A Feature Extraction Pipeline for Enhancing Lightweight Neural Networks in sEMG-based Joint Torque Estimation

2026-01-23

AI Summary

This study addresses the challenge of accurately estimating elbow and shoulder joint torques in robot-assisted rehabilitation, which is critical for delivering personalized assistance yet remains difficult under lightweight modeling and limited data conditions. The authors propose a feature extraction pipeline based on 8-channel surface electromyography (sEMG), combining handcrafted features with torque references derived from static equilibrium assumptions. This approach is integrated into both a multilayer perceptron (MLP) and a temporal convolutional network (TCN) for evaluation. Experimental results demonstrate that the proposed method enables a structurally simple MLP to achieve performance comparable to that of a TCN under data-scarce conditions, with mean root-mean-square errors (RMSE) of 0.963, 1.403, and 1.434 N·m for elbow, anterior shoulder, and lateral shoulder torque estimation, respectively—significantly reducing both model complexity and data requirements.

Technology Category

• Intelligent Robots: State Estimation
• Machine Learning: Learning with Manifolds
• Natural Language Processing: Safety and Robustness

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Robot-assisted rehabilitation offers an effective approach, wherein exoskeletons adapt to users'needs and provide personalized assistance. However, to deliver such assistance, accurate prediction of the user's joint torques is essential. In this work, we propose a feature extraction pipeline using 8-channel surface electromyography (sEMG) signals to predict elbow and shoulder joint torques. For preliminary evaluation, this pipeline was integrated into two neural network models: the Multilayer Perceptron (MLP) and the Temporal Convolutional Network (TCN). Data were collected from a single subject performing elbow and shoulder movements under three load conditions (0 kg, 1.10 kg, and 1.85 kg) using three motion-capture cameras. Reference torques were estimated from center-of-mass kinematics under the assumption of static equilibrium. Our offline analyses showed that, with our feature extraction pipeline, MLP model achieved mean RMSE of 0.963 N m, 1.403 N m, and 1.434 N m (over five seeds) for elbow, front-shoulder, and side-shoulder joints, respectively, which were comparable to the TCN performance. These results demonstrate that the proposed feature extraction pipeline enables a simple MLP to achieve performance comparable to that of a network designed explicitly for temporal dependencies. This finding is particularly relevant for applications with limited training data, a common scenario patient care.

Problem

• sEMG-based joint torque estimation
• lightweight neural networks
• feature extraction
• robot-assisted rehabilitation
• elbow and shoulder torque prediction

Innovation

• feature extraction pipeline
• lightweight neural networks
• sEMG-based torque estimation
• MLP vs TCN
• robot-assisted rehabilitation