Abstract – The aim of this scenario was to build an Electromyography (EMG) sensor capable of measuring muscle activity, and demonstrating muscle movement and strength through motion, audio and visual outputs. This report introduces the theory behind the EMG sensor, describes the methodologies applied throughout this scenario, presents the results from the work and summarises the meanings of the outputs. The purpose of this paper is to illustrate the various methodologies and algorithms for EMG signal analysis to provide efficient and effective ways of understanding the signal and its nature. We further point up some of the hardware implementations using EMG focusing on applications related to prosthetic hand control, grasp recognition, and human computer interaction. A comparison study is also given to show performance of various EMG signal analysis methods. This paper provides researchers a good understanding of EMG signal and its analysis procedures. This knowledge will help them develop more powerful, flexible, and efficient applications. An optimized circuit for processing of EMG signals has been designed and presented in this paper. This circuit acquires EMG signals from surface of the skin using bipolar electrodes and enables the amputee to control the prosthetic hand in an efficient manner. An optimised circuit for processing of EMG signals has been designed and presented in this report.
1. Introduction
The nervous system in our body controls muscle activity. During muscle contraction and relaxation, the cells are electrically activated thereby generating an electric potential. Electromyography (EMG) is an electrodiagnostic medicine technique for evaluating and recording the electrical activity produced by muscles, such as clenching a fist for example. EMG is performed using an instrument called an electromyograph to produce a record called an electromyogram. The signals can be analysed to detect medical abnormalities, activation level, or recruitment order, or to analyse the biomechanics of human or animal movement.
The EMG is further used in many clinical and biomedical applications like assessing low back pain, neuromuscular diseases and disorders of motor control. It can also be used for other practical purposes, as to control electronic devices such as assisting robotic hands and mobile phones. It is a daunting challenge for robotic designers to emulate human movement in the applications of prosthetics and robot planning due to the complex neurophysiology of human body. This is where the EMG can really become an important tool to understand and replicate the neurophysiology of the human body.
2. Theory
Each muscle cell generates electric potential. The EMG potential is in range between 50 μV to 30 mV, depending on the muscle type and conditions during the observation process. The bulk of the EMG signal can be detected in the 0-500 Hz frequency band; it is quite a complex signal, and to understand its formation it is important to appreciate where the signal originates.
2.1 EMG Signal Conditioning
The signal generated is normally a function of time and is desirable in terms of its amplitude, frequency and phase.
Fig. 1 Schematic showing conventional myoelectric signal processing [9]
2.1.1 Electrodes
The circuit designed acquires EMG Signals from the surface of the skin using bipolar electrodes. These electrodes convert ionic current from the muscles to electric current, inadvertently producing two types of transducer noise. One of them is caused due to the difference in the impedance between the electrode surface and that of the skin, and the alternating voltage, is generated due to the fluctuations in impedance between the skin and the electrode.
To remove the effect of this noise, we require the EMG to undergo the process of amplification, band limiting & rectification before it can be fed to the Arduino and subsequently to audiovisual outputs.
2.1.2 Differential Amplifier
The first step in processing the signal, the Differential Amplifier rejects the common elements by subtracting the two signal inputs from each electrode and amplifies the difference obtained.
2.1.3 Common Mode Rejection
Common Mode Rejection Ratio (CMRR) is the measure of the ability of a differential circuit to suppress the signals, which are common in both the inputs and are same in their phase and frequency with respect to the reference or GND voltage. As a guiding principle possibly the highest common mode rejection ratio (CMRR) is desirable [10]. Expression of CMRR is carried out in logarithmic form:
CMRR = 20 Log (Vout / Vin) dB (1)
There is always trade off between the gain and CMRR. An increase in the CMRR leads to a decrease in the gain and vice-verca. Another way to increase common mode rejection is to use a Differential amplifier with higher CMRR. The two signals, will therefore, have their similar values suppressed. The reason for a similar value of these signals can be power sources, electromagnetic devices and EMG signals due to the remote muscles.
