Home › Research
These are my first-author projects, newest first. Papers I coauthored are listed under Publications.
Transformer Driven Visual Servoing for Fabric Texture Matching Using Dual-Arm Manipulator
Two patterned fabric pieces are stacked so that their patterns line up. A Transformer-based model estimates the pose offset from images, and dual-arm impedance control keeps the fabric taut during alignment and placement.
Robotic Sewing System for Fully Automated Fabric Picking, Stacking, Loading, Sewing, and Unloading
One system picks, stacks, loads, sews and unloads fabric pieces in sequence. It uses no dedicated fixtures and runs from a single PC.
Fixture-Free 2D Sewing Using a Dual-Arm Manipulator System
The two arms hold the fabric from both sides, adjust its tension and follow a seam line tracked by a camera. Two stacked pieces are sewn together without a fixture, including along curved seams.
CNN-based Visual Servoing for Simultaneous Positioning and Flattening of Soft Fabric Parts
A CNN takes camera images and force/torque measurements and outputs dual-arm motions that bring a fabric piece to its target position while removing wrinkles.
Convolutional Neural Network-Based Visual Servoing for Eye-to-Hand Manipulator
The robot compares a target image with the current image to estimate the motion needed to reach the target pose. This was my doctoral work and uses a camera fixed outside the robot (eye-to-hand).
I am combining generative models and imitation learning with visual feedback to make deformable object manipulation more precise. I am also looking at how visual feedback can be built into vision-language-action (VLA) models.
Telecommunications Advancement Foundation research grant, 2026–2028
I am working on combining mesh-based shape estimation with visual feedback control so that deformable objects can be handled more accurately.
NSK Foundation for the Advancement of Mechatronics research grant, 2026–2028