Epidermal Sensing of Muscle Compensation Pei-Shin Hwang∗ Yi-Ting Wu∗ Polly Huang Department of Electrical Engineering, Department of Electrical Engineering, Department of Electrical Engineering, National Taiwan University National Taiwan University National Taiwan University b06901013@ntu.edu.tw b07901095@ntu.edu.tw pollyhuang@ntu.edu.tw ABSTRACT Towards effective and injury-free body building, we seek to explore the use of Electromyography (EMG) sensor for the detection of muscle compensation, a phenomenon often observed in physical exercises due to malcompliance or fatigue. KEYWORDS EMG, Muscle Compensation, Exercise Compliance Figure 1: Malcompliance in dumbbell bicep curl. 1 MOTIVATING APPLICATION Weight training is one of the most popular modern exercises to build muscle strength and resistance. It benefits not only the physical but 2 RELATED WORK also mental health, where stronger muscles allow body movement The surface EMG (sEMG) refers to the sensing of biopotentials at a higher caliber, and being in shape raises also self-confidence extracted from the electrodes on skin surface, and it has been used and social attention. Muscle building appears straightforward, in- widely for sports analysis, physical rehabilitation and as means volving lifting, pulling, or pressing of weights, and more repetitions of human-computer interface. In particular, Ke Ma et al. [5] pro- or heavier weights result in more robust muscles. The dilemma is posed an sEMG-based trunk compensation detection (sEMG-bTCD) however that incorrect postures or lifting excessive weights might mechanism, which was validated by the sEMG signals collected compromise the training effectiveness and, in the worst case, lead from nine superficial trunk muscles of the stroke participants dur- to injuries [6]. One common cause of injuries is muscle compensa- ing their rehabilitation therapy. Liu et al. [4] devised a miniature, tion – the phenomenon in which one overuses muscles that are not mobile EMG patch to monitor muscle fatigue during isotonic con- subject to training when the target muscles are exhausted and can traction. The system was shown capable of computing the median no longer complete the training independently. Figure 1 shows frequency of sEMG signal in real time and displayed the level of how the ill positioned elbow may activate the upper trapezius to muscle fatigue through a smartphone APP. Moreover, Biagetti et al. compensate the load supposedly for the bicep. Given the difficulty [2] presented a wireless system for sEMG and accelerometer signal spotting muscle compensation with bare eyes, these types of exer- acquisition to detect, monitor and recognize the human activity cise malcompliance often go unnoticed. being performed. To this end, we find opportunities in utilizing electromyography (EMG) signals collected from widespread body surface to assess 3 SYSTEM PROTOTYPE muscle compensation. EMG signaling is a common and effective method to evaluate muscle performance. Past works have focused Our hardware consists of a microcontroller, two muscle sensors, the use of EMG signals from a few targeted muscles and in the and a Bluetooth module. We use an Arduino Uno 1 as the micro- laboratory settings [1]. With the rise of multi-point epidermal EMG controller for the prototype. The EMG signal is extracted from sensors that are soft and bendable [3], it is no longer far fetched to Myoware 3-lead muscle sensors 2 . We choose this particular muscle observe muscle interactions in everyday training. As a motivating sensor because it provides not only raw EMG signals but also recti- application and a feasibility study, we share in this abstract a pre- fied and integrated EMG signals, which works well with Arduino’s liminary investigation of a wearable IoT that might grow capable analog-to-digital converter. The gain of signal rectification is ad- of identifying specific patterns between EMG signals and issuing justable on-board. We also connect an additional HC-06 Bluetooth feedback in sign of muscle compensation. module 3 for data transmission. ∗ Both authors contributed equally to this research. 