基于表面肌电图的人体运动意图识别研究进展

doi:10.3969/j.issn.1006-9771.2021.05.013

《中国康复理论与实践》 ›› 2021, Vol. 27 ›› Issue (5): 595-603.doi: 10.3969/j.issn.1006-9771.2021.05.013

基于表面肌电图的人体运动意图识别研究进展

曹梦琳^1,^2,³,陈宇豪^1,^2,³,王珏^1,^2,³,刘天^1,^2,³()

^1.生物医学信息工程教育部重点实验室，西安交通大学生命科学与技术学院健康与康复科学研究所，陕西西安市 710049
^2.国家医疗保健器具工程技术研究中心，广东广州市 510500
^3.神经功能信息学与康复工程民政部重点实验室，陕西西安市 710049

收稿日期:2019-12-29 修回日期:2021-03-11 出版日期:2021-05-25 发布日期:2021-05-26
通讯作者: 刘天 E-mail:tianliu@xjtu.edu.cn
作者简介:曹梦琳(1997-)，女，汉族，山西晋城市人，硕士研究生，主要研究方向：基于运动意图预测的下肢主动神经康复方法研究|刘天(1983-)，男，汉族，河北人，副教授，主要研究方向：生物医学信号处理及康复工程。
基金资助:
陕西省自然科学基金面上项目(2018JM7080);中国博士后科学基金(64批)二等资助项目(2018M643672);中央高校基本科研业务费专项资金项目(xjh012019049)

Advance in Human Motion Intention Recognition Based on Surface Electromyography (review)

Meng-lin CAO^1,^2,³,Yu-hao CHEN^1,^2,³,Jue WANG^1,^2,³,Tian LIU^1,^2,³()

^1.The Key Laboratory of Biomedical Information Engineering of Ministry of Education, Institute of Health and Rehabilitaion Science, School of Life Science and Technology, Xi'an Jiaotong University, Xi'an, Shaanxi 710049, China
^2.National Engineering Research Center of Health Care and Medical Devices, Guangzhou, Guangdong 510500, China
^3.The Key Laboratory of Neuro-informatics & Rehabilitation Engineering of Ministry of Civil Affairs, Xi'an, Shaanxi 710049, China

Received:2019-12-29 Revised:2021-03-11 Published:2021-05-25 Online:2021-05-26
Contact: Tian LIU E-mail:tianliu@xjtu.edu.cn
Supported by:
Shannxi Natural Science Foundation (General)(2018JM7080);China Postdoctoral Science Foundation(2018M643672);Fundamental Research Funds for the Central Universities(xjh012019049)

摘要/Abstract

摘要： 目的

分析基于表面肌电图人体运动意图识别的研究方法和识别效果。

方法

检索PubMed、Web of Science、中国知网、万方数据、维普数据库建库至2020年12月文献，筛选基于表面肌电图的人体运动意图识别实验研究，提取相关数据，进行描述性分析。

结果

根据采用的方法，运动意图识别可分为3类：基于肌肉骨骼模型的运动意图识别、基于传统机器学习的运动意图识别和基于深度学习的运动意图识别。

结论

单一基于表面肌电信号的方法难以完全彻底地估计所有运动意图。开发精确和实时的人体运动意图识别方法仍有待进一步研究。

关键词: 康复机器人, 人机交互, 表面肌电信号, 运动意图识别, 综述

Abstract: Objective

To summarize the methods and results of human motion intention recognition based on the surface electromyography.

Methods

Literatures were retrieved and reviewed from the databases of PubMed, Web of Science, CNKI, Wanfang and VIP until December, 2020. The experimental researches about human motion intention recognition based on surface electromyography were summarized.

Results

The methods of motion intention recognition were divided into three models: musculoskeletal model, traditional machine learning model and deep learning model.

Conclusion

It is difficult to fully estimate human motion intention using surface electromyography in a single way. More researches are needed to develop more accurate and real-time human motion intention recognition methods.