2.1.4 Noise & quality of EMG Signal
EMG signals acquire noise while traveling through different tissues. This is because of:
Inherent noise in electronic equipments: This noise can never be eliminated, but can be reduced using high quality electronic components.
Inherent instability of the signal: Caused by the random nature of the EMG Signal. It is essential to remove this noise and its done through processes including pre-amplification, bandlimitting & rectification – only after which the signal becomes usable & is an accurate illustration.
Ambient noise: Caused because of electromagnetic radiation. The surfaces of our bodies are constantly inundated with electric-magnetic radiation and it is impossible to avoid exposure to it on the surface of earth.
Motion artifact: Causes irregularities in the data. This is because of the electrode interface and the electrode cable. Proper circuit design and set-up can reduce its effect.
Quality of EMG signal, is dependent generally on the signal conditioning processes, and primarily on the features of the pre-amplification process. To maximise the quality of the EMG signal
Keep the Signal-to-noise (SNR) ratio as high as possible i.e it should contain the highest amount of information lowest amount of noise possible.
Avoid unnecessary filtering and notch filters to ensure signal distortion is minimal.
2.1.5 Band Pass Filtering
A high pass filter is used to remove the low frequency (typically less than 10 Hz) components produced by Motion artifacts. Similarly in order to remove the effects of aliasing high frequency components the signal is passed through a Low pass filter. The use of notch filter was considered to remove Ambient Noise, however, the EMG signal has a significant proportion of signal contributions around this frequency, and would lead to heavy signal distortion. This is why a Band Pass Filter was used in place of a Notch Filter.
3. Methodology
The circuit in Fig. 2 has been developed keeping in view all the required criteria. The designed circuit was initially simulated on MultiSim version 12.0 and on the basis of the outputs obtained, the circuit was chosen as the circuit to build the EMG Sensor.
3.1 Circuit Design
Fig. 2. Diagram for the circuit designed and implemented for this scenario.
Instrumentation Amplifier with a Driven Right Leg (DRL): The intrinsic EMG signals are amplified by the amplifier having a CMRR of 46 dB and this circuit uses a DRL circuit for further enhancement of the CMRR. The DRL circuit feedbacks the common mode voltage back into the human body, and thus further raises the CMRR. Noise which is produced by the power lines is common in both readings, and in order to remove it, a high CMRR is essential.
For this scenario, LMC660CN has been used as an Instrumentation Amplifier. It is a low power instrumentation amplifier and has high accuracy. Additionally, it provides high CMRR (120 dB when Gain (G) is greater than 100). Its gain may be set from 1 to 1000 V/V. It has Resistance R = 2.2 kΩ, thereby making the input gain of the circuit:
G = 1+(49.4 kΩ / 2.2 kΩ) = 23.45 V/V (2)
2. Band Pass Filter: The output of the Instrumentation Amplifier is then passed through a band pass filter designed for the frequencies of 20 Hz to 650 Hz. The first one is a first order High Pass (HP) Filter having a limiting frequency of 20 Hz, it helps attenuate the small frequencies caused by the motion artifacts in between the skin and the electrodes. The Gain of the HP Filter is set at 12.2 V/V.
The signal then passes through the second stage of band pass filter, it is a second order Low Pass (LP) Filter having a limiting frequency of 650 Hz. The limit of the LP filter is set at 650 Hz since a significant portion of signals are found between 50 Hz – 350 Hz. Therefore, by keeping the cutoff frequency at 650 Hz, optimum performance is ensured since there is no signal loss. The Gain of the LP Filter is kept at 10 V/V.
Fig. 3. Schematic diagram of the Band Pass Filter used [12].
4. Arduino: Chosen as the micro-controller for this scenario due to its structure. It includes a 5 V regulator, a burner, an oscillator, a micro-controller, serial communication interface, Light emitting diode (LED), headers for the connections and relatively inexpensive.
The output from the filter circuit is connected as the analog input to the Arduino. The Arduino constantly reads from the input and maps the result to a range from 0 to the 8 LEDs present in the Bar LED. It then loops over the LED array – if the array element’s index is lesser than the LED level, its corresponding pins turn on, thereby turning on all the LEDs bellow the level. Simultaneously, all the pins higher than the LED level are turned off. This process is repeated for all the outputs used.