4 PRELIMINARY RESULT A small number of participants are invited to the prototype trial. Copyright 2021 for this paper by its authors. Use permitted under Creative Commons As depicted in Figure 3, each participant is instructed to perform License Attribution 4.0 International (CC BY 4.0). the bicep curl with dumbbells of specific weights. Within the par- ticipants’ capabilities, 5 weights, 2 kilograms apart are lifted. The 1 https://store.arduino.cc/usa/arduino-uno-rev3 2 http://www.advancertechnologies.com/p/myoware.html 3 https://components101.com/wireless/hc-06-bluetooth-module-pinout-datasheet CHIIoT 1, February 17, 2021, Delft, The Netherlands Hwang and Wu, et al. Figure 2: Bicep and trapezius muscle signals in 6kg vs. 8kg case. Figure 3: Prototype and experiment setting. Figure 4: Normalized peak values of the trapezius muscles. participants perform each set of exercise with constant speed within a 20 second duration. Figure 2 shows the rectified sEMG signals (i.e, the slope) of signal ratios as well as the sheer amplitudes in of one participant lifting 6kg and 8kg weights. The blue lines dis- characterizing types of muscle compensation. play the signals of the bicep muscles, on which five clusters that correspond to five sets of bicep curl are clearly visible. The red 5 WORK IN PROGRESS lines display the signals of upper trapezius muscles. As the partici- We are currently pursuing work in two dimensions. First is comput- pant lifts a 6kg weight, this signal remains at a relatively low and erization of the process to quantify the level of muscle compensa- stable level. On the contrary, when lifting an 8kg weight, one can tion and to characterize the risk of posture deformation, excessive observe apparent peaks in the signal, which implies the substantial weight, or fatigue. In the meantime, we seek to stream line the com- use of the trapezius muscles. This showcases the plausibility of an putation tasks on board the epidermal wearable to enable real-time EMG-based wearable for muscle compensation assessment. user feedback. This may require an upgrade to Raspberry Pi or Furthermore, Figure 4 shows the normalized peak values of processors of higher computation capability. trapezius muscles of eight different participants. The peak EMG amplitude signals of the trapezius muscles in each exercise clus- REFERENCES ter are divided by the mean EMG amplitude signal during rest. [1] V. Andersen, M. Fimland, Dag-André Mo, V. M. Iversen, T. M. Larsen, Fredrik As the weight increases, this normalized ratio increases as well. Solheim, and A. Saeterbakken. 2019. Electromyographic comparison of the barbell deadlift using constant versus variable resistance in healthy, trained men. PLoS Generally, a surge of ratio occurs at 6kg and above, whereas in- ONE 14 (2019). dividual differences can be observed. (1) The yellow and sky blue [2] Giorgio Biagetti, Paolo Crippa, Laura Falaschetti, and Simone Orcioni. 2018. Hu- lines, which increase slightly and stably, are collected from athletic man activity monitoring system based on wearable sEMG and accelerometer wireless sensor nodes. BioMedical Engineering OnLine 17 (11 2018). participants. (2) The brown and grey lines, which show a surge at [3] Jun-An Chen, Cynthia Yun-Shin Liu, Hong Jhih Chen, Hsin-Yuan Chen, Hsea- 4kg but remain flat through 6kg, are the only data collected from Ching Hsueh, Yu-Chia Huang, Koji Yatani, Hao-Hua Chu, and Polly Huang. 2019. Position: Smart KT Tape – A Bendable Wearable System for Muscle Fatigue Sensing. male participants. (3) The dark blue line shows a unique pattern– ACM WearSys ’19, Seoul, Korea. not increasing much through 8kg but shooting up at 10kg, which [4] Shing-Hong Liu, Chuan-Bi Lin, Ying Chen, Wenxi Chen, Tai-Shen Huang, and Chi- suggests that the participant might have no strength left at all for Yueh Hsu. 2019. An EMG Patch for the Real-Time Monitoring of Muscle-Fatigue Conditions During Exercise. Sensors 19 (07 2019), 3108. lifting the 10kg weight. These distinctions imply the distinctive [5] Ke Ma, Yan Chen, Xiaoya Zhang, Haiqing Zheng, Song Yu, Cai Siqi, and Longhan physical abilities of individual participants, thus various extents of Xie. 2019. sEMG-Based Trunk Compensation Detection in Rehabilitation Training. muscle compensation. Frontiers in Neuroscience 13 (11 2019), 1250. [6] Tom Erik Solstad, Vidar Andersen, Matthew Shaw, Mogstad Hoel, Andreas Von- As for further utilization of the results, we find that there is no heim, and Atle Saeterbakken. 2020. A Comparison of Muscle Activation between binary classification but rather multiple levels of muscle compen- Barbell Bench Press and Dumbbell Flyes in Resistance-Trained Males. Journal of sports science & medicine 19 (08 2020), 645–651. sation. It is also more sensible to take into account the change of