Un enfoque con aprendizaje profundo para el análisis de señales electromiográficas de superficie: hacia el control de órtesis robótica de codo
A deep learning approach for surface electromyographic signal analysis: toward robotic elbow orthosis control
DOI:
https://doi.org/10.59420/remus.2.2025.323Palabras clave:
Inteligencia Artificial, Ortesis, EMGResumen
Las discapacidades motoras derivadas de enfermedades neuromusculares afectan significativamente la calidad de vida de los pacientes. En este trabajo se propone un modelo híbrido CNN-LSTM con enfoque regresivo para estimar en tiempo real el ángulo articular del codo a partir de señales electromiográficas superficiales (sEMG), con el objetivo de controlar de forma precisa y continua una órtesis robótica. Se emplearon señales sEMG de bíceps y tríceps de una base de datos validada, las cuales fueron sometidas a un riguroso preprocesamiento. El modelo fue entrenado y evaluado utilizando métricas como MAE (0.0828), RMSE (0.1132), R² (0.9517) y coeficiente de Pearson (0.9757), evidenciando alta precisión y robustez. Además, se desarrolló una interfaz gráfica para visualizar las señales y predicciones en tiempo real. Los resultados confirman la viabilidad del enfoque propuesto para aplicaciones clínicas, destacando su potencial para mejorar la rehabilitación funcional del miembro superior y promover la autonomía del paciente mediante un control intuitivo de órtesis activas.
Descargas
Citas
Gandolla M, Antonietti A, Longatelli V, Pe- drocchi A. The effectiveness of wearable upper limb assistive devices in degenera- tive neuromuscular diseases: a systematic review and meta-analysis. Front Bioeng Biotechnol. 2020 Jan 24;7;450. https://doi. org/10.3389/fbioe.2019.00450
Khalid S, Alnajjar F, Gochoo M, Renawi A, Shimoda S. Robotic assistive and rehabil- itation devices leading to motor recovery in upper limb: a systematic review. Disabil Rehabil Assist Technol. 2021;18(5):658– 672. http://doi.org/10.1080/17483107.2021.1906960
Cai S, Chen Y, Huang S, Wu Y, Zheng H, Li X, et al. SVM-Based Classification of sEMG Signals for Upper-Limb Self-Rehabilitation Training. Front. Neurorobot. 2019, 13:31. https://doi.org/10.3389/fnbot.2019.00031
Comella CL, Stebbins GT, Brown-Toms N, Goetz CG. Physical therapy and Parkinson’s disease: a controlled clinical trial. Neuro- logy. 1994;44(3 Pt 1):376–378. https://doi. org/10.1212/wnl.44.3_part_1.376
Freeman J, Langdon DW, Hobart JC, Thomp- son AJ. The impact of inpatient rehabilita- tion on progressive multiple sclerosis. Ann Neurol. 1997;42(2):236–244. https://doi. org/10.1002/ana.410420216
Gittler M, Davis AM. Guidelines for adult stroke rehabilitation and recovery. JAMA. 2018;319(8):820–821. http://doi.org/10.1001/jama.2017.22036
Maciejasz P, Eschweiler J, Gerlach-Hahn K, Jansen-Toy LJ, Leonhardt S. A survey on robotic devices for upper limb rehabil- itation. J Neuroeng Rehabil. 2014;11(1):3. https://doi.org/10.1186/1743-0003-11-3
Wei Z, Zhang ZQ, Xie SQ. Continuous mo- tion intention prediction using sEMG for upper-limb rehabilitation: a systematic review of model-based and model-free approaches. IEEE Trans Neural Syst Re- habil Eng. 2024;32:1487–1504. http://doi. org/10.1109/TNSRE.2024.3383857
Guerrero-Mendez CD, Blanco-Díaz CF, Ja- ramillo-Isaza S, Bastos-Filho TF, Ruiz-Ola- ya AF. Artificial intelligence applied to neuromotor rehabilitation engineering: ad- vances and challenges. In: Computational Approaches in Biomaterials and Biomedical Engineering Applications. 2024:212–244. http://doi.org/10.1201/9781032699882-10
Morimoto SYU, Cabral AKPDS, Sanguinetti DCDM, Freitas EDSRD, Merino GSAD, Costa JÂPD, et al. Upper limbs orthosis and pros- theses printed in 3D: an integrative review. Cad Bras Ter Ocup. 2021;29:e2078. https:// doi.org/10.1590/2526-8910.ctoAO2078
Shehata AW, Williams HE, Hebert JS, Pilars- ki PM. Machine learning for the control of prosthetic arms: using electromyographic signals for improved performance. IEEE Signal Process Mag. 2021;38(4):46–53. https://doi.org/10.1109/MSP.2021.3075931
Cardoso LRL, Bochkezanian V, Forner-Cordero A, Melendez-Calderon A, Bo APL. Soft robotics and functional electri- cal stimulation advances for restoring hand function in people with SCI: a narrative re- view, clinical guidelines and future direc- tions. J Neuroeng Rehabil. 2022;19(1):66. https://doi.org/10.1109/MSP.2021.3075931
Marquez-Chin C, Popovic MR. Functional electrical stimulation therapy for restora- tion of motor function after spinal cord in- jury and stroke: a review. Biomed Eng Onli- ne. 2020;19(1):34. https://doi.org/10.1186/ s12938-020-00773-4
Zhang A, Li Q, Li Z, Li J. Multimodal fusion convolutional neural network based on sEMG and accelerometer signals for inter- subject upper limb movement classification. IEEE Sens J. 2023;23(11):12334–12345.
