Worker Activity Recognition in Smart Manufacturing Using IMU and sEMG Signals with Convolutional Neural Networks

Tao, Wenjin; Lai, Ze-Hao; Leu, Ming C.; Yin, Zhaozheng

doi:10.29007/bld3

Cited by 15 publications

(9 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Human body modeling can be used in combination with motion capture systems and FSR sensors to identify muscle or joint contact force [ 116 ]. Finally, with regard to IMU motion reconstruction techniques, the Myo armband is commonly used because it captures both EMG and IMU signals [ 67 , 117 , 119 ]. As identified before, most of the studies are carried out in simulated environments and have a short duration.…”

Section: Results: Emg Applications In Production Engineeringmentioning

confidence: 99%

“…Tao et al [ 67 ] develop a system for evaluating worker activities using inertial sensors and a Myo sEMG signal acquisition device. Data are acquired from eight subjects who perform six different activities, such as removing a tool from a holder, hammering, or tightening nuts.…”

Section: Results: Emg Applications In Production Engineeringmentioning

confidence: 99%

“…The use of CNN networks for the analysis of EMG signals is quite recent. Tao et al [ 67 ] use a convolutional network to identify a worker’s activity using a Myo bracelet, which provides data from an IMU sensor and eight electromyographic channels. The activity maps of the IMU have a 30% higher performance compared to the rest of inputs, having more discriminatory components associated with the activity that the EMG signals.…”

Section: Signal Analysismentioning

confidence: 99%

See 2 more Smart Citations

EMG Characterization and Processing in Production Engineering

Olmo

Domingo

2020

Materials

View full text Add to dashboard Cite

Electromyography (EMG) signals are biomedical signals that measure electrical currents generated during muscle contraction. These signals are strongly influenced by physiological and anatomical characteristics of the muscles and represent the neuromuscular activities of the human body. The evolution of EMG analysis and acquisition techniques makes this technology more reliable for production engineering applications, overcoming some of its inherent issues. Taking as an example, the fatigue monitoring of workers as well as enriched human–machine interaction (HMI) systems used in collaborative tasks are now possible with this technology. The main objective of this research is to evaluate the current implementation of EMG technology within production engineering, its weaknesses, opportunities, and synergies with other technologies, with the aim of developing more natural and efficient HMI systems that could improve the safety and productivity within production environments.

show abstract

Section: Results: Emg Applications In Production Engineeringmentioning

confidence: 99%

Section: Results: Emg Applications In Production Engineeringmentioning

confidence: 99%

Section: Signal Analysismentioning

confidence: 99%

See 1 more Smart Citation

EMG Characterization and Processing in Production Engineering

Olmo

Domingo

2020

Materials

View full text Add to dashboard Cite

show abstract

“…Asif [ 28 ] investigated the effect of hyper-parameters on each hand gesture, providing a robust and stable hand gesture design scheme for the deep learning-based approach. Tao [ 30 ] combined the IMU signal and the sEMG signal to improve the accuracy of gesture recognition. Further study [ 31 ] explained the ability of a convolutional neural network to combine data from two different modalities.…”

Section: Related Workmentioning

confidence: 99%

Surface EMG-Based Instantaneous Hand Gesture Recognition Using Convolutional Neural Network with the Transfer Learning Method

Zhao

Wang

et al. 2021

Sensors

View full text Add to dashboard Cite

In recent years, surface electromyography (sEMG)-based human–computer interaction has been developed to improve the quality of life for people. Gesture recognition based on the instantaneous values of sEMG has the advantages of accurate prediction and low latency. However, the low generalization ability of the hand gesture recognition method limits its application to new subjects and new hand gestures, and brings a heavy training burden. For this reason, based on a convolutional neural network, a transfer learning (TL) strategy for instantaneous gesture recognition is proposed to improve the generalization performance of the target network. CapgMyo and NinaPro DB1 are used to evaluate the validity of our proposed strategy. Compared with the non-transfer learning (non-TL) strategy, our proposed strategy improves the average accuracy of new subject and new gesture recognition by 18.7% and 8.74%, respectively, when up to three repeated gestures are employed. The TL strategy reduces the training time by a factor of three. Experiments verify the transferability of spatial features and the validity of the proposed strategy in improving the recognition accuracy of new subjects and new gestures, and reducing the training burden. The proposed TL strategy provides an effective way of improving the generalization ability of the gesture recognition system.

show abstract

“…Tendo em vista a grande quantidade de dados coletados por meio dos dispositivos electrónicos e com o intuito de processar esta informação de forma rápida e autónoma, torna-se necessária a aplicação de processos baseados em aprendizagem computacional. Estudos académicos com foco no reconhecimento de atividades/ações humanas passaram pelo desenvolvimento de algoritmos baseados em Machine Learning [11][12][13] e seu subtipo o Deep Learning [14][15]. Como classificadores baseados em Machine Learning, pode-se destacar a aplicação de Decision Tree (DT), Discriminant Analysis (DA), Support Vector Machine (SVM), k-Nearest Neighbour (kNN), Naïve Bayes (NB), Multilayer perceptron (MP) [14], [16][17].…”

Section: Monitorização Autónoma Do Desempenhounclassified

BIM WD: A Dimensão dos Trabalhadores Integrada nos Modelos de Informação

Calvetti

Sanhudo

Mêda

et al. 2020

3º Congresso Português De 'Building Information Modelling'

View full text Add to dashboard Cite

ResumoAdicionar "dimensões" ou informação a modelos BIM tem o potencial de ampliar a compreensão sobre um projeto de construção. As dimensões BIM retratam a forma como esta metodologia é aplicada a todo o ciclo de vida da construção, nomeadamente: 4D (temporal); 5D (financeira); 6D (sustentabilidade); 7D (gestão das instalações); para além das três dimensões do espaço euclidiano. A segurança em obra e a produtividade na indústria da construção podem ser melhoradas a partir da análise detalhada das tarefas realizadas no estaleiro. Os avanços tecnológicos, designadamente os dispositivos vestíveis (Wearables), propiciam novas formas de recolher este tipo de informação, enquanto que o recurso à inteligência artificial permite agilizar a sua análise. Estudos em desenvolvimento no CONSTRUCT/Gequaltec (FEUP) recorrem a dispositivos electrónicos para a colheita de dados relativos aos movimentos dos trabalhadores da construção. À posteriori, a análise de dados é efectuada por meio de algoritmos baseados em aprendizagem computacional. Os resultados desta análise permitem monitorizar a ocupação dos trabalhadores durante as suas atividades diárias (e.g. serrar, pintar, caminhar, descansar), permitindo uma avaliação detalhada dos processos produtivos, e demonstrando um elevado potencial para gestão de obras, em particular nos domínios da segurança e do controlo da produtividade. Conceitualmente, estes novos dados são processados por meio de tecnologias e técnicas de informação, sendo posteriormente integrados no BIM e dando origem a uma nova dimensão, aqui denominada WD Workers Dimension. Apresenta-se neste artigo os conceitos e modelos desta nova dimensão/uso BIM, potenciados pela utilização de Wearables e a aplicação de Machine Learning.

show abstract

Worker Activity Recognition in Smart Manufacturing Using IMU and sEMG Signals with Convolutional Neural Networks

Cited by 15 publications

References 7 publications

EMG Characterization and Processing in Production Engineering

EMG Characterization and Processing in Production Engineering

Surface EMG-Based Instantaneous Hand Gesture Recognition Using Convolutional Neural Network with the Transfer Learning Method

BIM WD: A Dimensão dos Trabalhadores Integrada nos Modelos de Informação

Contact Info

Product

Resources

About