Received 26 December 2022, accepted 13 January 2023, date of publication 16 January 2023, date of current version 23 January 2023.

Digital Object Identifier 10.1109/ACCESS.2023.3237557

Development of Consumer-Friendly Surface

Electromyography System for
Muscle Fatigue Detection
RYUTA KINUGASA AND SHIMPEI KUBO
Faculty of Human Sciences, Kanagawa University, Yokohama, Kanagawa 221-8686, Japan
Corresponding author: Ryuta Kinugasa ([email protected])
This work was supported by the Japan Keirin Autorace Foundation under Grant 2020M-125.
This work involved human subjects or animals in its research. Approval of all ethical and experimental procedures and protocols was
granted by the Human Research Ethics Committee at Kanagawa University in 2022, and performed in line with the Declaration of Helsinki.

ABSTRACT In this study, a low-cost, wireless, and smartphone-controlled surface electromyography

(EMG) system was designed and developed for consumers, and the recorded EMG signals were evaluated
against a reference laboratory EMG system during fatiguing contraction. Using commercially available
inexpensive components, the components of the EMG signal-acquisition circuit were optimized, and a
microcontroller was combined with a Bluetooth module. The EMG signals were then converted from analog
to digital signals and transmitted to a smartphone via Bluetooth serial communication. EMG signals from
the biceps brachii of six healthy subjects were recorded separately using two EMG systems during sustained
submaximal isometric contraction until the endurance limit was reached. The root mean square (RMS)
and mean power frequency (MPF) of the EMG signals were calculated. The results indicated that both
the EMG systems exhibited a characteristic progressive increase in EMGRMS and decrease in EMGMPF
during sustained isometric contraction. The relative agreement between the two EMG systems, assessed by
intraclass correlation coefficient (ICC), was excellent for EMGRMS (ICC 0.933, P < 0.001) and moderate
for EMGMPF (ICC 0.662, P = 0.049). The cost of the sensor components in the hardware was ¥8,486
per unit. The proposed consumer-friendly EMG system, which is inexpensive and highly versatile in terms
of wireless and smartphone accessibility, can detect the phenomenon associated with increased amplitude
and low-frequency components during muscle fatigue contraction with a magnitude similar to that of the
commercially available laboratory EMG systems.

INDEX TERMS Amplitude, EMG, low-cost, smartphone, spectral analysis, sustained isometric contraction.

I. INTRODUCTION and decreased frequency [2]. For EMG systems to be useful

Accurate muscle fatigue estimation is essential for various in frequently detecting muscle fatigue, they must be highly
sports and rehabilitation. It also aids significantly in achiev- versatile, cost-effective, non-invasive, and wireless for porta-
ing efficient training and rehabilitation outcomes. Surface bility and ease of use.
electromyography (EMG) is a technique used to estimate Surface EMG systems have traditionally been used in
muscle fatigue. Using sustained submaximal isometric con- laboratory environments. However, they are typically expen-
tractions that induce muscle fatigue, defined as a progressive sive and not portable; moreover, the electrode and amplifier
exercise-induced decline in the force-generation capacity [1], module are often wired. Some commercial EMG devices
EMG signals are expected to exhibit increased amplitude are reportedly expensive, with costs ranging from approxi-
mately e (EUR) 15,000 to e20,000 (approximately ¥(JPY)
The associate editor coordinating the review of this manuscript and 2,200,000 to ¥2,942,000) [3] or up to $ (USD) 20,000
approving it for publication was Guido Lombardi . (approximately ¥2,750,000) [4]. For this reason, in several

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
6394 VOLUME 11, 2023
R. Kinugasa, S. Kubo: Development of Consumer-Friendly Surface EMG System for Muscle Fatigue Detection

