The manual lifting of heavy loads by personnel is susceptible to the development of muscle fatigue, which, in severe cases, can result in the irreversible impairment of muscle function. This study proposes a novel method of signal fusion to analyse muscle fatigue during manual lifting. Furthermore, this study represents the inaugural application of the back-propagation neural network and bidirectional encoder representation from the transformer (BP + BERT) algorithm to the fusion of two sensor inputs for the analysis of muscle fatigue. Lifting action fatigue tests were carried out on 16 testers in this study, with both surface electromyography (sEMG) and mechanomyography (MMG) signals collected as part of the process. The mean power frequency (MPF) eigenvalues were extracted separately for the two signals, and the results of muscle fatigue labelling according to the trend of the MPF eigenpeak were merged to produce three datasets. Subsequently, the three datasets were employed to categorise muscle fatigue classes using the support vector machine and radial basis function (SVM + RBF), support vector machine and bidirectional encoder representation from transformer (SVM + BERT), back-propagation neural network (BP), and back-propagation neural network and bidirectional encoder representation from transformer (BP + BERT) algorithms, respectively. The results of the muscle fatigue classification model demonstrated that the sEMG and MMG fused dataset, imported into the BP + BERT algorithm, exhibited the highest average accuracy of 98.10% for the muscle fatigue classification model. This study indicates that the fusion of sEMG and MMG signals is an effective approach, and the performance of the BP + BERT muscle fatigue classification model is also enhanced.
(1) Background: Exoskeleton systems are being used increasingly in medical rehabilitation and industrial assistance. Accurate evaluation of their performance is crucial for optimising design and improving performance. In this paper, we propose a performance evaluation index system for exoskeletons and a quantitative evaluation method based on multi-source signal fusion. (2) Methods: The system utilises a variety of sensors for signal collection, encompassing sEMG and MMG, which are indicative of muscular fatigue, along with IMU signals, which demonstrate the movement of both the exoskeleton and the human body. This is then combined with the subjective evaluation scale to constitute the exoskeleton effectiveness assessment index system. (3) Results: This study evaluates two types of material handling exoskeletons. The findings indicate that the ultimate scores for the single-hip and hip-knee exoskeletons are 0.625 and 0.845, respectively. (4) Conclusions: A comparison of the indexes of the two exoskeletons and the subjective evaluations of the testers reveals that this method effectively quantifies the exoskeleton’s human metabolism, human–machine synergy and performance in working conditions, as well as the perceived effectiveness of assistance in different working conditions.
Misalignment between the rotation axis of a lower-limb exoskeleton knee joint and the human knee joint in the sagittal plane introduces additional human-exoskeleton interaction forces, affecting gait and assistance efficiency. This study proposes an equivalent stiffness-damping scheme that simultaneously accounts for the serial characteristics of both the exoskeleton’s strapping and human soft tissues, and establishes a human-exoskeleton coupling model based on OpenSim that allows for precise adjustment of the misalignment magnitude along the sagittal coordinate system. Simulation results indicate that the effects of misalignment exhibit significant directional dependence: at a misalignment of 0.05 m, the peak increase in extension torque in the positive y-axis direction reached 38.9%, while the peak increase in flexion torque in the negative x-axis direction reached 211.2%. The trends in human-exoskeleton interaction torque and electromyographic signals measured experimentally were consistent with the simulation results. Considering that, at a misalignment of 0.02 m, the maximum difference in torque in the negative x-axis direction was 54.65% and the assist efficiency in the positive x-axis direction dropped to −5.13%, it is recommended that the misalignment of the exoskeleton knee joint be controlled within 0.02 m. This provides quantitative evidence for the structural design and wear calibration of the exoskeleton.
In recent years, advances in exoskeleton technology have increased the demand for human motion classification in terms of both movement diversity and recognition accuracy, making the effective classification of more complex asymmetric movements increasingly important. This study integrated surface electromyography (sEMG) and inertial measurement unit (IMU) signals and employed mutual information (MI) and muscle symmetry features (MSF) to analyze the characteristic differences between symmetrical muscles in both legs during asymmetric movements, with the aim of improving the accuracy of asymmetric motion classification. sEMG and IMU signals were collected from six movements, including three asymmetric postures: asymmetrical stance, single-knee kneeled position, and crouching advance. The acquired signals were processed through energy envelope analysis, active segment extraction, empirical mode decomposition (EMD), feature extraction, MI extraction, and MSF extraction. The relevance based on weight feature selection (RWFS) combined with conditional mutual information (CMI) method was then applied to reduce feature dimensionality, prioritize features with significant fluctuations, and preserve key characteristics. Finally, the CNN-LSTM-Attention algorithm was used for classification. Experimental results showed that fusing sEMG and IMU signals achieved 96.30% accuracy in lower limb motion recognition. The proposed method improves asymmetric movement classification and may provide a potential basis for exoskeleton motion classification in special environments.