Key words: rehabilitation robot, human-machine interaction, surface electromyography, motion intention recognition, review

中图分类号:

R496

曹梦琳,陈宇豪,王珏,刘天. 基于表面肌电图的人体运动意图识别研究进展[J]. 《中国康复理论与实践》, 2021, 27(5): 595-603.

Meng-lin CAO,Yu-hao CHEN,Jue WANG,Tian LIU. Advance in Human Motion Intention Recognition Based on Surface Electromyography (review)[J]. 《Chinese Journal of Rehabilitation Theory and Practice》, 2021, 27(5): 595-603.

图/表 5

图1

表1

表2

表3

表4

参考文献 48

1	SINGH R M, CHATTERJI S, KUMAR A. A review on surface EMG based control schemes of exoskeleton robot in stroke rehabilitation [C]. International Conference on Machine Intelligence & Research Advancement, IEEE (ICMIRA), 2013: 310-315.
2	GOPURA R A R C, BANDARA D S V, KIGUCHI K, et al. Developments in hardware systems of active upper-limb exoskeleton robots: a review [J]. Robot Auton Syst, 2016, 75: 203-220.
3	HUO W, MOHAMMED S, MORENO J C, et al. Lower limb wearable robots for assistance and rehabilitation: a state of the art [J]. IEEE Syst J, 2014, 10(3): 1068-1081.
4	MASAYUKI Y, RYOHEI K, MASAO Y. An evaluation of hand-force prediction using artificial neural-network regression models of surface emg signals for handwear devices [J]. J Sens, 2017, 2017: 1-12.
5	ARTEMIADIS P. EMG-based robot control interfaces: past, present and future [J]. Adv Robot Autom, 2012, 1(2): 1-3.
6	SARTORI M, REGGIANI M, FARINA D, et al. EMG-driven forward-dynamic estimation of muscle force and joint moment about multiple degrees of freedom in the human lower extremity [J]. PLoS One, 2012, 7(12): e52618.
7	XIA P, HU J, PENG Y. EMG-based estimation of limb movement using deep learning with recurrent convolutional neural networks [J]. Artif Organs, 2018, 42(5): E67-E77.
8	丁其川,熊安斌,赵新刚,等. 基于表面肌电的运动意图识别方法研究及应用综述[J]. 自动化学报, 2016, 42(1): 13-25.
	DING Q C, XIONG A B, ZHAO X G, et al. A review on researches and applications of sEMG-based motion intent recognition methods [J]. Acta Autom Sin, 2016, 42(1): 13-25.
9	SHEN H, SONG Q, DENG X, et al. Recognition of phases in sit-to-stand motion by Neural Network Ensemble (NNE) for power assist robot [C]. IEEE International Conference on Robotics and Biomimetics (ROBIO), 2007: 1703-1708.
10	KARAVAS N, AJOUDANI A, TSAGARAKIS N, et al. Tele-impedance based assistive control for a compliant knee exoskeleton [J]. Robot Auton Syst, 2015, 73: 78-90.
11	LLOYD D G, BESIER T F. An EMG-driven musculoskeletal model to estimate muscle forces and knee joint moments in vivo [J]. J Biomech, 2003, 36(6): 765-776.
12	HAN J, DING Q, XIONG A, et al. A state-space EMG model for the estimation of continuous joint movements [J]. IEEE Trans Ind Electron, 2015, 62(7): 4267-4275.
13	DING Q, XIONG A, ZHAO X, et al. A novel EMG-driven state space model for the estimation of continuous joint movements [C]. IEEE International Conference on Systems, Man, and Cybernetics (SMC), 2011: 2891-2897.
14	陈江城,张小栋,李睿,等. 利用表面肌电信号的下肢动态关节力矩预测模型[J]. 西安交通大学学报, 2015, 49(12): 26-33.
	CHEN J C, ZHANG X D, LI R, et al. Prediction model for dynamic joint torque of lower limb with surface EMG [J]. Acad J Xi'an Jiaotong Univ, 2015, 49(12): 26-33.
15	郭浩. 基于 EMG 的人体下肢运动意图感知与预测研究[D]. 哈尔滨:哈尔滨工业大学, 2019.
	GUO H. Research on human lower limb motion intention detection and prediction based on EMG [D]. Harbin: Harbin Institute of Technology, 2019.
16	PENG F, PENG W, ZHANG C. Evaluation of sEMG-based feature extraction and effective classification method for gait phase detection [C]. Cognitive Systems and Signal Processing, 2018: 138-149.
17	NAIK G R, SELVAN S E, ARJUNAN S P, et al. An ICA-EBM-based sEMG classifier for recognizing lower limb movements in individuals with and without knee pathology [J]. IEEE Trans Neural Syst Rehabil Eng, 2018, 26(3): 675-686.
18	曹祥红,刘磊,杨鹏,等. 利用多源信息和极限学习机的人体运动意图识别[J]. 传感技术学报, 2017, 30(8): 1171-1177.
	CAO X H, LIU L, YANG P, et al. The research of locomotion-mode recognition based on multi-source information and extreme learning machine [J]. Chin J Sens Actuators, 2017, 30(8): 1171-1177.
19	CAI S, CHEN Y, HUANG S, et al. SVM-based classification of sEMG signals for upper-limb self-rehabilitation training [J]. Front Neurorobot, 2019, 13: 31.
20	郑潇. 肌电信号多类特征分析及在步态识别中的应用[D]. 杭州:杭州电子科技大学, 2016.
	ZHENG X. Analysis of multi-features and application of gait recognition based on electromyographic signals [D]. Hangzhou: Hangzhou Dianzi University, 2016.
21	WEI P, XIE R, TANG R, et al. sEMG based gait phase recognition for children with spastic cerebral palsy [J]. Ann Biomed Eng, 2019, 47(1): 223-230.
22	BARBERI F, APRIGLIANO F, GRUPPIONI E, et al. Fast online decoding of motor tasks with single sEMG electrode in lower limb amputees [C]. International Symposium on Wearable Robotics (WeRob), 2018: 110-114.
23	XIE P, CHEN X, MA P P, et al. Identification method of human movement intention based on the fusion feature of EEG and EMG [J]. Proc World Congr Eng, 2013, 2205(1): 1340-1344.
