This work introduces a silent speech interface (SSI), proposing a few-layer graphene (FLG) strain sensing mechanism based on thorough cracks and AI-based self-adaptation capabilities that overcome the limitations of state-of-the-art technologies by simultaneously achieving high accuracy, high computational efficiency, and fast decoding speed while maintaining excellent user comfort. We demonstrate its application in a biocompatible textile-integrated ultrasensitive strain sensor embedded into a smart choker, which conforms to the user’s throat. Thanks to the structure of ordered through cracks in the graphene-coated textile, the proposed strain gauge achieves a gauge factor of 317 with <5% strain, corresponding to a 420% improvement over existing textile strain sensors fabricated by printing and coating technologies reported to date. Its high sensitivity allows it to capture subtle throat movements, simplifying signal processing and enabling the use of a computationally efficient neural network. The resulting neural network, based on a one-dimensional convolutional model, reduces computational load by 90% while maintaining a remarkable 95.25% accuracy in speech decoding. The synergy in sensor design and neural network optimization offers a promising solution for practical, wearable SSI systems, paving the way for seamless, natural silent communication in diverse settings.

Efficient operation of control systems in robotics or autonomous driving targeting real-world navigation scenarios requires perception methods that allow them to understand and adapt to unstructured environments with good accuracy, adaptation, and generality, similar to humans. To address this need, we present a memristor-based differential neuromorphic computing, perceptual signal processing, and online adaptation method providing neuromorphic style adaptation to external sensory stimuli. The adaptation ability and generality of this method are confirmed in two application scenarios: object grasping and autonomous driving. In the former, a robot hand realizes safe and stable grasping through fast ( ~ 1 ms) adaptation based on the tactile object features with a single memristor. In the latter, decision-making information of 10 unstructured environments in autonomous driving is extracted with an accuracy of 94% with a 40×25 memristor array. By mimicking human low-level perception mechanisms, the electronic neuromorphic circuit-based method achieves real-time adaptation and high-level reactions to unstructured environments.

Background: Ambulatory heart rate (HR) monitors that acquire electrocardiogram (ECG) or/and photoplethysmographm (PPG) signals from the torso, wrists, or ears are notably less accurate in tasks associated with high levels of movement compared to clinical measurements. However, a reliable estimation of HR can be obtained through data fusion from different sensors. These methods are especially suitable for multimodal hearable devices, where heart rate can be tracked from different modalities, including electrical ECG, optical PPG, and sounds (heart tones). Combined information from different modalities can compensate for single source limitations. Methods: In this paper, we evaluate the possible application of data fusion methods in hearables. We assess data fusion for heart rate estimation from simultaneous in-ear ECG and in-ear PPG, recorded on ten subjects while performing 5-min sitting and walking tasks. Results: Our findings show that data fusion methods provide a similar level of mean absolute error as the best single-source heart rate estimation but with much lower intra-subject variability, especially during walking activities. Conclusion: We conclude that data fusion methods provide more robust HR estimation than a single cardiovascular signal. These methods can enhance the performance of wearable devices, especially multimodal hearables, in heart rate tracking during physical activity.

This paper presents a novel wireless silent speech interface (SSI) integrating multi-channel textile-based EMG electrodes into headphone earmuff for real-time, hands-free communication. Unlike conventional patch-based EMG systems, which require large-area electrodes on the face or neck, our approach ensures comfort, discretion, and wearability while maintaining robust silent speech decoding. The system utilizes four graphene/PEDOT:PSS-coated textile electrodes to capture speech-related neuromuscular activity, with signals processed via a compact ESP32-S3-based wireless readout module. To address the challenge of variable skin-electrode coupling, we propose a 1D SE-ResNet architecture incorporating squeeze-and-excitation (SE) blocks to dynamically adjust per-channel attention weights, enhancing robustness against motion-induced impedance variations. The proposed system achieves 96% accuracy on 10 commonly used voice-free control words, outperforming conventional single-channel and non-adaptive baselines. Experimental validation, including XAI-based attention analysis and t-SNE feature visualization, confirms the adaptive channel selection capability and effective feature extraction of the model. This work advances wearable EMG-based SSIs, demonstrating a scalable, low-power, and user-friendly platform for silent communication, assistive technologies, and human-computer interaction.

It is predicted that the global shipment of smart wearables will approach 302.2 million devices in 2023, increasing from 222.9 million devices in 2019 [1] . It was also forecast in 2019 that the number of ear-worn devices—so-called hearables—would rise to 105.3 million in 2023, from 72 million in 2019 [2] . Given the relatively fixed position of the head with respect to the brain and vital organs in most of the daily activities, hearables provide more consistent recordings compared to more mobile parts of the body, such as the wrists. This allows for robust recordings of the both electroencephalogram (EEG) [3] , [4] , [5] , electrocardiogram (ECG) [6] , and photoplethysmogram (PPG) [7] , together with the derived measures including the heart rate (HR) [4] , respiratory rate [8] , blood oxygen saturation (SpO2) [7] , and blood glucose levels [9] .