studies, a low-cost EMG system has been developed. For A. CUSTOM-MADE PRODUCTS
example, Bawa [5] developed an EMG system using commer- The biggest advantage of custom-made products is cost effec-
cially available specialized modules, MyoWare and Arduino, tiveness. In previous studies, clarity regarding the cost of
for EMG measurement at a total cost of $150 (approximately all components used to fabricate the EMG systems is very
¥20,500). Yassin [6] developed EMG system using com- limited. Fortune et al. [7] developed an open-source active
mercially available microcontrollers (MCUs) and Bluetooth EMG circuit using thirteen components, and the total cost
chips without using specialized modules for EMG measure- of hardware was $112 (approximately ¥15,000) for one and
ment, at a total cost of $69.99 (approximately ¥9,500). $750 (approximately ¥100,300) for ten. This device allowed
Although these two studies do not mention all the compo- a large full scale input range, low baseline noise, and good
nents used, the total costs are roughly equivalent to that in interference suppression, thus bridging the gap between high
other studies listing all components: Fortune et al. [7] at $112 quality EMG and affordability. McKenzie et al. [8] designed,
(approximately ¥15,000 and McKenzie et al. [8] at $145 developed, and tested unique EMG, able to record EMG
(approximately ¥19,500). To the best of our knowledge, only signal concurrently to electrical stimulation. This design used
two studies has attempted to evaluate the variations in the total 59 components for fabrication, and the total cost was
EMG signals associated with muscle fatigue using low-cost $145 (approximately ¥19,500 for one and $548 (approxi-
EMG system [3], [5]. To analyze the amplitude and frequency mately ¥75,000) for five. Muraoka et al. [9] developed a
of EMG signals, del Toro et al. [3] used low-cost modules low-cost EMG device comprising an EMG amplifier, a PC
combinations of MyoWare sensor chip and Arduino board. with a microphone port, an electrode, and their cables and
However, these two studies using developed low-cost EMG reported the cost of the 11 components used to develop their
system have qualitatively compared the amplitude and fre- system at ¥5,200. The cost of these custom-made products is
quency of EMG signals during fatiguing contraction against a clearly lower than that of the commercial products described
laboratory-standard EMG system, which is yet to be compre- below.
hensively validated using the degree of agreement between In contrast, most custom-made products utilize some type
two EMG systems. In addition, the EMG system proposed of lead wire (i.e., wired connection) from the electrode to
in these studies were not versatile as the connection between the post-processing apparatus [7], [8], [9], [10], [11], [12],
electrode and board were wired, and they required postpro- [13], [14], [15], [16], [17]. To the best of the authors’
cessing using MATLAB (a highly specialized analysis soft- knowledge, none of these systems are completely wireless.
ware) on a personal computer (PC) having low portability. Zhao et al. [17] developed a readout circuit for acquiring
Therefore, unless the drawbacks of not only the high cost but EMG signals with six surface electrodes, comprising of wired
also the completely wired connection, postprocessing, and connections between the EMG acquisition circuit and elec-
portability are resolved, the general population is unlikely to trodes as well as the EMG acquisition circuit and battery.
use them on a daily basis. Although Yang et al. [16] eliminated the wires by placing
Therefore, a consumer-friendly surface EMG system that the substrate and electrodes in the same case, they used lead
consumers can use on a daily basis was designed and devel- wires to connect the substrate to the reference electrode.
oped in this study to assess muscle fatigue. This study This is a concern owing to its potential artifacts owing to
is the first attempt in developing a highly versatile low- cable movement as well as restriction of body movement and
cost, fully wireless, and smartphone-accessible EMG sys- clothing during EMG measurement.
tem. Low cost was achieved by using only commercially For EMG signal post-processing, several studies have uti-
available inexpensive components. Furthermore, the EMG lized MATLAB or LabVIEW on a PC [7], [9], [10], [11], [12],
signals recorded during fatiguing contractions are com- [13], [14], [16], [17]. For example, Geethanjali and Ray [12]
pared and validated against a reference laboratory EMG used a PC for processing the stored EMG data in MATLAB.
system. One of the issues was that a power adapter was required
to use the desktop PC, and specialized knowledge was
II. LITERATURE REVIEW needed to work with MATLAB and LabVIEW. Conversely,
The development of low-cost EMG systems can be divided Yassin et al. [6] developed a handheld EMG biofeedback
into two categories: custom-made and commercial products. device comprising a physiological signal-acquisition circuit
In this review, commercial products defined as those that and a smartphone with a cloud server. The smartphone col-
the plug-in-play EMG device such as Trigno as well as lected the EMG signals from the acquisition circuit and
dedicated components specialized for the EMG measure- uploaded them to the cloud. Suzuki et al. [18] visualized the
ment such as the MyoWare and BITalino while custom-made EMG wave form on a smartphone. Wu et al. [15] transmitted
products as those manufactured by combining components data using a Bluetooth low-energy module to a smartphone,
that are not modules specialized for the EMG measurement. which then used an infinite impulse response filter to remove
In this section, we have reviewed recent studies that devel- noise and display the EMG wave form on the screen. How-
oped devices aimed at measuring EMG signals noninvasively, ever, to the best of the authors’ knowledge, none of the
focusing on aspects related to versatility, such as wireless aforementioned studies analyzed the EMG signals on the
connection, EMG signal post-processing, and cost. smartphone itself.
VOLUME 11, 2023 6395
 R. Kinugasa, S. Kubo: Development of Consumer-Friendly Surface EMG System for Muscle Fatigue Detection

Finally, comparing the EMG signals measured by the for EMG measurement is approximately $38 (approximatel
custom-made products with those of a commercially avail- ¥5,200) to ¥50,000 [4], no studies that employed wireless
able EMG system is essential. Although few studies have connectivity and post-processing on the smartphone were
developed such systems, only three studies have quantita- found.
tively verified their validity with correlation coefficients of
over 0.8 [6], [15], [16] representing moderate-to-high relative III. METHODS
agreement between the custom-made and commercial EMG A. SYSTEM DESCRIPTION
systems. For the former, the EMG signals must also be veri- The proposed EMG system comprises of two parts, a sensor
fied to be consistent with those of commercial products. module and an Android-based smartphone (Fig. 1a). The
sensor module captures analog EMG signals via surface
B. COMMERCIAL PRODUCTS electrodes placed on the skin over the muscles of interest.
Two types of commercial products have been observed: After amplification and filtering, the analog signals are trans-
systems with plug-in-play device and specialized board mitted to an MCU that digitizes the EMG signals using an
infrastructure for EMG measurement. BiostampRC [2], analog-to-digital (A–D) converter for sampling. A universal
mDurance [19], and Duelite [20] have been primarily used serial bus (USB) on-the-go interface is implemented in the
in studies as systems with inexpensive plug-in-play device. MCU to communicate with an Android-based smartphone
Doheny et al. [2] used the BiostampRC, a low-power MCU for data control and transmission. An Android application
that acquires signals from the EMG sensors as well as was also developed to process EMG signals and enable user
a 3-axis accelerometer and gyroscope. It then processed, interaction.
sampled, and transmitted data to flash memory or broadcasted
wirelessly via Bluetooth. The data collected during fatiguing 1) HARDWARE
contraction were then post-processed in MATLAB on a PC. The hardware comprises an electrode, instrumentation ampli-
While the BiostampRC was completely wireless, the mDu- fier, operational amplifier, a Bluetooth module, lithium poly-
rance and Duelite utilized a wired connection between the mer rechargeable battery, and small commercially available
electrodes and bridge amplifier or probe for sampling. For container (SW-53S, Takachi, Saitama, Japan) with dimen-
post-processing, the mDurance Android mobile application sions 35 × 11 × 53 mm (w × h × d) and weight 19.0 g.
receives the EMG data from the acquisition unit and transmits Furthermore, commercially available stainless-steel balls of
it to a cloud service. The other plug-in-play devices analyze diameter 5 mm (SZ54, Hikari, Osaka, Japan) were used as
the EMG signals via commercial programming software, electrodes at distances of 30 mm. The EMG signals obtained
such as MATLAB, necessitating programming skills and spe- from the differential electrodes are amplified using a differen-
cific knowledge. The cost of these plug-in-play devices range tial amplifier circuit consisting of an instrumentation ampli-
from e538 (approximately ¥78,600) to ¥680,000. fier (AD8237ARMZ-R7, Analog Devices, Tokyo, Japan).
For specialized board infrastructure for EMG meas- Differential noise was removed, and the gain of the instru-
urement, the MyoWare [5], [21], the Arduino board [3], [22], mentation amplifier was set to 20 dB to eliminate the in-phase
[23], and the BITalino [24] have often been used. components of the positive and negative electrodes. The input
Heywood et al [21] used MyoWare board and incorporated impedance and common-mode rejection ratio (CMRR) of
an Analog Device operational amplifier. There was a cable the differential amplifier were 100 M and 120 dB, respec-
connection between board and electrode. All analyses were tively. Active electrodes ideally present a significantly high
performed using LabVIEW on a PC. They reported cor- input impedance and low output impedance, resulting in the
relation coefficients of 0.65–0.99, indicating moderate-to- acquisition systems being considerably robust to power line
high relative agreement with those of commercial system. interference and high electrode-skin impedance values [26].
Tecchino et al. [23] built a low-cost EMG system with In this study, the sampling frequency was set to the high-
an Arduino Uno ethernet board and an integrated signal est frequency at which no missing data would occur on
conditioning block for biological signals. This system saved the smartphone. The EMG signals were bandpass filtered
data automatically in a secure digital card, and postprocess (5–500 Hz) with a sampling frequency of 2,500 Hz, in accor-
analysis was done by MATLAB on a PC. During maxi- dance with the recommendations of the Surface EMG for
mal voluntary contraction, an excellent relative agreement Non-Invasive Assessment of Muscles (SENIAM) consortium
was observed (correlation coefficient = 0.30–0.98). Donisi project [27]. The amplitude-frequency characteristic of the
et al. [24] used BITalino sensor, a purpose-built sensor, primary amplifier acting as a bandpass filter led to the gain
to measure EMG signals. Several cable codes were required being reduced to 20 dB (1/10) outside the 5–500 Hz band.
to connection between electrode and main platform. Data Consequently, two operational amplifiers (LMV358IDR,
were analyzed using MATLAB on a PC. Batista et al. [25] Texas Instruments, Texas, USA) were placed in series, one as
provided evidence validating BITalino, and the computation a low-pass and the other as a high-pass filter with threshold
of the linear envelope resulted in a small difference and frequencies of 500 Hz and 5 Hz, respectively. Their max-
high correlation between the BITalino and reference device. imum gain was 1,000 (60 dB), which was calculated by
Although the price for these specialized board infrastructure multiplying the gains of the instrumentation and operational