24	ASTUDILLO F, CHARRY J, MINCHALA I, et al. Lower limbs motion intention detection by using pattern recognition [C]. IEEE Third Ecuador Technical Chapters Meeting (ETCM), 2018: 1-6.
25	MORBIDONI C, CUCCHIARELLI A, FIORETTI S, et al. A deep learning approach to EMG-based classification of gait phases during level ground walking [J]. Electronics, 2019, 8(8): 894-908.
26	KARANTARAT O, KITJAIDURE Y. The walking assistance system using the lower limb exoskeleton suit commanded by backpropagation neural network [C]. 11th Biomedical Engineering International Conference (BMEiCON), 2018: 1-5.
27	LUH G C, CAI J J, LEE Y S. Estimation of elbow motion intension under varing weight in lifting movement using an EMG-angle neural network model [C]. International Conference on Machine Learning and Cybernetics (ICMLC), IEEE, 2017: 640-645.
28	ZHANG F, LI P, HOU Z, et al. sEMG-based continuous estimation of joint angles of human legs by using BP neural network [J]. Neurocomputing, 2012, 78(1): 139-148.
29	XIE H, LI G, ZHAO X, et al. Prediction of limb joint angles based on multi-source signals by GS-GRNN for exoskeleton wearer [J]. Sensors, 2020, 20(4): 1104.
30	CHANDRAPAL M, CHEN X, WANG W, et al. Investigating improvements to neural network based EMG to joint torque estimation [J]. Paladyn J Behav Rob, 2011, 2(4): 185-192.
31	SHI Y, WANG S, LI J, et al. Prediction of continuous motion for lower limb joints based on sEMG signal [C]. IEEE International Conference on Mechatronics and Automation (ICMA), 2020: 383-388.
32	PARK K H, LEE S W. Movement intention decoding based on deep learning for multiuser myoelectric interfaces [C]. 4th International Winter Conference on Brain-Computer Interface (BCI), IEEE, 2016: 1-2.
33	CÔTÉ-ALLARD U, FALL C L, Drouin A, et al. Deep learning for electromyographic hand gesture signal classification using transfer learning [J]. IEEE Trans Neural Syst Rehabil Eng, 2019, 27(4): 760-771.
34	WEI W, DAI Q, WONG Y, et al. Surface-electromyography-based gesture recognition by multi-view deep learning [J]. IEEE Trans Biomed Eng, 2019, 66(10): 2964-2973.
35	BU N, FUKUDA O, TSUJI T. EMG-based motion discrimination using a novel recurrent neural network [J]. J Intell Inf Syst, 2003, 21(2): 113-126.
36	SONG J, ZHU A, TU Y, et al. Effects of different feature parameters of sEMG on human motion pattern recognition using multilayer perceptrons and LSTM neural networks [J]. Appl Sci, 2020, 10(10): 3358.
37	CHENG H R, CAO G Z, LI C H, et al. A CNN-LSTM hybrid model for ankle joint motion recognition method based on sEMG [C]. 17th International Conference on Ubiquitous Robots (UR), 2020: 339-344.
38	BETTHAUSER J L, KRALL J T, KALIKI R R, et al. Stable electromyographic sequence prediction during movement transitions using temporal convolutional networks [C]. 9th International IEEE/EMBS Conference on Neural Engineering (NER), 2019: 1046-1049.
39	BAO T, ZAIDI A, XIE S, et al. Surface-EMG based wrist kinematics estimation using convolutional neural network [C]. IEEE 16th International Conference on Wearable and Implantable Body Sensor Networks (BSN), 2019: 1-4.
40	RANE L, DING Z, MCGREGOR A H, et al. Deep learning for musculoskeletal force prediction [J]. Ann Biomed Eng, 2019, 47(3): 778-789.
41	AMERI A, AKHAEE M A, SCHEME E, et al. Regression convolutional neural network for improved simultaneous EMG control [J]. J Neural Eng, 2019, 16(3): 036015.
42	YANG W, YANG D, LIU Y, et al. Decoding simultaneous multi-DOF wrist movements from raw EMG signals using a convolutional neural network [J]. IEEE Trans Hum-Mach Syst, 2019, 49(5): 411-420.
43	DAO T T. From deep learning to transfer learning for the prediction of skeletal muscle forces [J]. Med Biol Eng Comput, 2019, 57(5): 1049-1058.
44	MA X, LIU Y, SONG Q, et al. Continuous estimation of knee joint angle based on surface electromyography using a long short-term memory neural network and time-advanced feature [J]. Sensors, 2020, 20(17): 4966.
45	MA C, LIN C, SAMUEL O W, et al. Continuous estimation of upper limb joint angle from sEMG signals based on SCA-LSTM deep learning approach [J]. Biomed Signal Process Control, 2020, 61: 102024.
46	XIA P, HU J, PENG Y. EMG-based estimation of limb movement using deep learning with recurrent convolutional neural networks [J]. Artif Organs, 2018, 42(5): E67-E77.
47	XU L, CHEN X, CAO S, et al. Feasibility study of advanced neural networks applied to sEMG-based force estimation [J]. Sensors (Basel), 2018, 18(10): 3226-3241.
48	GAUTAM A, PANWAR M, BISWAS D, et al. MyoNet: A transfer-learning-based LRCN for lower limb movement recognition and knee joint angle prediction for remote monitoring of rehabilitation progress from sEMG [J]. IEEE J Transl Eng Health Med, 2020, 8: 1-10.

