µ
Neuromorphic devices
Emerging ferroelectric and antiferroelectric transistors for compact neuronal dynamics, low-energy switching, and scalable integration.
Research themes
From device physics to intelligent systems.
Our work connects emerging nanoelectronic devices, compact circuits, event-driven sensing, and spiking neural networks for real-world edge intelligence.
↯
Brain-inspired circuits
CMOS and emerging-device neuron circuits that support leakage, integration, spike generation, reset, adaptation, and dynamic thresholding.
⌁
Spiking neural networks
Hardware-aware SNN models and benchmarks for sparse event-driven inference, low latency, and energy-efficient edge deployment.
✋
Neuromorphic sensing
Artificial tactile and wearable sensing systems that encode physical interactions into efficient temporal representations.
⚕
Wearable digital technologies
AI-assisted sensing for gait measurement, rehabilitation assessment, and digital biomarkers for clinical and health applications.
AI
Embodied and edge AI
System-level intelligence for robots, prosthetics, wearables, and physical-world AI constrained by energy, latency, robustness, and size.
Approach
Cross-layer co-design is the central principle.
We treat material properties, device dynamics, circuit behavior, and algorithmic requirements as a coupled design space. This allows us to transform physical device effects into useful computational primitives for compact, adaptive, and energy-efficient neuromorphic systems.
01
Materials and devices
Tune hysteresis, leakage, relaxation, steep-slope behavior, endurance, variability, and device operating windows.
02
Circuits and architectures
Design compact neuronal and synaptic building blocks that exploit device physics instead of compensating for it.
03
Sensors and interfaces
Build tactile and wearable sensor systems that generate sparse, temporal, and biologically inspired signal representations.
04
Algorithms and benchmarks
Evaluate accuracy, energy per inference, latency, robustness, scalability, and relevance for edge and embodied AI.
Selected projects
Research directions.
The lab develops core neuromorphic hardware while applying event-driven intelligence to robotics, wearables, and health-related sensing.
Flagship direction
Brain-like artificial neurons with circuit-device co-design
We develop compact artificial neuron circuits using antiferroelectric/ferroelectric transistor dynamics to achieve rich neuronal functions, high efficiency, and scalable hardware implementation for spiking neural networks.
Neuromorphic tactile intelligence
Artificial touch systems for robotic and prosthetic perception using tactile sensors, artificial neurons, receptive-field processing, and SNNs.
AI-driven digital biomarkers
Wearable sensing and machine learning for gait measurement, rehabilitation assessment, and longitudinal monitoring of disease progression.
Sustainable edge intelligence
Low-power sensors and event-driven AI for physical-world monitoring, sustainable manufacturing, and resource-aware intelligent systems.