6396 VOLUME 11, 2023

R. Kinugasa, S. Kubo: Development of Consumer-Friendly Surface EMG System for Muscle Fatigue Detection

the baseline value to just before the beginning of the next

active phase. The application displayed real-time dynamic
charts of the EMG signals during voluntary contraction and
automatically detected the amplitude of the EMG signals
based on phase classification (Fig. 1b). The EMG data were
stored in the internal memory of the smartphone.

IV. EXPERIMENT
A. SUBJECTS
Six healthy male subjects (age: 25.3 ± 9.8 years; height:
170.3 ± 5.6 cm; weight: 69.7 ± 10.8 kg) with no history of
neurological disorders participated in this study. Since men
fatigue more easily than women during elbow flexion [9],
which is the target joint in this study, only male subjects
were included. The participants provided informed consent,
and all the experimental procedures were approved by the
local ethics committee (2022–22). The study was conducted
in accordance with the tenets of the Declaration of Helsinki.

B. FORCE AND EMG RECORDINGS

FIGURE 1. Overall design of the proposed EMG system. (a) Magnified Each subject was seated upright in a neutral position on a
view of the sensor. (b) Smartphone screenshot showing the automatic
detection of peak amplitude of EMG signal amplitude during muscle chair with the left upper arm and forearm in the vertical and
contraction. (c) Image of the circuit board in the sensor showing each horizontal positions, respectively, and the elbow joint flexed
functional module: (1) universal serial bus terminal for charging,
(2) Bluetooth low-energy module including an advanced reduced
at 90◦ (Fig. 2). The elbow was supported by a pad to restrain
instruction set computer (RISC) Machine (ARM) processor and arm movements. The forearm was tightly fixed with a Velcro
analog/digital converter, (3) connections to the lithium polymer battery, strap attached to a dynamometer (Biodex System 4, Biodex,
(4) filter implemented by two operational amplifiers, (5) reference voltage
generation, (6) electrode, and (7) instrumentation amplifier. New York, USA). The upper arm and forearm were aligned
with the principal axes of the dynamometer such that elbow
flexion moment primarily produced a force in the upward
amplifiers as 10 (20 dB) × 100 (40 dB). An A–D converter
direction. Before sustained contraction to induce fatigue, the
and wireless control were used in a Bluetooth low-energy
subject performed maximum voluntary contraction (MVC)
module (EYSKJNZWB, Taiyo Yuden, Tokyo, Japan) con-
using the elbow flexor muscles by gradually increasing force
taining a 12-bit ARM processor as the central processing unit
from rest to maximum over 3 s, sustaining maximum force
along with a Bluetooth 5.0 compliant chip and antenna man-
for 2–3 s, and then decreasing the force back to baseline.
ufactured by Nordic. The analog EMG signals were digitized
with 12-bit accuracy and transmitted to a smartphone (Pixel
4a, Google, Mountain View, CA, United States of America)
via Bluetooth serial communication. Owing to the wide range
of the EMG signal amplitude, a high dynamic range and
variable gain are required for the 12-bit A–D converter based
on the recommendations of the SENIAM [28].The dynamic
range of the A–D converter was 3 V, and its amplification
was varied for the EMG signal to reach maximum value. The
system was powered by a 110 mAh lithium polymer battery
(DTP401525, Data Power Technology, Tokyo, Japan).

2) SOFTWARE
An Android application was developed to visualize and
acquire the EMG signals. The application detected the base-
line, active, and relaxed phases during muscle contraction
bouts. The baseline phase was defined as a resting state
for 2–3 s long when the subjects were instructed to fully
relax their arm, and the active phase as the time when the
EMG signals exceeded four standard deviations (SDs) of the FIGURE 2. Overview of experimental setup. The subject was seated with
arms slightly abducted from the trunk, elbow flexed at 90◦ , and forearm
baseline value to just before the relax phase. The relax phase in a neutral position. The electrode was placed over the biceps brachii
was defined as the time when the EMG signals were less than muscle.