https://doi.org/10.1109/JSEN.2023.3266872
Fleming A, Stafford N, Huang S, Hu X, Fer- ris DP, Huang H. Myoelectric control of robotic lower limb prostheses: a review of electromyography interfaces, control par- adigms, challenges and future directions. J Neural Eng. 2021;18(4):041004. https:// doi.org/10.1088/1741-2552/ac1176
Li X, Tian L, Zheng Y, Samuel OW, Fang P, Wang L, et al. A new strategy based on feature filtering technique for improving the real-time control performance of myo- electric prostheses. Biomed Signal Pro- cess Control. 2021;70:102969. https://doi. org/10.1016/j.bspc.2021.102969
Xie C, Yang Q, Huang Y, Su S, Xu T, Song
R. A hybrid arm-hand rehabilitation ro- bot with EMG-based admittance con- troller. IEEE Trans Biomed Circuits Syst. 2021;15(6):1332–1342. https://doi. org/10.1109/TBCAS.2021.3130090
Khasnabis C, Motsch K, Achu K, Al Jubah K, Brodtkorb S, Chervin P, Coleridge P, et al., editors. Community-Based Rehabilitation: CBR Guidelines. Assistive Devices. Geneva: World Health Organization; 2010. https:// www.ncbi.nlm.nih.gov/books/NBK310951/
Velásquez E, Cornelis J, Omelina L, Jan- sen B. Muscle classification via hybrid CNN-LSTM architecture from surface EMG signals. In: 2023 24th Int Conf on Digital Signal Processing (DSP). Rhodes, Greece; 2023:1–5. https://2023.ic-dsp.org/wp-con- tent/uploads/2023/05/DSP2023-25.pdf
Ajit A, Acharya K, Samanta A. A review of convolutional neural networks. In: 2020 International Conference on Emerging Trends in Information Technology and En- gineering (ic-ETITE). Vellore, India: IEEE; 2020:1–5. https://doi.org/10.1109/ic-ETI- TE47903.2020.049
Ameri A, Akhaee MA, Scheme E, Engle- hart K. A deep transfer learning approach to reducing the effect of electrode shift in EMG pattern recognition-based con- trol. IEEE Trans Neural Syst Rehabil Eng. 2020;28(2):370–379. https://doi. org/10.1109/tnsre.2019.2962189
Kim KT, Guan C, Lee SW. A subject-trans- fer framework based on single-trial EMG analysis using convolutional neural net- works. IEEE Trans Neural Syst Reha- bil Eng. 2020;28(1):94–103. https://doi. org/10.1109/tnsre.2019.2946625
Soto Alapont J. Analysis of feedforward and LSTM deep neural network architectu- res for self-supervised regression of sEMG signals for multi-grasp robot hand control [tesis de maestría]. Valencia, Spain: Univer- sitat Politècnica de València; 2023. https:// riunet.upv.es/handle/10251/196516
Ghislieri M, Cerone GL, Knaflitz M, Agostini
V. Long short-term memory (LSTM) recur- rent neural network for muscle activity de- tection. J Neuroeng Rehabil. 2021;18:153. https://doi.org/10.1186/s12984-021- 00945-w
Basheer I, Hajmeer M. Artificial neural networks: fundamentals, computing, de- sign, and application. J Microbiol Methods. 2000;43(1):3–31. https://doi.org/10.1016/ s0167-7012(00)00201-3
Graves A. Generating sequences with re- current neural networks [preprint]. 2013. http://arxiv.org/abs/1308.0850
Bittibssi TM, Zekry AH, Genedy MA, Maged SA. sEMG pattern recognition based on recurrent neural network. Biomed Signal Process Control. 2021;70:103048.
