Research areas

1. Dexterous Robot Hands & Adaptive Grippers

Anthropomorphic robot hand
Flexure-based compliant joint
Underactuated mechanism
Variable stiffness gripper
Active friction modulation

The HANDS Lab develops next-generation robotic hands and grippers that achieve human-level dexterity through mechanically intelligent design. Our research spans flexure-based anthropomorphic hands with enhanced payload and shock resistance, underactuated wire-driven grippers with kinematic contact detection, and variable-stiffness Fin Ray grippers for handling delicate and irregular objects. We also explore bioinspired approaches — such as fluid-based active friction modulation inspired by human sweat and fingerprints — to enable stable grasping across diverse objects and environments.

2. Physical AI & Embodied Intelligence

Physical AI
Reinforcement learning (manipulation & dexterous hand control)
Sim-to-real transfer
Vision-Language-Action (VLA) models
Visual navigation / Aerial manipulation

We build embodied AI systems where intelligence emerges from the tight coupling of body, perception, and action. Our research includes reinforcement learning for manipulator trajectory generation and dexterous robot hand control, where policies are trained in NVIDIA Isaac Sim-based digital twin environments and deployed to physical hardware through sim-to-real transfer. We also explore Vision-Language-Action (VLA) models for language-conditioned manipulation, alongside monocular topological visual navigation (mono-MORP) and vision-guided aerial manipulation with UAV-mounted grippers. Combined with our mechanically intelligent hands and grippers, this line of work aims to realize Physical AI — robots that learn, adapt, and act intelligently in unstructured real-world environments.

3. Wearable Robots for Human Augmentation

Supernumerary robotic finger
Soft wearable glove
Hand function assistance
EMG-based control
Human–robot interface

Our wearable robotics research augments and restores human hand function. We developed a flexure-based supernumerary robotic finger that extends the capability of the human hand, and a ring-pull type soft wearable glove that assists grip strength for daily tasks. These systems are driven by intuitive bio-signal interfaces, including EMG-based intention recognition, building on our earlier work in brain–computer interfaces. Our goal is lightweight, comfortable wearable robots that feel like a natural extension of the body.

4. Soft Actuators & Bioinspired Mechanisms

Pneumatic soft actuator
Force enhancement design
Compliant mechanism
Bioinspired robot design
Soft robotics

We design soft actuators and compliant mechanisms that deliver both adaptability and force. Representative work includes a two-chamber pneumatic soft actuator with an expansion limit line for force enhancement, and adaptive compliant hinge structures for shape-conforming applications. Drawing inspiration from biological systems — from snake locomotion to human joint structures — we develop mechanisms that achieve complex, robust motion with simple and lightweight designs.

5. Soft Tactile Sensing & Contact Intelligence

Soft tactile sensor
Slip detection
Sensor-integrated joint
Multi-axis force sensing
Jamming-based contact enhancement

Robots need a sense of touch to grasp reliably. We develop soft tactile sensors that detect slip and multi-axis contact forces, and integrate them directly into 3D-printed flexible joints of robotic hands — enabling proprioceptive and tactile feedback without bulky external sensors. Our surface-conforming modular jamming pads further enhance contact stability and friction at the gripper–object interface. Together, these technologies form the sensing foundation for closed-loop, contact-intelligent manipulation.

Page updated

Google Sites

Report abuse