The outputs used for the EMG sensor were: Bar LED, LED Dimmer, Speaker and a Motor. Reference [5] accurately depicts how the Bar LED worked as the muscles contracted.
Fig. 4. Output circuit illustrating how the Bar LED is connected using the Arduino.
3.2 Input vs Output
During the simulation on MultiSim, the values obtained for the inputs and outputs are as follows and the salient features are shown in figure 5.
Input at V1 – 2 mV
Input at V2 – 5 mV
Output – 3.456 V
Fig. 5. Output graph produced by the circuit in fig 2.
4. Physical Implementations & Results
The circuit was then designed using DIY Layout creator, as shown in Figure 6.
The circuit designed was later soldered on to the Veroboard and tested using the Ag-AgCl electrodes placed at a distance of 20 mm from each other, as shown in Figure 7. It produced the expected output of 3.456 V.
Fig. 6. Circuit designed using the DIY Layout creator.
Fig. 7. Electrode arrangement. Note: the EMG signals (m1 and m2) are considered to be different, whereas the noise (n) is similar.
The first few tests were done on the arms of a human being. The output was produced on the oscilloscope. Using the scope the data was then recorded and using LabView 2017 the graphical representations were produced. The test with the graphical representation of the unambiguous results obtained are shown in Figure 8.
Fig.8. Output graph of EMG signal obtained after placing electrodes
on the arm of a human while processing the signal through the circuit.
Following the tests from the arms, the next few series of tests were done on fingers of a human being. The idea behind these tests were to a) prove the EMG sensor built
can produce outputs by different (relatively weaker) muscles and b) compare the strength of each of the fingers in order to identify the strongest finger i.e finger produced the highest variation in amplitude in every hand.
Fig. 9. Output graph of EMG signal obtained after placing electrodes
on different fingers of a human while processing the signal through the circuit.
Every subject was asked to exert their maximum force for every test and every finger. It is realised that this is not an optimum method, however, it is the most straightforward and prompt way to ensure the tests were fair.
5. Improvements
The disadvantage of the circuit built is that the CMRR of Instrumentation Amplifier is not high enough and can be further increased using alternative techniques. These alternate solutions could help keep the dimensions of the circuit smaller as well.
The circuit has been developed keeping in view the circuit designed for the project and drawbacks of proposed circuits in the research papers [11] – [15]. The circuit was initially simulated on MultiSim version 12.0 and on the basis of the outputs obtained, this circuit seems to be have been the optimum design to build the EMG Sensor.
The circuit has been simulated on Proteus 7.6 version, salient features of the circuit are as follows:
Fig. 10. Output graph produced by the circuit in fig. 2.
1. Diodes for Safety of Circuit: Two diodes have been employed in the circuit at the input to provide safety to the circuitry and the device. Usually the voltage across a muscle is of the tune of mV. A diode will switch on at 0.7 V. If ever more than 0.7 V is present across the inputs (may be due to short circuit), a current passage with much less resistance than our body will be provided by the diodes. The resistance of our body is approximately 300Ω and 1500Ω, hence when a diode is ON, it will offer very less resistance as compared to our body, hence current will pass through the circuit.
2. Instrument Amplifier: AD620 is a differential amplifier, which removes ambient noise. Pre-amplification is essential since the EMG signal is most influenced by it. Pre-amplification stage has certain parameters which are important [16], and they characterise the pre-amplification process, these are:
High common mode rejection ratio
A very high input impedance
Short distance to the signal source
Strong DC signal suppression
The AD620 is used as the preamplifier since it has a low input voltage noise. It is an Instrumentation Amplifier which has a high accuracy of the range of 40 ppm maximum nonlinearity, it has low offset voltage of 50 μV max and an offset drift of 0.6 μV/°C max. Moreover, less input bias current, low noise and low power (1.3 mA max supply) of the AD620 made it most suitable for this battery powered project. Characteristics of AD620 are as follows:
CMRR – 120 to 130 with Gain 1 to 1000
Input Impedance – 10 GΩ 2 pF
Fig. 11. Graph of CMRR Vs Frequency of AD620
Fig. 12. Output graph produced by the circuit in fig. 10.
