sEMG-Based Joint Angle Estimation via Hierarchical Spiking Attentional Feature Decomposition Network

2024-04-11

IEEE Robotics and Automation Letters

AI Summary

To address the challenge of cross-subject real-time joint-angle estimation from surface electromyography (sEMG) in lower-limb rehabilitation exoskeletons, this paper proposes the Spiking Attention-based Feature Decomposition Network (SAFE-Net). SAFE-Net introduces, for the first time, a spiking sparse attention encoder and a spiking attention-based feature decomposition module, explicitly disentangling kinematic dynamics from subject-specific variability. It jointly supports joint-angle regression and subject identification. Leveraging spiking neural networks (SNNs) for temporal sEMG modeling, SAFE-Net integrates sparse attention mechanisms and feature-decoupled learning. Evaluated on two public datasets, it reduces inference energy consumption by 39.1% and 37.5% compared to Informer and Spikformer, respectively, while improving angular prediction accuracy (MAE reduced by 12.3%) and subject identification accuracy (increased by 4.8%). The model thus achieves low-power operation, strong cross-subject generalization, and effective multi-task synergy.

Technology Category

• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Natural Language Processing: Safety and Robustness
• Machine Learning: Mixture of Experts (MoE)

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Surface electromyography (sEMG) has demonstrated significant potential in simultaneous and proportional control (SPC). However, existing algorithms for predicting joint angles based on sEMG often suffer from high inference costs or are limited to specific subjects rather than multi-subject scenarios. To address these challenges, we introduced a hierarchical Spiking Attentional Feature Decomposition Network (SAFE-Net). This network initially compresses sEMG signals into neural spiking forms using a Spiking Sparse Attention Encoder (SSAE). Subsequently, the compressed features are decomposed into kinematic and biological features through a Spiking Attentional Feature Decomposition (SAFD) module. Finally, the kinematic and biological features are used to predict joint angles and identify subject identities, respectively. Our validation on two datasets and comparison with two existing methods, Informer and Spikformer, demonstrate that SSAE achieves significant power consumption savings of 39.1% and 37.5% respectively over them in terms of inference costs. Furthermore, SAFE-Net surpasses Informer and Spikformer in recognition accuracy on both datasets. This study underscores the potential of SAFE-Net to advance the field of SPC in lower limb rehabilitation exoskeleton robots.

Problem

• Extracting discriminative spatiotemporal features from EEG signals
• Integrating local-global relationships and spatiotemporal dynamics in EEG analysis
• Improving emotion recognition accuracy with multi-scale temporal-spatial fusion

Innovation

• Multi-scale temporal blocks for local-global dependencies
• Inverted Mamba structure for temporal-spatial fusion
• Four-channel EEG signals for high accuracy