VOLUME 11, 2023 6397

 R. Kinugasa, S. Kubo: Development of Consumer-Friendly Surface EMG System for Muscle Fatigue Detection

The MVCs were repeated three times, and the subject was into 10 s epochs. For each epoch, the root mean square
verbally encouraged to produce maximal effort. A 60 s rest (RMS) and mean power frequency (MPF) of the EMG sig-
period was incorporated between trials. The greatest force nals (EMGRMS and EMGMPF , respectively) were calculated
achieved by the subject was taken as the MVC force and used based on the Fourier transform. EMGRMS was calculated to
as the reference to calculate the 50% target level for sustained examine muscle fatigue-induced changes in the amplitude of
contraction. muscle activity, defined as follows
Sustained isometric elbow flexion was performed at a tar- sP
n
get value of 50% of the MVC force. This protocol success- i=1 |rawDatai |
EMGRMS = , (1)
fully induced muscle fatigue in a previous study [30]. The n
subject was required to match the target torque as displayed
on the monitor and verbally encouraged to sustain the torque where i is the order number of the dealing sample point, raw
for as long as possible. The sustained contraction was termi- data is the value of the ith sample point, and n is the total
nated when the force declined by 10% of the target value for number of data points.
longer than ∼0.1 s despite verbal encouragement to maintain To examine the muscle fatigue-induced changes in the
the task. The force was recorded using an A–D converter characteristic spectral frequency of EMG signals, EMGMPF ,
(PowerLab 16/35, AD Instruments, Sydney, Australia) at a calculated as the frequency centroid of the spectrum,
sampling frequency of 2,000 Hz. is defined as follows:
R∞
Two measurements were separately obtained using the pro- S (f ) · f · d (f )
posed and commercial EMG systems for research purposes EMGMPF = 0 R ∞ , (2)
0 S (f ) · fd
(Trigno, Delsys, Boston, MA, United States of America). The
order of measurement was randomized with a rest interval of where f is the frequency, S(f ) is the power at frequency f ,
at least 2 h between each session. After carefully cleaning and d(f ) is the frequency resolution.
the skin with alcohol, an electrode was placed on the skin EMGRMS and EMGMPF were calculated for 10 s in steps
over the short head of the biceps brachii according to the of 1 s [31]. The initial and final values for EMGRMS and
guidelines of surface EMG [27]. The placement of the elec- EMGMPF were defined as the beginning and ending values of
trode was marked with a permanent marker to ensure that the the sustained isometric contraction, respectively. The relative
electrode was completely placed at the same spot during each changes in EMGRMS and EMGMPF were also defined as the
measurement. The EMG signals were measured using stan- final value normalized by the initial value. The signal-to-
dard Trigno wireless sensors (10 mm interelectrode distance, noise ratio (SNR) was calculated as the value of the raw
CMRR > 80 dB @60 Hz, gain of 100, transmission range EMG signal during sustained muscle contraction relative to
of 40 m, durable battery lifetime of 4–8 h, 16-bit resolution, the value of the baseline noise. Battery lifetime was tested for
500µs inter-sensor latency) and amplified and bandpass fil- continuous EMG signal measurement and transmission to a
tered (20–450 Hz) using an A–D converter (PowerLab 16/35, smartphone via Bluetooth.
ADInstruments) at a sampling frequency of 2,000 Hz. The
price of a Trigno EMG system package (16 channels) is D. STATISTICS
¥4,751,250 (Price at time of purchase). Table 1 shows a All data are presented as mean ± standard deviation. The
comparison between the key features of our proposed and the difference in the SNR of the EMG signal and the normalized
referenced systems. EMGRMS and EMGMPF between the two EMG systems was
evaluated using paired t-tests. EMGRMS and EMGMPF were
TABLE 1. Feature comparison of proposed and referenced systems. analyzed using repeated measures two-way analysis of the
variance (time × EMG system). Post-hoc tests (Bonferroni
pairwise comparisons) were used to investigate the even-
tual differences between the two EMG systems. The relative
agreement of the given variables as assessed by both the EMG
systems was tested using the intra-class correlation coeffi-
cient (ICC)(2,1). Estimates of correlation were interpreted as
excellent (0.75–1), modest (0.4–0.74), or poor (0–0.39) [32].
The significance level was set at P < 0.05.

V. RESULTS
The mean time to complete a sustained isometric contraction
for all the subjects was 83.5 ± 12.4 s (range: 68–96 s)
C. DATA ANALYSIS for the proposed EMG system and 87.2 ± 23.1 s (range:
The EMG signals acquired via both the systems were ana- 52–108 s) for the commercial EMG system (Fig. 3). The
lyzed using an in-house MATLAB scrip (R2017a, Math- SNR was significantly lower in the proposed EMG sys-
works, Natick, MA, USA). The EMG signals were segmented tem (77.4 ± 6.8 dB) than in commercial EMG systems

6398 VOLUME 11, 2023

R. Kinugasa, S. Kubo: Development of Consumer-Friendly Surface EMG System for Muscle Fatigue Detection

FIGURE 5. Comparisons for EMGRMS (left) and EMGMPF (right) of the raw
EMG in the initial and final values of sustained isometric contraction
between the two EMG systems. The cross represents the average value.
Data from the proposed EMG system are denoted with red boxplots while
those from the commercial EMG system are denoted with black boxplots;
∗ indicates the main effect of time, P < 0.05, and # indicates the main
effect of the EMG systems, P < 0.05.

FIGURE 6. Normalized, defined as the final value normalized by the

FIGURE 3. Typical example (n = 1) of progressive changes in torque, raw initial value, EMGRMS (left), and EMGMPF (right) of the raw EMG
EMG, EMGRMS , and EMGMPF during sustained isometric contraction of the measured by the proposed and commercial EMG systems. The dashed
elbow flexor muscles at 50% of MVC till the endurance limit was reached. line is the line of identity, and the solid line is the line of best fit.
Two measurements were performed: the red indicates our proposed EMG
system and black indicates the commercial EMG system. EMGRMS and
EMGMPF were calculated over a 10 s window along the signal in 1 s steps. TABLE 2. List and cost of hardware components.