作者	年份	对象	动作	提取特征	方法	时间	准确率
Xie等^[23]	2013	脑卒中、周围神经损伤、健康人	伸膝/屈膝	有效PF的多尺度熵	ELM	—	80.5% (脑卒中) 79.6% (周围神经损伤) 93.4% (健康人)
郑潇^[20]	2016	健康人	支撑前期/中期/后期/摆动前期/后期	MAV和VAR(SVM)；非线性分形维数(K-Means)	SVM、改进K-Means	—	SVM > 95% K-Means 92.1%
Barberi等^[22]	2018	截肢者	步行/上坡/下坡/上下楼梯	MFL	LDA	< 100 ms	100%
Wei等^[21]	2018	痉挛性脑瘫患儿	站立中期/站立末期/摆动前期/摆动中期/摆动末期	多种时域或频域特征的组合，最终选取MAV和ZC	SVM	—	89.40%
Peng等^[16]	2018	健康男性	站立前态/中间态/末态/前摆/中摆/终摆	SSC、WL、Wamp、Logvar、DB7-MAV	SV-LDA、WT-LDA、XGBoost、LightGBM	85 ms	94.3% (LightGBM)
Karantarat等^[26]	2018	患者、健康人	步行/坐着/站立	时域特征	BPNN	—	99.39%
Astudillo等^[24]	2018	健康人	两种动作	RMS、导数	ANN	(12.6±10) ms	94.88%
Morbidoni等^[25]	2019	健康人	站立/摆动	EMG信号的包络线	MLP	—	93.4%

