Researchers develop battery-free wearable device powered by human movement
Researchers develop battery-free wearable device powered by human movement
Researchers at Dongguk University in South Korea have developed a flexible, battery-free electronic device that harvests energy from human movement to power neuromorphic sensing and learning functions. The device demonstrates artificial memory capabilities and activity recognition with 88.05% accuracy, potentially enabling self-powered wearable AI systems for health monitoring and human-machine interfaces.
Source: Robotics & Automation News · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
Reported accuracy
88.05%
View original evidence
ation could also move the system toward a longer-term memory state lasting more than two seconds. The device was additionally tested using an artificial neural network designed to recognize six human activities, including walking, sitting, standing and climbing stairs. Using the experimentally measured behavior of the device, the system achieved 88.05 percent accuracy in classifying the activities while the flexible
Open source S6short-term memory duration
0.2–0.45 seconds
View original evidence
ate 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.” The researchers demonstrated several forms of artificial memory, ranging from sensory memory lasting around 70 milliseconds to short-term memory lasting 0.2-0.45 seconds. Repeated stimul
Open source S5sensory memory duration
≈70
View original evidence
ate 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.” The researchers demonstrated several forms of artificial memory, ranging from sensory memory lasting around 70 milliseconds to short-term memory lasting 0.2-0.45 seconds. Repeated stimul
Open source S5number of human activities recognized
6
View original evidence
ation could also move the system toward a longer-term memory state lasting more than two seconds. The device was additionally tested using an artificial neural network designed to recognize six human activities, including walking, sitting, standing and climbing stairs. Using the experimentally measured behavior of the device, the system achieved 88.05 percent accuracy in classifying the activities while the flexible
Open source S6research publication year
2,026
View original evidence
broadly, our work points toward self-powered neuromorphic electronics with integrated sensing, memory, learning, and information processing in a single flexible platform.” The research was published in Advanced Materials in July 2026.
Open source S8
- paper: not_reported
- code: not_reported
- weights: not_reported
- dataset: not_reported
What remains unknown
Not established in the collected evidence: Environment, Control, Data origin.
Reported performance applies to the described task. It does not establish general autonomy or deployment readiness.
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
ation could also move the system toward a longer-term memory state lasting more than two seconds. The device was additionally tested using an artificial neural network designed to recognize six human activities, including walking, sitting, standing and climbing stairs. Using the experimentally measured behavior of the device, the system achieved 88.05 percent accuracy in classifying the activities while the flexible
Open source S6
ate 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.” The researchers demonstrated several forms of artificial memory, ranging from sensory memory lasting around 70 milliseconds to short-term memory lasting 0.2-0.45 seconds. Repeated stimul
Open source S5
broadly, our work points toward self-powered neuromorphic electronics with integrated sensing, memory, learning, and information processing in a single flexible platform.” The research was published in Advanced Materials in July 2026.
Open source S8
Source：Robotics & Automation News · roboticsandautomationnews.com