FIGURE 4. Typical example (n = 1) of the power spectral density of the

surface EMG signals during sustained isometric contraction of the elbow
flexor muscles at 50% of MVC, till the endurance limit was reached. Two
measurements were performed: the red indicates our proposed EMG
system and black indicates the commercial EMG system.
the two EMG systems, while normalized EMGMPF indicated
(116.5 ± 18.8 dB) (P < 0.001). No significant difference in modest agreement (ICC = 0.662, P = 0.049) (Fig. 6). No sig-
power spectral density was observed between the two EMG nificant difference was observed in the normalized EMGRMS
systems (Fig. 4). The interaction between time and EMG and EMGMPF between the proposed and commercial EMG
systems was not significant for EMGRMS and EMGMPF . systems. Our system operated for over 8 hours on a 110 mAh
However, it was significant for EMGRMS (P = 0.002) and battery. Lastly, Table 2 lists all the hardware components
EMGMPF (P < 0.001), suggesting that regardless of the EMG and their costs. The total cost of hardware components was
system, EMGRMS was significantly higher in the final value ¥41,376 per unit, and the price of the smartphone was 79.5%
than in the initial value, while EMGMPF exhibited the oppo- of the total cost.
site trend (Fig. 5). The effect of the EMG system was signif-
icant for EMGMPF (P = 0.002), suggesting that, regardless VI. DISCUSSION
of time, EMGMPF showed significantly lower values in the This is the first study to develop a versatile EMG for con-
proposed EMG system. Normalized EMGRMS demonstrated sumer use and to validate EMG signals during fatiguing
an excellent agreement (ICC = 0.933, P < 0.001) between contraction. As raw EMG signals are significantly sensitive

VOLUME 11, 2023 6399

 R. Kinugasa, S. Kubo: Development of Consumer-Friendly Surface EMG System for Muscle Fatigue Detection

TABLE 3. Key characteristics comparison of our proposed system with commercial devices.

to noise from the acquisition device and measurement envi- 116.5 ± 18.8 dB. Sensor location and skin preparation, which
ronment, investigating the EMG signals obtained by the influence SNR, are completely identical for our proposed and
custom-made product to verify their similarity to those commercial EMG systems. Consequently, the influence of
by a reference product is important. The obtained results physical factors for the SNR difference between the two EMG
demonstrate that the relative agreement between the proposed systems is eliminated. Hence, the hardware components may
and commercial EMG systems is significant for EMGRMS be involved. The CMRR determines the amount of common
(ICC = 0.933, P < 0.001) and EMGMPF (ICC = 0.662, noise that can be removed from the signal. Additionally,
P = 0.049). Few studies have validated the use of custom- it has a direct impact and is approximately proportional to the
made, low-cost EMG systems [6], [15], [16]. For exam- SNR [11]. Nevertheless, the SNR of our proposed system is
ple, Yang et al. [16] implemented a low-cost surface EMG significantly lower than that of commercial systems, possibly
acquisition system combined with an MCU and a Wi-Fi due to differences in input impedance. To prevent attenuation
module. They calculated the mean absolute value of raw and distortion of the detected signal due to the effects of input
EMG during comfortable walking. The ICC between custom- loading, the input impedance of the differential amplifier
mode and commercial device was 0.86 for the vastus lateralis should be as large as possible [33]. Fu et al. [11] reported
and 0.81 for the biceps femoris. To our best knowledge, that an SNR between 37.0–46.2 dB using a low-cost EMG
no studies have compared custom-made and commercial amplifier with discrete operational amplifiers. In addition,
EMG systems during fatiguing contraction. According to Prakash et al. [13] compared the SNR between self-made
one of the EMG device-manufacturing companies, 1.2 dB EMG sensors (18.9 dB) and commercially available
is an acceptable SNR [33]. A higher SNR indicates a better MyoWare sensor (4.8 dB). Therefore, this proves that the
signal quality. The SNR of the proposed EMG system was proposed EMG system not only detects fatigue accurately,
77.4 ± 6.8 dB, while that of the commercial EMG system was but also achieves adequate noise immunity.