作者	年份	对象	回归目标	方法	结果
Zhang等^[28]	2012	健康人	脚踝、膝盖、臀部关节角度	BPNN	平均误差< 9°
Luh等^[27]	2017	健康人	肘关节角度	BPNN	有较高的精度
Chandrapal等^[30]	2011	健康人	膝关节扭矩	MLP	平均最低估计误差10.46%
Masayuki等^[4]	2017	健康人	握力	ANN	平均CC = 0.84
Xie等^[29]	2020	—	髋、膝、踝关节角度	GS-GRNN	髋CC = 0.9985，膝CC = 0.9925，踝CC = 0.9873

作者	年份	研究对象	分类动作	网络输入	分类方法	时间	准确率
Bu等^[35]	2003	健康人	6种手部运动	原始EMG	递归对数线性化高斯混合网络	—	88.4%
Park等^[32]	2016	健康人	6种手部运动	原始EMG	CNN	—	90%
C?té-Allard等^[33]	2019	健康人	手势识别	原始EMG，频谱图，CWT	ConvNet	—	98.31%
Wei等^[21]	2019	11个数据库，部分数据库中有截肢者	手势识别	通过穷举，找到3种特征，输入网络	multi-view深度学习框架	—	较高的识别精度
Song等^[36]	2020	健康人	7种常见运动模式	时域、频域特征	MLP、LSTM	—	MLP 95.53%， LSTM 96.57%
Cheng等^[37]	2020	—	4种踝关节动作	原始EMG	CNN-LSTM	—	97.55%±1.93%
Betthauser等^[38]	2019	—	3种手部运动	MAV	TCN	—	准确率与LSTM相当，稳定性高与LSTM

作者	年份	研究对象	回归目标	网络输入	回归方法	时间	回归精度
Bao等^[39]	2019	健康人	腕关节角度	原始EMG信号、时域、频谱	CNN	—	频谱图效果最好
Rane等^[40]	2019	健康人、骨关节炎患者	骨骼肌力	原始EMG信号	CNN	71 ms	结果优于比赛中6个获奖作品中的4个
Ameri等^[41]	2019	健康人	关节自由度	EMG信号矩阵	CNN	6 ms	误差< 10%
Yang等^[42]	2019	健康人	关节自由度	EMG信号	CNN	比SVM快	比SVM模型准确
Dao^[43]	2019	健康人	骨骼肌力	原始EMG信号	LSTM、权重转移策略	—	RMSE < 10%， CC = 0.95~0.99
Xia等^[46]	2018	健康人	上肢运动轨迹	时频信号	RCNN	—	平均CC 0.903
Xu等^[47]	2018	健康人	上肢力	NMF信号提取	CNN、LSTM、CNN-LSTM	—	平均RMSE(%) CNN：12.13±1.98 LSTM：9.07±1.29 CNN-LSTM：8.67±1.14
Ma等^[44]	2020	健康人	膝关节角度	sEMG的RMS及其时间提前特征	LSTM	—	平均RMSE (3.4726±0.6162)°
Ma等^[45]	2020	健康人	上肢关节角度	EMG信号包络	SCA-LSTM	—	平均CC (0.957±0.013)
Gautam等^[48]	2020	健康人和膝关节病患者	膝关节角度	EMG信号	LRCN	—	MAE：健康人为8.1%，患者为9.2%