3. Band Pass Filter: Same band pass filter is being used as in the project [See Section 3.1].
4. The values of output and input given by the circuit during testing are as under:
Input at V1 – 2 mV
Input at V2 – 5 mV
Output – 3.499 V.
6. Conclusion
The obtained results clearly proved high quality and well noise cancellation of dedicated amplifier with better measurement results in comparison to other conventionally used devices, and it is indicated that mostly the area of active electrode is very important factor. The overall size of electrodes is very critical in signal-to-noise ratio. To improve quality of output signal we proposed digital narrowband filters (mostly in range of 55-95 Hz), but for lab testing purposes we have considered wider frequency range: 2 Hz – 300 Hz.
In the future work a smart system will be further enhanced with focus on its continuous miniaturisation and investigation of other electrode surface materials (Au, Pt, Al, ZnO, ITO) with accumulation of the number of real-time wireless connected probes. More advanced methods of on-probe digital pre-processing and noise filtering will be investigated, leading to development of more simple two-electrode probe. We will also investigate some new geometrical configurations of electrodes (Interdigital array of microelectrodes, etc.).
7. Future developments
1. Contactless sensors: The traditional contact electrodes are not very useful for textile integration: good skin contact with conductive gel is needed for this type of electrodes. In the literature [11-13], contactless sensors for ECG are proposed. These sensors detect an electric displacement current by means of a capacitive coupling to the body instead of detecting a Nernstian current; therefore, they require no electrical contact with the skin. The avoidance of direct skin contact gives the opportunity to wear the vest above other clothing. The contactless sensors consist of a conducting plate covered by an insulating layer, forming a parallel plate capacitor with the skin [14].
A typical setup for measuring sEMG is a bipolar setup. Because of the capacitive coupling, the impedance of the electrode is very high. Therefore an impedance converter is placed directly on top of the electrode. An extra reference electrode is also used to reduce environmental noise.
In contraction, the contactless sensor shows the same characteristics as the contact electrode. In rest however, the level of the measurement with the contactless sensor is about 20dBV higher than with the traditional contact electrode. This indicates a lower signal-to-noise ratio for the newly developed sensors. It must be mentioned that it is not possible to do the two measurements at the same time on the same position.
EMG can be used to sense isometric muscular activity (type of muscular activity that does not translate into movement). This feature makes it possible to define a class of subtle motionless gestures to control interface without being noticed and without disrupting the surrounding environment. The device for this purpose includes a high input impedance amplifier connected to electrodes, an anti-aliasing filter, a micro-controller to sample and process the EMG signal, and a Bluetooth communication module to transmit the processing results. When activation is detected, the controller sends a signal wirelessly to the main wearable processing unit, such as a mobile phone or PDA. Using EMG, the user can react to the cues in a subtle way, without disrupting their environment and without using their hands on the interface. The EMG controller does not occupy the user’s hands, and does not require them to operate it; hence it is “hands free” (74).
The US Air Force and other military branches increasingly use unmanned vehicles for surveillance missions. One way to control these systems from the field is a wearable cockpit. One could use a wearable computer with a wireless link and display goggles, and then employs EMG-based gestures to manipulate the switches control inputs. A space-based application could let astronauts’ type into a computer despite being restricted by a spacesuit. If a depressurisation accident occurred on a long-term space mission and astronauts needed to access onboard computers, they could use EMG electrodes in their spacesuits to replicate a computer interface (76).
Unvoiced speech recognition
Mime Speech Recognition – recognizes speech by observing the muscle associated with speech. It is not based on voice signals but EMG. It will realize unvoiced communication, which is a new communication style. Because voice signals are not used, it can be applied in noisy environments; it can support people without vocal cords and aphasics (77).
REFERENCES
[1] S. Ali and I. Javaid, “Optimized Circuit for EMG Signal Processing.” National University of Sciences and Technology, College of E&ME, Pakistan, 2012.
