Design of sEMG-Detecting Circuit for EMG-Bridge Xiao-Bin Chen, Yu-Xuan Zhou, Hai-Peng Wang, Xiao-Ying Lü, Zhi-Gong Wang, Senior Member, IEEE * Abstract— A surface electromyography (sEMG) signal typically results from the electrical activities of many muscle fibers, and can be utilized as a signal source in prostheses due to its abundance of movement information. This paper proposes an sEMG--detection circuit for the acquisition of the controlling signal in EMG-Bridge (EMGB) systems. The detection circuit mainly comprises a preamplifier, a driven right leg (DRL) circuit, a high-pass filter (HPF), a low-pass filter (LPF), and a gain adjustable amplifying circuit. The common-mode rejection ratio (CMRR) of the circuit is higher than 120 dB, the input impedance is greater than 100 MΩ, the passband range is 20~450 Hz, and the frequency attenuation in stopband is not less than 120dB/dec. I. INTRODUCTION Limb paralysis is a severe loss of muscle function often caused by stroke and spinal cord injury (SCI). For example, nearly 80% of stroke survivors suffer from limb paralysis due to the brain damages , an increasing problem with aging (b) High-Pass (a) Preamplifier population. The recovery of motor function after stroke and SCI has been a difficult problem in the field of rehabilitation .The electromyography bridge (EMGB) proposed by our researching team  is an advanced technique to help patients rebuild their limb function. The principle is using the real-time sEMG signals of a healthy limb to control functional electrical stimulation (FES) on the paralytic limb, allowing voluntary and flexible rehabilitation. Because sEMG signals are normally the sum of the electrical contributions of the active motor units (MUs) as detected by electrodes placed on the skin superficial to the muscle, they have the following two main characteristics: 1. Small amplitudes: Most sEMG signals are relatively weak: their amplitude range is 0.01~5 mV, with peak to peak and root mean square values of less than 6 mV and, 1.5 mV, respectively. 2. Low frequency: The frequency range is 0~500 Hz, with most of the energy concentrated between 50~150 Hz . Therefore, a key component of the EMGB system is the sEMG Filter (HPF) (c) Low-Pass Filter (LPF) OPA132 E L E C T R O D E R1 U1 INA128 C2 U6 R7 C3 U7 R9 C5 C4 OPA132 U2 U3 C1 R3 OPA132 R8 U8 C6 R11 R12 R13 OPA132 R14 C7 R10 U9 C8 OPA132 C9 R2 OPA132 R16 R5 U5 R18 OPA130 R21 U10 U11 U12 U4 R15 R4 R6 OPA132 R19 R17 R22 OPA132 OPA132 (e) Gain (d) Driven Right Leg (DRL) Circuit OPA132 Adjustable Amplifier R20 TLC2252 (f) Level C10 A D C Lift Circuit Fig.1 The schematic of the sEMG-signal detecting circuit. * This work was supported by National Natural Science Foundation of China (No. 61534003) and the Jiangsu Province Science and Technology Project of China (No. BE2016738) Xiao-Bin Chen, is with State Key Laboratory of Bioelectronics, Southeast University, Nanjing, P. R. China. ([email protected]) Yu-Xuan Zhou, is with State Key Laboratory of Bioelectronics, Southeast University, Nanjing, P. R. China. Hai-Peng Wang, is with the Institute of RF- & OE-ICs, Southeast University, Nanjing, P. R. China. Xiao-Ying Lü, corresponding author, is the vice director of State Key Laboratory of Bioelectronics, Southeast University, Nanjing, 210096, P.R. China (phone: 025-83793430; fax: 025-83793430;e-mail: [email protected]). Zhi-Gong Wang, corresponding author, is the director of the Institute of RF- & OE-ICs, Southeast University, Nanjing, 210096, P. R. China (phone: 025-83792882; fax: 025-83792882; e-mail: [email protected]). Xiao-Ying Lü and Zhi-Gong Wang are also with the Co-innovation Center of Neuroregeneration, Nantong University, 226001 Nantong, China. 978-1-5090-2809-2/17/$31.00 ©2017 IEEE detection circuit. In this work, we present a novel design with high performance in terms of the correct decoupling of the sEMG signals and implementation of the channel mapping. II. SEMG DETECTING CIRCUIT Fig. 2 shows the block diagram of the sEMG detection circuit, and the detailed schematic we designed is presented in Fig. 1. The circuit includes five parts: a preamplifier, a DRL circuit, a high-pass filter (HPF), a gain adjustable amplifier, and a low-pass filter (LPF). 382 Fig.2 The block diagram of the sEMG-detecting system. A. Preamplifier In order to improve the anti- interference performance of the circuit, the preamplifier and the sEMG detecting electrodes should be located as close as possible. Thus, we combined the sEMG electrodes and preamplifier as a sensor by packaging them in one rectangular plastic box with a size of 20×35×7 mm3. There are two kinds of sEMG electrodes: dry metal electrodes  and wet gel electrodes . The wet gel electrodes are mostly disposable electrodes, whereas the dry metal electrode can be repeatedly used. Therefore, we used two filamentary silver electrodes with cylindrical shapes of s of 10-mm length, 1-mm diameter and 10-mm separation interval(Fig. 3) Fig.3 Illustration of the sEMG sensor with placement on the forearm (not to scale). common mode interferences on the skin and match the body's reference potential to the circuit's reference voltage. The input of the DRL can be obtained from the node of In-Amp’s gain resistors R1 and R2 as shown in Fig.2(a). C. Filtering and Gain Circuits The task of the sEMG detecting circuit is to extract clean EMG signals from a noisy background. In this work, we used an In-Amp as the first-stage amplifier (preamplifier) to achieve significant rejection of the common mode noises and interferences. Because