6400 VOLUME 11, 2023

R. Kinugasa, S. Kubo: Development of Consumer-Friendly Surface EMG System for Muscle Fatigue Detection

Accordingly, the proposed EMG system was successfully duction velocity decreases, and the low-frequency compo-
fabricated using easily available and low-cost components; nent of EMG signals increases, increasing the amplitude
moreover, its circuit board was developed in-house, and the of EMG signals by allowing more energy to reach the
system is controlled with a smartphone. The list of com- electrodes owing to the low-pass filter effect [43]. The
mercial devices in Table 3 was refers based on the previous proposed EMG system developed in this study succeeded
studies [4], [34]. Only MyoWare and our proposed system in reproducing these changes in EMG signals associated
cost less than ¥10,000. Although limited studies mention the with muscle fatigue in accordance with the aforementioned
market prices of the packaged EMG systems, some commer- theory.
cial products are significantly expensive, with prices rang- The proposed EMG system opens the possibility of contin-
ing from approximately e15,000 to e20,000 (approximately uous muscle fatigue monitoring during everyday life of the
¥2,200,000 to ¥2,942,000) [3] or up to $20,000 (approxi- user. One advantage of our system is that it sends the data of
mately ¥2,750,000) [4]. In contrast, several EMG sensors the patien’s or athlete to the physiotherapists or clinicians in
are sold in the market. del Toro et al. [3] reported the cost real time. This system will reduce treatment time and costs as
of their system in the range of e100–e150 (approximately it will allow the patient or athlete to implement rehabilitation
¥14,700–¥22,000), which is not significantly different from exercises remotely and reduce the number of routine visits
the price of the proposed EMG system. However, EMG sen- to the clinic. Our findings indicate an excellent agreement
sors face several challenges. One problem is that postpro- between our proposed and commercial EMG systems dur-
cessing must be performed by the user, which is difficult. ing fatiguing contraction, although the various limitations
Additionally, the cables extending from the sensors are wired to this study must be noted. Firstly, as we aimed to pro-
to the amplifier, which are detrimental to the EMG measure- duce a low-cost EMG sensor using commercially available
ments. In contrast, the proposed EMG system is completely inexpensive components, a single channel and low battery
wireless and accompanied by a smartphone application that life are observed. In the future, multi-channel systems and
handles the postprocessing, making it superior to the EMG extended battery life will be required for daily use. Sec-
systems available in the market. Consequently, our proposed ondly, to design a controlled experimental environment, the
system has better usability than other commercial devices results presented here were limited to the biceps brachii
considering highly versatile due to wireless and smartphone muscle and isometric contraction in a relatively small group
accessibility. of primarily younger healthy subjects, without considering
Previous studies have indicated that relative anatomical changes associated with aging or pathological conditions.
alignment and movement of the muscle fibers, location of The features of the surface EMG signal and changes with
the innervation zone, and crosstalk between muscles can fatigue may differ across muscles, with subject anatomy and
influence EMG signals [38]. Different electrode positions and subcutaneous fat thickness. In particular, the tissue between
interelectrode distances can also affect EMG signals [39]. the electrode and muscle fibers has a spatial low-pass filter-
In this study, both EMG systems were derived the signal from ing effect on the EMG signal detected at the skin surface,
the exact same anatomical location; therefore, the observed with decreases both the amplitude and frequency content of
differences cannot be attributed to their structural factors. The the signal [44]. Despite these limitations, establishing signal
observed differences are primarily attributed to the differ- properties under isometric conditions in healthy subjects is
ences in the electronic components. One potential possibility a critical first step before progressing to a wider range of
is the difference in inter-electrode distances. The 30 mm inter- conditions.
electrode distance of the proposed EMG system is generally
recommended for bipolar surface EMG [27] since it is larger VII. CONCLUSION
than that of the Trigno sensor. Theoretical, experimental, and The novelty of the present study is that we have developed a
simulation investigations have established that the frequency consumer-friendly EMG system that is easy to use (entirely
content of a surface EMG signal decreases with increasing wireless, smartphone-controlled, and battery-powered) and
inter-electrode distance because of increased detection vol- highly accurate, and low-cost (less than $100 (approximately
ume and reduced spatial filtering of the signal [40], [41], ¥10,000 yen)). Experimental results indicate that the pro-
[42]. Furthermore, the commercial EMG system has a larger posed EMG system, which is highly versatile owing to
gain (100 vs. 20 dB) and higher quality A–D converter its wireless design and smartphone accessibility, can accu-
(16 vs. 12 bits), thereby enhancing signal amplification rately detect changes in EMG amplitude and frequency
when EMG activity is relatively low. This suggests that during fatiguing isometric contractions. This EMG system
to investigate the EMG signal during submaximal contrac- can monitor EMG signals associated with muscle fatigue
tions, attention must be paid when using low-cost EMG during daily activities of the user. This study enables the
systems. possibility of further evaluating stiff shoulders, back pain,
Furthermore, EMGRMS progressively increases and and swallowing difficulties. Additionally, we aim to support
EMGMPF decreases during sustained submaximal isomet- the early prediction of diabetes among healthy individu-
ric contraction, where the force level is constant until als, which has been difficult to evaluate with conventional
exhaustion. When the muscles are fatigued, the nerve con- EMG devices.

