A Parallel Ultra-Low Power Silent Speech Interface based on a Wearable, Fully-dry EMG Neckband

2025-09-26

AI Summary

Addressing the challenge of simultaneously achieving user comfort and ultra-low power consumption in silent speech recognition, this work presents a fully dry-electrode, ultra-low-power electromyography (EMG) system integrated into a textile neckband. Built upon the BioGAP-Ultra platform, the system features a 14-channel fully differential biosignal acquisition circuit with a total power consumption of only 22 mW and integrated wireless transmission capability. Gel-free dry electrodes enable non-invasive, long-term wear without skin irritation or signal drift. In binary classification tasks distinguishing vocalized versus silent articulation, the system achieves session-wise average accuracies of 87±3% and 68±3%, respectively; cross-session evaluation demonstrates robust performance at 64±18% and 54±7%. To the best of our knowledge, this is the first demonstration of deep integration of an ultra-low-power, fully dry-electrode EMG system into a functional wearable textile platform, establishing a novel paradigm for practical silent speech interfaces.

Technology Category

• Natural Language Processing: Speech
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Machine Learning: Hardware-aware ML

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search

Abstract

We present a wearable, fully-dry, and ultra-low power EMG system for silent speech recognition, integrated into a textile neckband to enable comfortable, non-intrusive use. The system features 14 fully-differential EMG channels and is based on the BioGAP-Ultra platform for ultra-low power (22 mW) biosignal acquisition and wireless transmission. We evaluate its performance on eight speech commands under both vocalized and silent articulation, achieving average classification accuracies of 87$pm$3% and 68$pm$3% respectively, with a 5-fold CV approach. To mimic everyday-life conditions, we introduce session-to-session variability by repositioning the neckband between sessions, achieving leave-one-session-out accuracies of 64$pm$18% and 54$pm$7% for the vocalized and silent experiments, respectively. These results highlight the robustness of the proposed approach and the promise of energy-efficient silent-speech decoding.

Problem

• Developing wearable EMG neckband for silent speech recognition
• Achieving ultra-low power consumption for biosignal acquisition
• Evaluating classification accuracy under vocalized and silent conditions

Innovation

• Wearable fully-dry EMG neckband for silent speech
• Ultra-low power BioGAP platform with wireless transmission
• Multi-channel EMG system achieving robust silent recognition