[1]	邵伟婷, 雷江华. 反应中断再定向干预孤独症谱系障碍儿童刻板语言的效果：Scoping综述[J]. 《中国康复理论与实践》, 2024, 30(1): 10-20.
[2]	王航宇, 葛可可, 范永红, 都丽露, 邹敏, 封磊. 基于ICD-11和ICF主动式音乐疗法改善认知障碍老年人认知功能的系统综述[J]. 《中国康复理论与实践》, 2024, 30(1): 36-43.
[3]	闻嘉宁, 金秋艳, 张琦, 李杰, 司琦. 认知参与型身体活动对发展儿童青少年执行功能的效果：基于ICF的系统综述[J]. 《中国康复理论与实践》, 2024, 30(1): 44-53.
[4]	葛可可, 范永红, 王航宇, 都丽露, 李长江, 邹敏. 失眠老年人正念干预健康效益的系统综述[J]. 《中国康复理论与实践》, 2024, 30(1): 54-60.
[5]	张婧雅, 邹敏, 孙宏伟, 孙昌隆, 朱峻同. 听障儿童青少年焦虑或抑郁情绪心理干预效果的系统综述[J]. 《中国康复理论与实践》, 2023, 29(9): 1004-1011.
[6]	王俊宇, 杨永, 袁逊, 谢婷, 庄洁. 高强度间歇训练对健康儿童青少年执行功能效果的系统综述[J]. 《中国康复理论与实践》, 2023, 29(9): 1012-1020.
[7]	魏晓微, 杨剑, 魏春艳. 特殊教育学校孤独症谱系障碍儿童参与适应性瑜伽活动的心理与行为效益的系统综述[J]. 《中国康复理论与实践》, 2023, 29(9): 1021-1028.
[8]	杨亚茹, 杨剑. 基于WHO-HPS架构学校身体活动相关健康服务及其健康效益：系统综述的系统综述[J]. 《中国康复理论与实践》, 2023, 29(9): 1040-1047.
[9]	史佳伟, 李凌宇, 杨浩杰, 王琴潞, 邹海欧. 预康复对全膝关节置换术后患者的有效性：系统综述的系统综述[J]. 《中国康复理论与实践》, 2023, 29(9): 1057-1064.
[10]	蔡华年, 费思先, 张忆晨, 孙青, 郭帅, 宋韬. 基于导纳控制的双边康复机器人运动辅助分析[J]. 《中国康复理论与实践》, 2023, 29(9): 1104-1109.
[11]	蒋长好, 黄辰, 高晓妍, 戴元富, 赵国明. 神经反馈训练对老年人认知功能效果的系统综述[J]. 《中国康复理论与实践》, 2023, 29(8): 903-909.
[12]	魏晓微, 杨剑, 魏春艳, 贺启令. 学校环境下适应性体育课程促进智力与发展性残疾儿童心理运动发展的系统综述[J]. 《中国康复理论与实践》, 2023, 29(8): 910-918.
[13]	张园, 杨剑. 基于世界卫生组织健康促进学校架构的学校健康服务及效果：Scoping综述[J]. 《中国康复理论与实践》, 2023, 29(7): 791-799.
[14]	王少璞, 陈钢. 基于世界卫生组织健康促进学校架构的心理行为健康服务及其健康效益：系统综述的系统综述[J]. 《中国康复理论与实践》, 2023, 29(7): 800-807.
[15]	蒋长好, 高晓妍. 短时身体活动对儿童认知功能影响的系统综述[J]. 《中国康复理论与实践》, 2023, 29(6): 667-672.

基于表面肌电图的人体运动意图识别研究进展

Advance in Human Motion Intention Recognition Based on Surface Electromyography (review)

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 5

参考文献 48

相关文章 15

Metrics

本文评价

推荐阅读 0