VOLUME 11, 2023 6401

 R. Kinugasa, S. Kubo: Development of Consumer-Friendly Surface EMG System for Muscle Fatigue Detection

ACKNOWLEDGMENT [19] A. Molina-Molina, E. J. Ruiz-Malagón, F. Carrillo-Pérez,

The authors would like to thank Hiroshi Ogawa, the President L. E. Roche-Seruendo, M. Damas, O. Banos, and F. García-Pinillos,
‘‘Validation of mDurance, a wearable surface electromyography system
of SaLusTek Company for developing the hardware; Toshio for muscle activity assessment,’’ Frontiers Physiol., vol. 11, Nov. 2020,
Suzuki, the President of Bifrostec Company for developing Art. no. 606287.
the wireless firmware; and Shoichiro Takagi, the President of [20] Y. Zhao, S. Zhang, T. Yu, Y. Zhang, G. Ye, H. Cui, C. He, W. Jiang,
Y. Zhai, C. Lu, X. Gu, and N. Liu, ‘‘Ultra-conformal skin electrodes
VS Games Company for developing the smartphone software with synergistically enhanced conductivity for long-time and low-motion
application and also would like to thank Yoshitaka Miyake at artifact epidermal electrophysiology,’’ Nature Commun., vol. 12, no. 1,
the Kihara Memorial Yokohama Foundation for the Advance- p. 4880, Aug. 2021.
[21] S. Heywood, Y. H. Pua, J. McClelland, P. Geigle, A. Rahmann,
ment of Life Sciences for introduction with companies.
K. Bower, and R. Clark, ‘‘Low-cost electromyography—Validation against
a commercial system using both manual and automated activation timing
REFERENCES thresholds,’’ J. Electromyography Kinesiol., vol. 42, pp. 74–80, Oct. 2018.
[22] S. F. del Toro, S. Santos-Cuadros, E. Olmeda, C. Álvarez-Caldas, V. Díaz,
[1] S. C. Gandevia, ‘‘Spinal and supraspinal factors in human muscle fatigue,’’
and J. L. San Román, ‘‘Is the use of a low-cost sEMG sensor valid to
Physiol. Rev., vol. 81, no. 4, pp. 1725–1789, Jan. 2001.
measure muscle fatigue?’’ Sensors, vol. 19, no. 14, p. 3201, Jul. 2019.
[2] E. P. Doheny, C. Goulding, M. W. Flood, L. Mcmanus, and M. M. Lowery,
[23] P. Tecchino, A. Monte, and P. Zamparo, ‘‘Low-cost electromyography:
‘‘Feature-based evaluation of a wearable surface EMG sensor against lab-
Validity against a commercial system depends on exercise type and inten-
oratory standard EMG during force-varying and fatiguing contractions,’’
sity,’’ Eur. J. Transl. Myol., vol. 31, no. 2, p. 9735, May 2021.
IEEE Sensors J., vol. 20, no. 5, pp. 2757–2765, Mar. 2020.
[24] L. Donisi, E. Capodaglio, G. Pagano, F. Amitrano,
[3] S. Fuentes del Toro, Y. Wei, E. Olmeda, L. Ren, W. Guowu, and
M. Cesarelli, M. Panigazzi, and G. D’Addio, ‘‘Feasibility of tree-
V. Díaz, ‘‘Validation of a low-cost electromyography (EMG) system via
based machine learning algorithms fed with surface electromyographic
a commercial and accurate EMG device: Pilot study,’’ Sensors, vol. 19,
features to discriminate risk classes according to NIOSH,’’ in Proc. IEEE
no. 23, p. 5214, Nov. 2019.
Int. Symp. Med. Meas. Appl. (MeMeA), Jun. 2022, pp. 1–6.
[4] H. Manzur-Valdivia and J. Alvarez-Ruf, ‘‘Surface electromyography in
clinical practice. A perspective from a developing country,’’ Frontiers [25] D. Batista, H. Plácido da Silva, A. Fred, C. Moreira, M. Reis, and
Neurol., vol. 11, Oct. 2020, Art. no. 578829. H. A. Ferreira, ‘‘Benchmarking of the BITalino biomedical toolkit against
an established gold standard,’’ Healthcare Technol. Lett., vol. 6, no. 2,
[5] A. Bawa and K. Banitsas, ‘‘Design validation of a low-cost EMG sensor
pp. 32–36, Apr. 2019.
compared to a commercial-based system for measuring muscle activity and
fatigue,’’ Sensors, vol. 22, no. 15, p. 5799, Aug. 2022. [26] S. Johnson, P. Lynn, J. Miller, and G. Reed, ‘‘Miniature skin-mounted
[6] M. M. Yassin, A. M. Saber, M. N. Saad, A. M. Said, and A. M. Khalifa, preamplifier for measurement of surface electromyographic potentials,’’
‘‘Developing a low-cost, smart, handheld electromyography biofeedback Med. Biol. Eng. Comput., vol. 15, no. 6, pp. 710–711, Nov. 1977.
system for telerehabilitation with clinical evaluation,’’ Med. Novel Technol. [27] B. Freriks, H. J. Hermens, C. Disselhorst-Klug, and G. Rau, ‘‘The rec-
Devices, vol. 10, Jun. 2021, Art. no. 100056. ommendations for sensors and sensor placement procedures for surface
[7] B. C. Fortune, C. G. Pretty, L. T. Chatfield, L. R. McKenzie, and electromyography,’’ in European Recommendations for Signal Pro-
M. P. Hayes, ‘‘Low-cost active electromyography,’’ HardwareX, vol. 6, cessing Methods for Surface ElectroMyoGraphy, 2nd ed. Enschede,
Oct. 2019, Art. no. e00085. The Netherlands: Roessingh Research and Development, 1999, pp. 15–53.
[8] L. R. McKenzie, C. G. Pretty, B. C. Fortune, and L. T. Chatfield, ‘‘Low-cost [28] R. Merletti, D. Farina, H. Hermens, B. Freriks, and J. Harlaar, ‘‘European
stimulation resistant electromyography,’’ HardwareX, vol. 9, Apr. 2021, recommendations for signal processing methods for surface electromyo-
Art. no. e00178. Graphy,’’ in European Recommendations for Surface Electromyography,
[9] Y. Muraoka, A. Ishio, and K. Takeda, ‘‘Low-cost 2ch EMG biofeedback 2nd ed. Enschede, The Netherlands: Roessingh Research and Develop-
device using a stereo microphone port,’’ Jpn. J. Compr. Rehabil. Sci., vol. 5, ment, 1999, pp. 57–70.
pp. 1–6, Jan. 2014. [29] K. G. Avin, M. R. Naughton, B. W. Ford, H. E. Moore,
[10] G. Biagetti, P. Crippa, L. Falaschetti, S. Orcioni, and C. Turchetti, M. N. Monitto-Webber, A. M. Stark, A. J. Gentile, and L. A. Frey
‘‘Wireless surface electromyograph and electrocardiograph system on Law, ‘‘Sex differences in fatigue resistance are muscle group dependent,’’
802.15.4,’’ IEEE Trans. Consum. Electron., vol. 62, no. 3, pp. 258–266, Med. Sci. Sports Exerc., vol. 42, no. 10, pp. 1943–1950, Oct. 2010.
Aug. 2016. [30] J. Duchateau, C. Balestra, A. Carpentier, and K. Hainaut, ‘‘Reflex reg-
[11] Z. Fu, A. Y. B. Hashim, Z. Jamaludin, and I. S. Mohamad, ‘‘Design of ulation during sustained and intermittent submaximal contractions in
a low cost EMG amplifier with discreet op-amps for machine control,’’ humans,’’ J. Physiol., vol. 541, no. 3, pp. 959–967, Jun. 2002.
ARPN J. Eng. Appl. Sci., vol. 11, no. 5, pp. 3345–3349, Mar. 2016. [31] J. Ushiyama, M. Katsu, Y. Masakado, A. Kimura, M. Liu, and J. Ushiba,
[12] P. Geethanjali and K. K. Ray, ‘‘A low-cost real-time research platform ‘‘Muscle fatigue-induced enhancement of corticomuscular coherence fol-
for EMG pattern recognition-based prosthetic hand,’’ IEEE/ASME Trans. lowing sustained submaximal isometric contraction of the tibialis anterior
Mechatronics, vol. 20, no. 4, pp. 1948–1955, Aug. 2015. muscle,’’ J. Appl. Physiol., vol. 110, no. 5, pp. 1233–1240, 2011.
[13] A. Prakash, B. Kumari, and S. Sharma, ‘‘A low-cost, wearable sEMG [32] J. L. Fleiss, ‘‘Reliability of measurement,’’ in The Design and Analysis of
sensor for upper limb prosthetic application,’’ J. Med. Eng. Technol., Clinical Experiments. Hoboken, NJ, USA: Wiley, 1986.
vol. 43, no. 4, pp. 235–247, May 2019. [33] Signal Quality Monitor, Delsys. Accessed: Oct. 19, 2022. [Online].
[14] T. Supuk, A. Skelin, and M. Cic, ‘‘Design, development and testing of a Available: https://delsys.com/emgworks/signal-quality-monitor/
low-cost sEMG system and its use in recording muscle activity in human improve/#6a368b2365534dedb
gait,’’ Sensors, vol. 14, no. 5, pp. 8235–8258, May 2014. [34] A. Palumbo, P. Vizza, B. Calabrese, and N. Ielpo, ‘‘Biopotential signal
[15] Y.-D. Wu, S.-J. Ruan, and Y.-H. Lee, ‘‘An ultra-low power surface EMG monitoring systems in rehabilitation: A review,’’ Sensors, vol. 21, no. 21,
sensor for wearable biometric and medical applications,’’ Biosensors, p. 7172, Oct. 2021.
vol. 11, no. 11, p. 411, Oct. 2021. [35] K. I. Jang, H. N. Jung, J. W. Lee, S. Xu, and Y. H. Liu, ‘‘Ferromag-
[16] Y.-H. Yang, S.-J. Ruan, P.-C. Chen, Y.-T. Liu, and Y.-H. Hsueh, netic, folded electrode composite as a soft interface to the skin for long-
‘‘A low-cost wireless multichannel surface EMG acquisition system,’’ term electrophysiological recording,’’ Adv. Funct. Mater., vol. 25, no. 26,
IEEE Consum. Electron. Mag., vol. 9, no. 5, pp. 14–19, Sep. 2020. pp. 7281–7290, Oct. 2016.
[17] Y. Zhao, J. Wang, Y. Zhang, H. Liu, Z. Chen, Y. Lu, Y. Dai, L. Xu, and [36] A. Gonzalez-Mendoza, A. I. Perez-SanPablo, R. Lopez-Gutierrez, and
S. Gao, ‘‘Flexible and wearable EMG and PSD sensors enabled locomotion I. Quinones-Uriostegui, ‘‘Validation of an EMG sensor for Internet of
mode recognition for IoHT-based in-home rehabilitation,’’ IEEE Sensors Things and robotics,’’ in Proc. 15th Int. Conf. Electr. Eng., Comput. Sci.
J., vol. 21, no. 23, pp. 26311–26319, Dec. 2021. Autom. Control (CCE), Sep. 2018, pp. 1–5.
[18] R. Suzuki, Y. Muraoka, and S. Okazaki, ‘‘Examination of validity of [37] J. Guerreiro, A. Lourenço, H. Silva, and A. Fred, ‘‘Performance compar-
low—Cost electromyog-raphy feedback device by using smartphone,’’ ison of low-cost hardware platforms targeting physiological computing
Jpn. J. Biofeedback Res., vol. 44, no. 1, pp. 15–20, May 2017. applications,’’ Proc. Technol., vol. 17, pp. 399–406, Jan. 2014.

