Battery-Free Wearable Sensor Enables Activity Recognition for Health Monitoring

Neuromorphic devices are engineered systems that emulate functions of biological neural networks. Wearable versions could enable low-power patient monitoring, but many existing designs still depend on external power sources, limiting their practicality in flexible, body-worn systems. To overcome this barrier, researchers have developed a battery-free neuromorphic sensing platform that converts mechanical movement into synaptic electrical activity.

The technology is a self-powered, flexible graphene-channel ion-gel-gated transistor, or g-IGT, driven by a triboelectric nanogenerator. The work was led by Dongguk University (Seoul, South Korea). The device is designed to reproduce tactile sensing processes in which mechanical stimuli are converted into electrical signals. It combines energy harvesting, sensing, memory, and learning functions within a flexible platform.

The system uses two triboelectric nanogenerators connected to a single g-IGT. One nanogenerator supplies pre-synaptic spikes to the gate, while the other supplies post-synaptic spikes to the drain side. When mechanical stimuli such as touch are detected, the nanogenerators produce voltage pulses that regulate the synaptic response. The circuit operates without an external power source by harvesting energy directly from mechanical stimuli.

The device demonstrated hierarchical memory behavior. It showed sensory memory with a decay time of about 70 milliseconds and short-term memory with decay times of 0.2 to 0.45 seconds. Repeated stimulation shifted short-term memory toward long-term memory, with a decay time exceeding 2 seconds. The device also demonstrated spike-rate-dependent plasticity, and this function remained stable under bending.

The researchers further evaluated learning capability by applying experimentally measured synaptic behavior in a single-layer artificial neural network for human activity recognition. Using publicly available human-motion data, the system classified walking, sitting, standing, lying, walking upstairs, and walking downstairs. It achieved 88.05% accuracy using TENG-driven synaptic behavior under bending. Under high-noise conditions, accuracy remained above 75%, although performance declined under extreme signal distortion.

The study, titled “Self-Powered Flexible Triboelectric-Gated Ion-Gel Transistor for Neuromorphic Tactile Sensing and Human Activity Recognition,” was published in Advanced Materials. Potential applications include self-powered wearable health-monitoring devices, electronic skin, smart prosthetics, human-machine interfaces, and intelligent motion-monitoring systems.

“In human tactile perception mechanoreceptors sense even minute mechanical disturbances and convert them into neural spikes. To replicate this process electronically, we integrated a triboelectric nanogenerator with a g-IGT that converts mechanical stimuli into electrical signals that directly regulate artificial synaptic behavior without requiring external power,” said Professor Sejoon Lee, Department of System Semiconductor, Dongguk University. 

“Our research could contribute to a new generation of wearable artificial intelligence systems that operate with minimal reliance on batteries or external computing resources. More broadly, our work points toward self-powered neuromorphic electronics with integrated sensing, memory, learning, and information processing in a single flexible platform,” added Prof. Lee.

Related Links
Dongguk University