[2] M.B.I. Raez, M.S. Hussain, and F. Mohd-Yasin, “Techniques of EMG signal analysis: detection, processing, classification and applications.” Faculty of Engineering, Multimedia University, Malaysia, 2006.
[3] G. Skov-Madsen and N. Rijkhoff, “Long-term stability of EMG recordings using a wireless EMG electrode.” Center for Sensory-Motor Interaction (SMI), Denmark, 2008.
[4] A.B. Jani, R. Bagree and A. K. Roy, “Design of a Low-power, Low-cost ECG & EMG Sensor for Wearable Biometric and Medical Application,” Dhirubhai Ambani Institute of Information and Communication Technology, India, 2017.
[5] Contactless EMG Sensors for continuous monitoring of muscle activity to prevent musculoskeletal disorders
[6] Gerdle B, Karlsson S, Day S, Djupsjöbacka M, “Acquisition, Processing and Analysis of the Surface Electromyogram. Modern Techniques in Neuroscience.” Chapter 26: 705-755. Eds: Windhorst U and Johansson H. Springer Verlag, Berlin, 1999.
[7] Burke, M.J., Gleeson, D.T. (2000), “A micropower dry-electrode ECG preamplifier”, IEEE Transactions on Biomedical Engineering, v. 47, n. 2, p. 155-162.
[8] Important Factors in Surface EMG Measurement By Dr. Scott Day Bortec Biomedical Ltd
Shahid S. Higher Order Statistics Techniques Applied to EMG Signal Analysis and Characterization. Ph.D. thesis, University of Limerick; Ireland, 2004.
[10] Basmajian JV, de Luca CJ. Muscles Alive – The Functions Revealed by Electromyography. The Williams & Wilkins Company; Baltimore, 1985.
[11] Cram JR, Kasman GS, Holtz J. Introduction to Surface Electromyography. Aspen Publishers Inc.; Gaithersburg, Maryland, 1998.
[12] Paulo Roberto Stefani Sanches André Frotta Müller, Luigi Carro Altamiro Amadeu Susin and Percy Nohama, “Analog reconfigurable technologies for EMG signal processing”
[13] Duchene J, Hogrel J. A Model of EMG Generation. IEEE Transactions on Biomedical Engineering. 2000;47(2):192–201.
[14] Ricamato AL, Absher RG, Moffroid MT, Tranowski JP. A time-frequency approach to evaluate electromyographic recordings. Proceedings of Fifth Annual IEEE Symposium on Computer- Based Medical Systems 1992; pp. 520-527.
[15] Manabe H, Hiraiwa A, Sugimura T. Unvoiced Speech Recognition using EMG-Mime Speech Recognition. Conference on Human Factors in Computing Systems 2003; pp. 794-795.
[16] de Luca CJ. The use of surface electromyography in biomechanics. J Appl Biomech. 1997;13:135–163.
[17] Nikias CL, Petropulu AP. Higher-Spectral Analysis: A Nonlinear Signal Processing Framework. Prentice Hall; New Jersey, 1993.
https://www.arduino.cc/en/Tutorial/BarGraph
De Luca, C.J. Electromyography. Encyclopedia of Medical Devices and Instrumentation, (John G. Webster, Ed.) John Wiley Publisher, 98-109, 2006.
Design of EMG wireless sensor system
Erik Vavrinský1, Martin Daříček2,1, Martin Donoval2, Karol Rendek1, František Horínek2,1, Martin Horniak2,1, Daniel Donoval1
1Department of Microelectronics, Slovak University of Technology in Bratislava Ilkovičova 3, SK-812 19 Bratislava, Slovakia 2Nanodesign s.r.o. Drotárska 19A, SK-811 04 Bratislava, Slovakia E-mail: erik.vavrinsky@stuba.sk
Intelligent Bio-Detector – Scientific Figure on ResearchGate. Available from: https://www.researchgate.net/LabVIEW-output-of-the-EMG-sensor_fig32_311936100 [accessed 28 Nov, 2018]
2018-11-28-1543419013
...(download the rest of the essay above)