6402 VOLUME 11, 2023

R. Kinugasa, S. Kubo: Development of Consumer-Friendly Surface EMG System for Muscle Fatigue Detection

[38] D. Farina, C. Cescon, and R. Merletti, ‘‘Influence of anatomical, physical, RYUTA KINUGASA received the Ph.D. degree in
and detection-system parameters on surface EMG,’’ Biol. Cybern., vol. 86, human sciences from Waseda University, Japan,
no. 6, pp. 445–456, 2002. in 2007. He was a Postdoctoral Fellow at the
[39] M. B. Lanza, T. G. Balshaw, and J. P. Folland, ‘‘Do changes School of Medicine, University of California at
in neuromuscular activation contribute to the knee extensor angle- San Diego, USA. He is currently a Professor at the
torque relationship?’’ Exp. Physiol., vol. 102, no. 8, pp. 962–973, Faculty of Human Sciences, Kanagawa University.
Aug. 2017. He has authored or coauthored more than 50 jour-
[40] R. Merletti, D. Farina, and M. Gazzoni, ‘‘The linear electrode array: A nal articles and 120 conference papers/abstracts.
useful tool with many applications,’’ J. Electromyogr. Kinesiol., vol. 13,
His research interests include signal processing of
no. 1, pp. 37–47, 2003.
electromyography and its applications in move-
[41] M. M. Lowery, C. L. Vaughan, P. J. Nolan, and M. J. O’Malley, ‘‘Spectral
compression of the electromyographic signal due to decreasing muscle ment and human biomechanics and physiology.
fiber conduction velocity,’’ IEEE Trans. Rehabil. Eng., vol. 8, no. 3,
pp. 353–361, Sep. 2000.
[42] J. Rodriguez-Falces, D. Neyroud, and N. Place, ‘‘Influence of inter-
electrode distance, contraction type, and muscle on the relationship SHIMPEI KUBO received the M.A. degree
between the sEMG power spectrum and contraction force,’’ Eur. J. Appl.
from The Open University of Japan, Japan,
Physiol., vol. 115, no. 3, pp. 627–638, Mar. 2015.
in 2016. From 2018 to 2020, he was a Visiting
[43] C. J. De Luca, ‘‘Spectral compression of the EMG signal as an index of
Researcher with Head Office for Information Sys-
muscle fatigue,’’ in Neuromuscular Fatigue, A. J. Sargeant, D. Kernell,
Eds. Amsterdam, The Netherlands: Royal Netherlands Academy of Art tems and Cybersecurity, RIKEN. He is currently
and Sciences, 1992, pp. 44–51. a Research Associate with the Faculty of Human
[44] M. M. Lowery, N. S. Stoykov, and T. A. Kuiken, ‘‘Independence Science, Kanagawa University. His research inter-
of myoelectric control signals examined using a surface EMG ests include signal processing of electromyogra-
model,’’ IEEE Trans. Biomed. Eng., vol. 50, no. 6, pp. 789–793, phy and biomechanics in human movement.
Jun. 2003.

VOLUME 11, 2023 6403