Toro-Ossaba A, Tejada JC. EMG elbow da- taset [dataset]. Zenodo; 2023. https://doi. org/10.5281/zenodo.7946782
Song T, Yan Z, Guo S, Li Y, Li X, Xi F. Review of sEMG for robot control: Techniques and applications. Appl Sci. 2023;13(17):9546. https://doi.org/10.3390/app13179546
Wang S, Zhu S, Shang Z. A novel combi- nation method of a convolutional neural network and energy operators for the de- tection of change-points in electromyo- graphic signals. Appl Sci. 2023;13(2):923. https://doi.org/10.3390/app13020923
Martinek R, Ladrova M, Sidikova M, Jaros R, Behbehani K, Kahankova R, et al. Ad- vanced bioelectrical signal processing me- thods: Past, present, and future approach— Part III: Other biosignals. Sensors (Basel). 2021;21(18):6064. https://doi.org/10.3390/ s21186064
Amazon Web Services. Dividir los datos en datos de entrenamiento y evaluación. AWS. https://docs.aws.amazon.com/es_es/ma- chine-learning/latest/dg/splitting-the-da- ta-into-training-and-evaluation-data.html
Kaur M, Mathur S, Bhatia D, Verma S. siG- num: Graphical user interface for EMG sig- nal analysis. J Med Eng Technol. 2014 Nov 11;39(1):19–25. https://pubmed.ncbi.nlm. nih.gov/25385355/
Khoshroo P. Control system and graphi- cal user interface design of an Upper-Ex- tremity rehabilitation robot. 2020. https:// uwspace.uwaterloo.ca/items/718d7daa- e0a3-444e-b131-b785b95d04fb
Hodson TO. Root mean square error (RMSE) or mean absolute error (MAE): When to use them or not. Geosci Model Dev Discuss. 2022;2022:1–10. https://doi.org/10.5194/ gmd-15-5481-2022
MathWorks. Coeficiente de determina- ción (R cuadrado). https://la.mathworks. com/help/stats/coefficient-of-determina- tion-r-squared.html
Chicco D, Warrens MJ, Jurman G. The coef- ficient of determination R-squared is more informative than SMAPE, MAE, MAPE, MSE and RMSE in regression analysis evaluation. PeerJ Comput Sci. 2021;7:e623. https://doi.org/doi:10.7717/peerj-cs.623
Alqahtani H, Ray A. Feature selection of surface topography parameters for fa- tiguedamage detection using Pear- son correlation method and neural net- work analysis. Fatigue Fract Eng Mater Struct. 2023;46(5):1810–20. https://doi. org/10.1111/ffe.13963
Uslu G, Baydere S, Tekin A, Subasi F. Ac- celerometer based classification of el- bow flexion and extension exercises. 2021 IEEE International Conference on Bioin- formatics and Biomedicine (BIBM). 2020 Dec 16;2657–63. https://doi.org/10.1109/ bibm49941.2020.9313155
Cai Z, Zhu Y. A hybrid CNN-LSTM network for hand gesture recognition with surface EMG signals. Proc SPIE 11878. 2021 Jun 30; Thirteenth International Conference on Digital Image Processing (ICDIP 2021):74. https://doi.org/10.1117/12.2601074
Chen Y, Zhang H, Wang C, Ang KK, Ng SH, Jin H, et al. A hierarchical dynamic Baye- sian learning network for EMG-based early prediction of voluntary movement inten- tion. Scientific Reports. 2023 Mar 23;13(1). https://doi.org/10.1038/s41598-023- 30716-7
Instituto Nacional de Estadística y Geo- grafía (INEGI). Estadísticas a propósito del Día Internacional de las Personas con Dis- capacidad. Aguascalientes: INEGI; 2024. https://www.inegi.org.mx
Descargas
Publicado
Cómo citar
Número
Sección
Licencia
Derechos de autor 2025 REMUS - Revista Estudiantil de Medicina de la Universidad de Sonora
Esta obra está bajo una licencia internacional Creative Commons Atribución-NoComercial-SinDerivadas 4.0.
