, it has a timing error of about 10-5s, which is negligible for the monitoring of the step rate. This operation mode is also optimal from the power management point of view, because both devices have a very low duty cycle, sleeping for most of the time as reported in Figure 9.
Figure 9: Programmable devices for operation with duty cycle.
It can be observed that a “sleeping” CC1110 has the RF module in idle state, but is capable of managing hardware interrupts. During receive and transmit modes the current consumption of the CC1110 is as low as 22 mA and 31 mA (at 0 dBm output power), respectively Indeed long listening period can significantly reduce CC battery lifespan. According to datasheet, CC1110 running in “Power mode 2” requires only 0.8 μA of current (32 kHz RC-oscillator and sleep timer running, RF module in idle state). The ultimate goal is to reduce duty cycle to 1.5-2% while maintaining the same accuracy update rate.
Graphical user interface on a smart watch and preliminary test
Finally, Figure 10 shows the GUI design of “Step Menu”. In this menu are displayed the step frequency with 3 digits with big size fonts and the total number of steps with smaller size fonts, and also the icon to monitor the status of wireless link with CC1110. The GUI provides a feedback to the user with a heart icon turning on each time the step rate is above a selected threshold. Threshold can be varied with UP/ DOWN button on the smart watch and total number of steps can be stored and cleared after a running session.
Figure 10: Developed GUI on TI EZ430-Chronos for the real time indication of frequency equal to step per second.

Conclusions and Future Developments

The technology of fabrication of conductive textile has been applied to study and develop a piezoresistive sensor in the form of a removable insole. The piezoresistive sensor is connected to flexible PCB technology with a portable read-out electronics that includes a CC1100 transceiver with microcontroller. This smart sensor transmits the count of voltage pulses to a smart watch. Finally the software on the TI EZ430-Chronos watch estimates the step rate and display on a custom GUI with an update rate of 1s. In a future development the sensors will be redesigned with the shape of electrodes capable to monitor the foot pressure spatial distribution for rehabilitation studies, gaming and optimization of runner’s conditions during sport activity. Another improvement is a “double-foot system”: at present the device counts the steps and estimates the frequency of only one foot, but the synchronous data acquisition from both feet is possible. While a good degree of optimization for the battery life span is possible by using power management techniques for portable devices, the reliable integration of electronics with smart textile sensors and the connections remains a challenging problem for wearable devices designers.

Acknowledgments

The author wish to acknowledge the master students team formed by Martina Rossi, Niccolò Bandini, Gianni Meoni and Matteo Lucarelli for development of the prototype and the support of Dr Riccardo Marchesi of Plug & Wear, Firenze Italy, for the production of samples of conductive textile fabrics and The Texas Instruments University Program.

References















international publisher, scitechnol, subscription journals, subscription, international, publisher, science

Track Your Manuscript

Awards Nomination

Associations