the frequency band of sEMG signals ranges from 20 Hz to 500 Hz and most of the energy is concentrated between 50~150 Hz, a high-pass filter and a low-pass filter (Fig. 1 (b) and Fig. 1 (c)) were designed to suppress the spectral components outside of that range. Two Sallen-Key filters are used as the basic unit for the fourth-order high pass and low pass filter circuits due to their high input impedance and low output impedance characteristics and lower requirements for operational amplifier and precision resistors and capacitors. The high pass and low pass filter circuits compose a pass band ranging from 20 Hz to 450 Hz, which can help to remove baseline wander and motion artifacts. The Sallen-Key filter circuits are shown in Fig. 4. The signal amplitude detected by the filamentary silver electrodes is very low, typically in the mV-range. In this work, a preamplifier (Fig.2 a) has been designed. It consists of an instrument amplifier and a voltage follower circuit. The gain was fixed using an INA128 instrumentation amplifier (In-Amp) and is calculated as： with RG = R1 + R2 ( R1 = R2) . The fixed gain in this stage provides the lowest value within our desired range, and higher gains are provided by the gain adjustable amplifying circuit. B. Driven Right Leg (DRL) Circuit The sEMG-detecting circuit can detect small EMG signals. Unfortunately, the patient's body also receives electric interferences from the environment, especially at the 50/60-Hz frequency of electrical power lines. The use of DRL has been suggested to eliminate the interferences by active cancellation . In this work, a DRL circuit is implemented to reduce the Fig. 4 Sallen-Key low-pass filter (a) and high-pass filter (b). In this work, the gain circuit is difficult to realize in the first stage due to large DC offset without AC coupling. Our design used a low gain amplifier in the first stage and then a high gain in the final stage. Therefore, we have chosen a two-stage gain adjustable amplifier (Fig.2 (e)) where the gain can be increased beyond 100. III. RESULTS AND DISCUSSION A. sEMG Detection Circuit In the fabrication of the sEMG detecting circuit, the electrodes and the preamplifier were packaged into one sensor, whereas the other circuits were made into an sEMG-signal processing unit. The components of the sEMG detection circuit are shown in Fig. 5. 383 Fig. 5 The components of sEMG detecting circuit B. Experiments and Results In order to verify the frequency selectivity of the filters, its amplitude-vs.-frequency characteristic was measured. The input signal is a sinusoidal signal with amplitude of 1 V. The signal frequency ranged from 0 Hz to 40 Hz and from 300 Hz to 4 kHz for the high-pass and low-pass filters, respectively. The transfer functions of the filters are shown in Fig. 6. For the high-pass filter, the transfer function tapers off at 17 Hz. For the low-pass filter, the transfer function tapers off at 450 Hz . (b) Fig.6. Amplitude-vs.-frequency characteristic of the highpass filter (a) and low pass filter (b) (a) The experimental procedure of detecting sEMG signals of the extensor carpi ulnaris (ECU) using the proposed circuit are described by following protocol: 1. Use ethanol to clean the right lower arm of the subject. 2. Attach the sensor to the skin superficial to the extensor carpi ulnaris (ECU). 3. The subject performs repeated free wrist extensions. 4. A 0.5kg load is lifted, then the subject performed wrist extensions The output of the circuit during the wrist extension is shown in Fig. 7 Analysis revealed a high signal noise ratio (SNR) of amplitude. Accordingly, our circuit demonstrated reliable sEMG signal detection, and the signals meet the requirements as input for the EMGB system. 384  (a) (b) Fig.7 sEMG signal obtained during wrist extension exercise. IV. CONCLUSION We described the complete sEMG detection system and presented a general circuit topology of a sEMG-signal detecting circuit with detailed description of each component. The experiment demonstrated that our circuit has a signal detection capability with a high signal-to-noise ratio can be reliably used in an EMGB system. In our future work, the sensor can be further improved with emphasis on further miniaturization of the circuit and the investigation of using other electrode surface materials (e.g., gold and copper). A wireless interface between the PCB and the computer may also be included in future designs.. REFERENCES       C. D. Mathers, T. Boerma and D. M. Fat, “Global and regional causes of death”, British Medical Bulletin, 2009, 92(1): 7-32. https://www.christopherreeve.org/research/history-of-spinal-cord-rese arch Zhou, Yuxuan, et al. Surface myoelectric signals decoding using the continuous wavelet transform singularity detection." Bioelectronics and Bioinformatics (ISBB), 2011 International Symposium on. IEEE, 2011. Qiu W, Hua Y. Performance analysis of the subspace method for blind channel identification ☆[J]. Signal Processing, 1996, 50(1–2):71-81. Imtiaz, U., et al. "Design of a wireless miniature low cost EMG sensor using gold plated dry electrodes for biomechanics research."Mechatronics and Automation (ICMA), 2013 IEEE International Conference on IEEE, 2013:957-962. Vijayan, A. E., et al. "High SNR EMG Acquisition System for Biofeedback Applications." IEEE International Conference on 385 Computational Intelligence & Communication Technology 2015:359-363. Winter B B, Webster J G. Driven-right-leg circuit design.[J]. IEEE Transactions on Biomedical Engineering, 1983, 30(1):62-6.